**3. Results**

### *3.1. Generation of a Floxed Dhdds Mouse Line*

The scheme used for the generation of a *Dhdds* conditional allele, employing a validated *Dhdds* construct (from KOMP) and mouse embryonic stem cells (ESCs), is shown in Figure 1 (see *Materials and Methods*, above). Clones from confirmed flippase recognition target (FRT)-excised alleles were used to generate *Dhdds* heterozygous and homozygous mouse lines on a C57Bl/6J background.

**Figure 1.** Generation of Cre-dependent *Dhdds* conditional knockout (KO) mice. (**A**) A validated *Dhdds* construct from the Knockout Mouse Project (KOMP) (U.C. Davis) was linearized and introduced into mouse ESCs. (**B**) Transformed cells were treated with FLP-FRT recombinase and PCR was used to verify lacZ cassette excision. (**C**) Clones from confirmed FRT-excised alleles were used to generate *Dhddsflx*/*flx-* mice. (**D**) Pups carrying the *Dhdds* floxed allele were identified by coat appearance.

### *3.2. Validation of Retinal Cell Type-Specific Cre-Expressing Mouse Lines*

To assess the specificity and efficiency of the Cre recombinase in the RPE, we cross-bred transgenic mice expressing Cre recombinase (bred to homozygosity) with a ZsGreen Ai6 reporter mouse line, and then evaluated ZsGreen expression in the retina by confocal fluorescence microscopy. As expected, WT mouse retinas (a negative control) did not exhibit ZsGreen expression (Figure 2A). However, ZsGreen fluorescence was detected in Cre recombinase-positive mice specifically in the RPE layer by 1 postnatal (PN) month (Figure 2B), with >90% of the RPE cells being labeled. We subsequently confirmed that Cre recombinase continued to be expressed robustly and specifically in the RPE for >3 months (data not shown).

**Figure 2.** Cre recombinase-dependent ZsGreen expression (green) in mouse retina. (**A**) Retina from a postnatal (PN) one-month old CreRPE × ZsGreen reporter mouse, demonstrating ZsGreen expression specifically in retinal pigment epithelium (RPE) cells. (**B**) Retina from an age-matched, wild type (WT) mouse retina, demonstrating lack of ZsGreen expression. Nuclei counterstained with DAPI (blue). Abbreviations: IS, photoreceptor inner segmen<sup>t</sup> layer; OS, photoreceptor outer segmen<sup>t</sup> layer; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars (both panels): 20 μm.

*3.3. RPE-Specific Ablation of Dhdds Causes a Geographic Atrophy-Like Phenotype and Retinal Degeneration, Involving Photoreceptors*

In vivo retinal imaging using SD-OCT (Figure 3) showed comparable normal layer stratification in WT, CreRPE and *Dhdds*+/*flx* CreRPE age-matched mice in each group. However, *Dhddsflx*/*flx* CreRPE mice showed altered hyper-reflectivity at all ages (indicated by red arrows, Figure 3), indicative of pathologic changes and a reduction in outer retinal layer thickness.

**Figure 3.** SD-OCT revealed structural changes in *Dhddsflx*/*flx* CreRPE mice at all ages. OCT scans were performed at 1, 2, and 3 months (m) postnatal for each genotype. (**A**) WT, (**B**) CreRPE, (**C**) *Dhdds*+/flx CreRPE all showed normal layer stratification. (**D**) *Dhddsflx*/*flx* CreRPE showed alterations of hyper-reflectivity at all ages, indicative of pathologic changes (red arrows) and reduction in layer thickness, particularly in the outer retina. IPL, inner plexiform layer, ROS, rod outer segment. Scale bar (shown in WT 1 m panel): 50 μm, applies to all panels.

"Spidergram" plots of average thickness values (in mm) vs. retinal eccentricity (distance from the optic nerve head (ONH) along with vertical meridian) for the outer nuclear layer (ONL) and FRT are shown in Figure 4. No significant differences were observed when comparing WT, heterozygous, or Cre-only mice for both ONL and FRT measurements. However, *Dhddsflx*/*flx* CreRPE mice showed a significant reduction (vs. WT) in ONL and F.R.T. values at all ages analyzed (n ≥ 4, all *p* < 0.001).

Histologically, at PN 3 months, light micrograph images revealed that all retina layers in mice lacking Cre expression appeared normal (Figure 5A and Supplementary Materials, Figure S1A). In contrast, the age-matched Dhddsflx/flx Cre RPE mice displayed a geographic atrophy-like RPE appearance with the most degeneration observed mid-centrally throughout the retina (Supplementary Materials, Figure S1B). There were regions of well-preserved Dhddsflx/flx Cre RPE retina that showed severe RPE pathology (Figure 5B). Notably, the descent of the external limiting membrane (ELM) [25] towards Bruch's membrane was also observed (Figure 5C,E). There was a near-total loss of photoreceptors and RPE within the most affected regions (Figure 5D). Very well-preserved regions were also observed in the periphery (Figure 5F).

