**4. Results**

### *4.1. Summary of Studies that Report PR Cell Loss*

The combined searches of MGI and PubMed databases yielded a total of 230 genes associated with PR cell loss. Ultimately, 3834 reports at MGI and 3325 at PubMed, which most typically characterized one mutant gene but on rare occasions described more than one, were used in the present review. The distribution of retrieved publications sorted by functional categories is summarized in Table S1. The genes identified in these models are summarized in Figure 3. Descriptions of gene and protein symbols used in the text, figures, and Table S1 are provided in Table S2.

**Figure 3.** Genes associated with PR cell loss in monogenic mouse models of retinal degeneration (RD). Genes identified by combined review of the Mouse Genome Informatics (MGI) database and articles from a PubMed query were assigned to the indicated functional categories as described in the text. Genes for which mutant alleles are available only in the conditional form are displayed in red. Conditional alleles were included only in instances where germline null alleles resulted in embryonic, prenatal or postnatal lethality. For additional details on inclusion/exclusion criteria, see Section 3.4.

### 4.1.1. PR Cell Loss Models

The mouse models described in Table S1 were either spontaneous (12%) or chemically induced mutants (11%), or those produced through genetic engineering approaches (77%). This latter group, which was by far the largest, utilized standard homologous recombination, gene-traps, nuclease mediated approaches such as CRISPR/Cas9, and conditionals to mediate genomic changes. Additionally, four models of inadvertent transgene insertion into a unique gene, whose disruption led to PR degeneration, were included within this group. Interesting examples of differences in the disease onset or rate of progression were demonstrated in different models of the same gene (e.g., *Aipl1*) that may be related to allelic differences, null versus missense mutations, or genetic background effects [21,39,40]. Most of the genetically engineered models in Table S1 tended to be in mixed, segregating genetic backgrounds that might impact phenotypic expression (discussed below).

Within the genetically engineered category, a relatively large group of models, 16%, were conditional models, representing 39 genes (Figure 3; red). Generation of conditional mutants is based on the Cre-Lox recombination approach, which requires a floxed gene and a cre-driver to excise the targeted genomic region in a spatial (e.g., cell/tissue specific) or temporal manner (e.g., induction by chemicals such as doxycycline). Since 30% of all null mutations lead to embryonic lethality, as they represent genes that are essential during development, conditionals are often used to examine the adult function of genes [41]. This was the case in 92% of the conditional models described here, as standard organism-wide removal of the genes was reported to be embryonic, perinatal, or postnatal lethal. Thus, conditionals allow us to learn the function of a gene post-developmentally. Conditionals are also sometimes used to determine the cellular contributions to a disease phenotype. If a gene is expressed in multiple retinal cell types, by removing them systematically and examining the consequent phenotype, one can learn how the loss of function of the gene within particular cell types affects the disease phenotype. For example, removal of *Arl3* from rod PRs using a Rho-icre driver shows a later and slower rate of degeneration than that found with Six3-cre, a Cre driver that expresses in early retinal development. This suggests that *Arl3* in rods is necessary for PR survival but that *Arl3* function in other retinal cell types also affects PR survival [42]. The most widely used Cre models include: for targeting retinal progenitor cells, Tg(rx3-icre)1Mjam, Tg(Six3-cre)69Frty, Tg(Chx10-EGFP/cre,-ALPP)2Clc, Tg(Crx-cre)1Tfur, and Tg(Pax6-cre,GFP)2Pgr; for targeting rods, Tg(Rho-icre)1Ck, Tg(RHO-cre)8Eap, and *Pde6gtm1(cre*/*ERT2)Eye*; for targeting M-cone PRs, Tg(OPN1LW-cre)4Yzl (also known as HRGP-cre); for targeting PRs, Tg(Rbp3-cre)528Jxm (also known as IRBP-cre); for targeting RPE, Tg(BEST1-cre)1Jdun, Tg(BEST1-rtTA,tetO-cre)1Yzl and *Foxg1tm1(cre)Skm*; and for targeting adult tissues using tamoxifen, Tg(CAG-cre/Esr1\*)5Amc.

### 4.1.2. Mouse Models from Phenotyping Programs

The models listed in Table S1 come from many sources. In addition to individual investigator-initiated efforts, currently the largest contributor to ocular models is the International Mouse Phenotyping consortium, in which 19 phenotyping centers from 11 countries participate to systematically characterize knockout mice generated in a standardized manner [41,43]. All centers do some eye phenotyping, thus providing a window into potential models. Although only a few models from this program are included in Table S1, as most are not ye<sup>t</sup> fully characterized, it is anticipated that this consortium will provide a wealth of models for individual laboratories to study. For example, in the MGI database, 39 IMPC models were identified with "reduced retinal thickness" that with further characterization may reveal PR degeneration.

At The Jackson Laboratory, the Eye Mutant Resource (EMR) and the Translational Vision Research Models (TVRM) programs are dedicated to screen for or generate mouse models with ocular diseases. The EMR has been screening retired breeders by slit lamp biomicroscopy, indirect ophthalmoscopy, and electroretinography since 1988. Retired breeders from the production and genetic resources colonies are screened. Heritable mutants are phenotypically and genetically characterized and the spontaneous mutants are distributed worldwide. The TVRM program arose from the JAX Neuromutagenesis Facility. Mice for this program are generated by chemical mutagenesis or genetic engineering. Carefully characterized mutants are also distributed. Examples of mutants from the EMR and TVRM programs are shown in Figures 4 and 5, respectively.

(**a**) 

(**b**) 

**Figure 4.** *Cont.*

(**c**) 

**Figure 4.** (**a**) Characterization of mouse models from the Eye Mutant Resource (EMR) program at JAX. Example optical coherence tomography (OCT) and fundus images were taken from rapid RD models: *Pde6brd1* (B6.C3-*Pde6brd1 Hps4le*/J, Stock No: 000002), *Pde6brd1-2J* (C57BL/6J-*Pde6brd1-2J*/J, Stock No: 004766), *Pde6brd10* (B6.CXB1-*Pde6brd10*/J, Stock No: 004297), *Rd4*/+ (STOCK In(4)56Rk/J, Stock No: 001379) and *Cep290rd16* (B6.Cg-*Cep290rd16*/Boc, Stock No: 012283) (**b**) Example images from slower RD models: *Prph2Rd2* (C3A.Cg-*Pde6b*+ *Prph2Rd2*/J, Stock No: 001979), *Rd3rd3* (B6.Cg-*Rd3rd3*/Boc, Stock No: 008627), *Lpcat1rd11* (B6.Cg-*Lpcat1rd11*/Boc, Stock No: 006947), *Rpe65rd12* (B6(A)-*Rpe65rd12*/J, Stock No: 005379) and *Prom1rd19* (B6.BXD83-*Prom1rd19*/Boc, Stock No: 026803). (**c**) Examples from slow and very slow RD models: *Mfrprd6* (B6.C3Ga-*Mfrprd6*/J, Stock No: 003684), *Nr2e3rd7* (B6.Cg-*Nr2e3rd7*/J, Stock No: 004643), *Crb1rd8* (STOCK *Crb1rd8*/J, Stock No: 003392), *RpgrRd9* (C57BL/6J-*RpgrRd9*/Boc, Stock No: 003391), and *Gnat1rd17* (B6.Cg-*Gnat1irdr*/Boc, Stock No: 008811). *Yellow bars* indicate full retinal thickness. Values correspond to the mouse age at the time of imaging (weeks).
