**5. Conclusions and Perspectives**

Prph2 plays a key role in the maintenance as well as the development of photoreceptor OS. While Prph2 is found to be vital for both rods and cones, there seem to be distinct differences in its function in these two types of photoreceptor cells. A complete knockout of Prph2 results in the complete absence of ROS while COS formation is still initialized even though the resulting COS are severely disorganized and lack lamellae and discs. This begs the question, which protein is mediating the formation of the COS and if it works in an interplay with Prph2. The OS of both rods and cones represent a highly modified primary cilium. The Prph2 mediated suppression of ectosomes shedding from the photoreceptor cilium was found to be a prerequisite for the formation of the ROS. Thus far there is no study demonstrating a similar mechanism in the formation of COS. The investigation of the formation of ectosomes and membrane dynamics in the development of COS in WT and *Prph2*−/− mice might be a potential approach for further studies aiming to unravel the different function of Prph2 in ROS versus COS. While the function of Prph2 in the initialization of ROS and COS varies, it seems to be indispensable for the correct shaping, sizing and stacking of discs and lamellae in both. Furthermore, not only the presence of Prph2 is needed for developing and maintaining these structures but also the precise regulation in the formation of the different Prph2/Rom1 and Prph2/Prph2 complexes. Changes in the ratio of the tetramers, intermediate complexes and higher order complexes were proven detrimental to the function as well as the structure of the photoreceptors. The transgenic and knockin mouse models discussed here proved the importance of the precise regulation of the complex formation. Each of the knockin mouse models carrying a pathogenic *Prph2* mutation displayed an altered ratio of the different Prph2/Rom1 and Prph2 complexes resulting in disorganized OS structure and decreased response found in ERG measurements. Interestingly rods and cones were differently affected by the alterations in complex formation, providing further evidence for the differential use of Prph2 by the two photoreceptor cell types. In addition to that, the mouse models used helped identify whether a mutation results in a loss-of-function or a gain-of-function effect.

While there is agreemen<sup>t</sup> in the numerous studies concerning the ratio between tetramers, intermediate and higher order complexes and its vital importance for OS structure and function, the precise roles of the different complexes are still not fully understood. In the C150S knockin mouse model, both Prph2 and Rom1 are found exclusively as tetramers, while intermediate and higher order complexes are absent. Still the formation of both ROS and COS is initiated, proving that the tetramers alone are sufficient to initiate OS formation. Homozygous animals fail to form disc and lamellae, indicating that the intermediate and higher order complexes are more likely to be involved in the formation of the rim, membrane closure and disc stacking. The knockout of Rom1 showed minor effects on disc alignment, sizing and stacking while the disc rim formation is unaffected. Rom1 is found in non-covalent hetero-tetramers and in intermediate complexes but it is excluded from higher order Prph2 complexes. It seems plausible that the Prph2 higher order complexes, which are

una ffected in the Rom1 knockout retinas, mediate the membrane curvature and rim closure, while the intermediate hetero complexes support disc spacing, sizing and alignment. Additional studies targeting the formation of the di fferent complexes specifically and analyzing their function in rods and cones are needed to further pinpoint the roles of the di fferent Prph2 and Rom1complexes.

Progress in the development of a feasible gene therapy of Prph2 related diseases has been made in recent years. Transducing the photoreceptors of *Prph2*−/− and *Prph2*+/− mice either via AAVs or NPs showed a partial rescue of the knockout phenotype, even though the magnitude of the rescue and (in case of the AAVs) the persistence of the rescue e ffect is still low. The small magnitude of the rescue is most likely due to a low transduction rate of photoreceptors by the vectors used. Improving the transduction rate achieved by the vector as well as its distribution in the eye is one obstacle which needs to be overcome in future attempts. A second obstacle is that pathogenic *PRPH2* mutations do not result in a knockout scenario, but instead in a scenario where the mutant protein is still expressed. The knockin mouse models summarized in this review have proven to be a valid approach in understanding how the presence of the mutant protein a ffects both rod and cone photoreceptors, and also how stable the mutant protein is. In addition, these models helped to identify mutations, which result in dominant effects. Using a sh-RNA-mediated knockdown of both WT and mutant proteins followed by the expression of a sh-RNA resistant form of the WT protein provided a promising approach in the therapy of dominant gain-of-function mutations of Rho and Prph2. The fact that rods and cones utilize Prph2 di fferently could lead to further complications; this issue needs to be taken into consideration when developing e ffective therapeutic strategies for *PRPH2* related diseases. Testing these strategies in the murine models adds another layer of complication due to the low percentage of cones, which makes it di fficult to address cone-dominant phenotypes in mouse models. Analyzing *Prph2* mutations in the *Nrl*−/− background provides an approach for studying functional and structural e ffects of mutation specifically on cones. The lack of a macula in the murine retina continues to be the rate limiting factor which prevents conclusive studies aimed at developing therapies for mutations associated with macular defects and pattern dystrophy using mouse models. In addition to this, several *PRPH2* mutations were found to cause secondary e ffects in the choroid and RPE. The reasons for these secondary defects as well as their impact on the pathogenesis of *PRPH2* associated diseases are not well understood. Characterizing these secondary defects has to be achieved in future studies in order to devise a successful therapy.

**Author Contributions:** L.T. wrote original draft of the manuscript. M.S.M. performed electroretinography and histology. M.K. performed immunohistochemistry, developed the tables and participate in the writing. M.I.N., M.R.A.-U. designed the experiments, reviewed and edited the manuscript. All authors have read and agreed on the final version of the manuscript.

**Funding:** This research is supported by a gran<sup>t</sup> from the National Institutes of Health (R01 EY10609-MIN, MRA).

**Conflicts of Interest:** The authors declare no conflicts of interest.