**Figure 4.** Quantitative morphometric analysis of WT, CreRPE, *Dhdds*+/flx CreRPE, and *Dhddsflx*/*flx* CreRPE mice SD-OCT data. Average OCT measurements at each eccentricity revealed no significant differences when comparing WT, CreRPE and *Dhdds*+/flx CreRPE mice for both outer nuclear layer (ONL) and F.R.T. thickness measurements (n = 4 for all genotypes, except for WT at one month (1 m; n = 8) and two months (2 m; n = 5). However, significant changes were observed when comparing WT and *Dhddsflx*/*flx* CreRPE mice at any age with respect to both ONL and F.R.T. values.

**Figure 5.** Pathology observed in *Dhddsflx*/*flx* Cre RPE mouse retina. (**A**) *Dhddsflx*/*flx* retina without Cre expression appears normal. (**B**–**F**) Five regions of a retina from a *Dhddsflx*/*flx* CreRPE mouse expressing Cre in RPE are shown. (**B**,**F**) Relatively well-preserved peripheral retina with some photoreceptor loss, outer segmen<sup>t</sup> (OS) shortening, and differing severity of RPE pathology are shown (arrow in panel **B** points to severely compromised RPE. (**C**,**E**) Transition zones of severe to mild retinal pathology showing loss of photoreceptors, severe compromise of RPE and external limiting membrane (ELM) descent (arrows). (**D**) A more central region of the retina showing severe cell loss in both the retina and RPE. Scale bars (all panels, except **F**): 20 μm; scale bar, panel **F**: 10 μm. ROS, rod outer segments; ONL, outer nuclear layer, INL, inner nuclear layer; IPL, inner plexiform layer, GCL, ganglion cell layer.

Additional pathology was revealed by higher magnification EM analysis (Figure 6, panels C–O). In contrast, *Dhddsflx*/*flx* retinas without Cre expression were indistinguishable from WT retinas (Figure 6A,B). The two areas most affected showed severe RPE dystrophy with concomitant degeneration and loss of photoreceptor cells (Figure 6C,D,F,G,J,M–O). ELM descent was apparent in areas where there was a transition from milder to more severe pathology (\* in panels D–F,J). RPE cell transmigration was apparent in the outer retina (arrows, panels G,J,O,L). Thus, compared to WT neural retina and RPE, the observed RPE anomalies including migration of nucleus and RPE melanosomes, displacement of the ELM, and shortened misshaped outer segments throughout the retina were seen in the *Dhddsflx*/*flx CreRPE* mice.

**Figure 6.** High magnification observation of *Dhddsflx*/*flx* CreRPE retina. EM images of 3m (**A**,**B**) *Dhddsflx*/*flx* and (**C**–**O**) *Dhddsflx*/*flx* CreRPE mice. Mice homozygous for the floxed *Dhdds* allele (*Dhddsflx*/*flx*) are not distinguishable from WT. (**A**) Rod outer segments are properly aligned and all retinal layers are intact (only rod inner and outer segments and the outer nuclear layer are shown). (**B**) The RPE shows normal thickness and melanin distribution. (**C**–**O**) Examples of mildly and severely compromised RPE and retina in *Dhddsflx*/*flx* CreRPE mice. Severe RPE and outer retina degeneration is concentrated mid-centrally on both sides of the optic nerve head (see histology, Figure 5, and Supplementary Materials, Figure S1). Severe RPE/PR atrophy is observed in the most affected central regions (**C**–**F**,**I**,**J**,**O**). External limiting membrane (ELM) descent (\*) towards the RPE is apparent in the transition regions from severely affected to more intact regions (**D**–**F**,**J**). Transmigration of RPE cellular material into the ROS space (arrows) is also seen (**G**,**J**,**L**,**O**). Extended RPE basal fenestrations (arrowheads, **I**) are apparent in the most affected regions (**D**,**I**). Outside of the central region of severe degeneration are regions with compromised but still apparent retinal and RPE layers (**K**–**N**). Scale bars (all panels, except **I**): 10 μm; panel **I**, 2.5 μm.

To further examine the effects of *Dhdds* deletion in the RPE, serial block face-scanning electron microscopy was used to examine a ~100 μ3 region of the *Dhddsflx*/*flx* retina across an area of transition from milder to more severe pathology (Supplementary Materials, Videos S1–3). In WT mice, the retina appeared normal in all layers across the entire block. Examination of two regions of the *Dhddsflx*/*flx* Cre RPE retina showed areas of severe compromise as well as areas where retinal histology was more well-preserved. Transmigration of RPE nuclei and melanosomes into the photoreceptor region (subretinal space and photoreceptor outer segmen<sup>t</sup> layer, and even deeper, into the ONL) was also apparent.

