Somatic ablation of the Lrat gene in the mouse retinal pigment epithelium drastically reduces its retinoid storage.

PURPOSE To generate a mouse model in which the Lrat gene is selectively disrupted in the retinal pigment epithelium (RPE). To evaluate the effects on the synthesis of retinyl esters and on the expression of other proteins involved in the continuation of the visual cycle. METHODS A mouse line in which part of the first exon of the Lrat gene has been flanked by loxP sites, was generated and used in the study (Lrat(L3/L3) mice). Heterozygous mice (Lrat(+/L3)) were crossed with mice expressing Cre-recombinase under control of the tyrosinase-related protein-1 (Tyrp1) promoter, which is active selectively in melanin-synthesizing cells such as RPE cells. Accordingly, mice obtained from these crosses should display an RPE-specific disruption of the Lrat gene (Lrat(rpe-/-)). In addition, by crossing CMV-Cre transgenic mice with Lrat(L3/L3) animals, a germline null Lrat knockout (Lrat(L-/L-) mice) was generated. RNA and protein expression, endogenous retinoid levels, and electroretinogram (ERG) analyses were performed on Lrat(rpe-/-) and Lrat(L-)/(L-) mice, to determine the effects of Lrat disruption. Retinoid levels in nonocular tissues were also analyzed for comparison. RESULTS Analysis of RPE tissues from Lrat(rpe-/-) mice showed absence of Lrat message, lack of Lrat protein expression and consequently a reduced light response in ERG recordings. In addition, RPE cells from Lrat(rpe-/-) showed a strong reduction in their ability to synthesize all-trans retinyl esters, whereas Lrat activity in other tissues known to process retinol was comparable to control Lrat(L3/L3) animals. The Lrat(L-/L-) mice showed no detectable Lrat message, lack of protein expression, and barely detectable ester formation in RPE cells or several other relevant tissues analyzed. CONCLUSIONS Three Lrat mouse lines with genetic modifications were generated. The Lrat(L-)/(L-) mice displayed features similar to equivalent models previously reported by others. The second mouse line (Lrat(rpe-/-)) displayed loss of Lrat function only in the RPE. The third line possesses functional Lrat in all tissues, but part of the Lrat coding gene was flanked by loxP sites (Lrat(L3/L3)). This feature allows the disruption of this gene in any tissue of choice, by intercrossing with mice in which Cre-recombinase expression is driven by an appropriate tissue-specific promoter.