### *3.4. Altered Scotopic and Photopic ERG Amplitudes in Dhddsflx*/*flx CreRPE Mice*

Analysis by ERG (Figure 7) showed that both homozygous and heterozygous *Dhddsflx*/*flx* CreRPE mice exhibited reduced scotopic ERG a- and b-wave responses. At PN 1 month, scotopic a-wave responses (289 ± 42 μV; n = 5) were significantly reduced in *Dhdds*flx/flx CreRPE mice compared to WT mice (395 ± 14 μV; n = 8; *p* < 0.05). Scotopic b-wave responses also were reduced for both *Dhdds*+/*flx* CreRPE (772 ± 42 μV; n = 20) and *Dhddsflx*/*flx* CreRPE (479 ± 69 μV; n = 5) mice, compared to WT mice (934 ± 38 μV; n = 8; *p* < 0.01). At PN 2 months of age, *Dhddsflx*/*flx* CreRPE scotopic ERG a-wave (223 ± 48 μV; n = 10) and b-wave (477 ± 88 μV; n = 10) responses and *Dhdds*+/*flx* CreRPE a-wave (305 ± 19 μV; n = 10) and b-wave (557 ± 64 μV; n = 10) responses were significantly reduced compared to the WT a-wave (366 ± 10 μV; n = 15) and b-wave (935 ± 27 μV; n = 15) responses (*p* < 0.005 for all comparisons). At PN 3 months, scotopic a-wave amplitudes were reduced only in *Dhdds*flx/flx CreRPE (60 ± 48 μV; n = 6) compared to WT mice (354 ± 16 μV; n = 18; *p* < 0.001); b-wave responses were reduced in both *Dhdds*+/*flx* CreRPE (662 ± 88 μV; n = 6) and *Dhddsflx*/*flx* CreRPE (150 ± 105 μV; n = 6) mice compared to WT mice (914 ± 35 μV; n = 18) *p* < 0.05 and *p* < 0.01, respectively).

**Figure 7.** Scotopic ERG analysis of WT, *Dhdds*+/*flx* CreRPE, *Dhddsflx*/*flx* CreRPE mice. Maximum responses to saturating light stimuli showed significant decreases in ( **A**) a-wave and (**B**) b-wave amplitudes for *Dhdds*+/*flx* CreRPE and *Dhddsflx*/*flx* CreRPE when compared to WT mice at 1, 2 and 3 postnatal months. Statistical significance: \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

The photopic a-wave and b-wave responses (Figure 8) at PN 1 month were not di fferent when comparing WT (n = 5) and *Dhdds*+/*flx* CreRPE (n = 13) mice, however the photopic responses of *Dhddsflx*/*flx* CreRPE mice (n = 4) were significantly lower. Photopic ERG responses were significantly di fferent only for *Dhdds*+/*flx* CreRPE a-wave (29 ± 2 μV; n = 4; *p* < 0.01), but not b-wave, responses at PN 2 months of age. In *Dhddsflx*/*flx* CreRPE mice, b-wave (86 ± 19 μV; n = 6) amplitudes were significantly reduced (*p* < 0.01), compared to WT mice (n = 13), but not the a-wave responses. At PN 3 months, *Dhddsflx*/*flx* CreRPE mice exhibited significant functional impairment in the photopic a-wave (2 ± 1 μV; n = 4) and b-wave (7 ± 1 μV; n = 4) responses compared to WT a-wave (14 ± 1 μV; n = 9) and b-wave responses (144 ± 8 μV; n = 9; *p* < 0.001 for all comparisons).

Optokinetic reflex (OKR) analysis (Supplementary Materials, Figure S3), a measure of retina-tobrain transmission (i.e., visual capacity), showed reductions in photopic (4–31%) and scotopic (8–29%) contrast sensitivity over the range of 0.031 to 0.272 c/d in PN 3-month old *Dhddsflx*/*flx* CreRPE mice (*p* < 0.05), compared to WT controls of the same age. However, no di fferences in spatial frequency (a measure of visual acuity) were observed between the di fferent mouse lines. This is partly evident in the OKR scotopic and photopic plots, which show a similar high-frequency cut-o ff for all three mouse lines examined.

**Figure 8.** Photopic ERG analysis of WT, CreRPE *Dhdds*+/*flx*, CreRPE *Dhddsflx*/*flx* mice. (**A**) a-wave responses were significantly lower in CreRPE *Dhdds-*/*-* at 1 (n = 4) and 3 (n = 4) months postnatal compared to WT (one month (1 m), n = 5; three months (3 m), n = 9) mice. (**B**) Photopic b-wave amplitudes for *Dhdds*+/*flx* CreRPE mice showed no statistically significant differences when compared to WT mice at 1 (n = 13) and 2 (n = 4) postnatal months, but were reduced by 3 months (n = 4). In contrast, the responses of *Dhddsflx*/*flx* CreRPE mice (two months (2 m), n = 6) were significantly lower when compared to WT mice at all postnatal ages examined. Statistical significance: \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.
