A) Similar expression of C1qtnf5 and Mfrp by reverse transcriptase-PCR in RPE/choroid from wild-type (wt) and heterozygous (wt/ki) or homozygous (ki/ki) C1qtnf5 Ser163Arg KI mice. B) Western blot of eye cups from wild-type and C1qtnf5 heterozygous and homozygous knock-in mice stained with anti-C1QTNF5 antibody (top) and anti-β tubulin antibody (bottom).

RPE damage was scored in wild-type and C1qtnf5 Ser163Arg knock-in (KI) mice at ages between 6 and 24 months old and no signficant difference was found between any of the genotypes (n (wt) = 6, n(het) = 15, n(mut) = 14). Overall, there is a signficant correlation of RPE damage with age (Pearson correlation n(wt,het,mut) = 35, p = 0.0026, r² = 0.244).

A, Dark-adapted (DA) ERGs evoked by 2.2 log scot-cd.s.m⁻² flashes (black traces) in 19-month-old heterozygote and homozygote, C1qtnf5 Ser163Arg knock-in mice compared to responses from a representative, wild-type (WT) control. DA cone function (bottom grey traces) was isolated in dark-adapted mice by recording ERGs within a short interval following a rod-suppressing bleaching light; recovery of the rod-mediated ERG response (grey traces overlapping DA rod traces) was assessed by recording ERGs 10 minutes following this bleach. B, Summary statistics of ERG parameters, showing a-wave (rods, left panel) and b-wave (cones, middle panel) in wild-type (WT), C1qtnf5 Ser163Arg heterozygous (Heter) and homozygous (Homoz) knock-in mice. Recovery of the a-wave (photoresponse) after the bleaching exposure is shown (right panel) and expressed as a fraction of the dark-adapted amplitude. The results for both eyes (right eye symbols are slightly displaced to the left) are plotted; circles are the 10–12 month old mice and hexagons are 15–18 month old mice. The grey line in each panel represents 2SD below the mean for WT mice.

(A) HeLa cells were infected with lentivirus contain wild-type C1QTNF5 (LNT.C1QTNF5-WT) or Ser163Arg mutant C1QTNF5 (LNT.C1QTNF5-Mut). At 72 hr post-infection, cells were stained with anti-CTRP5 (green) and the endoplasmic reticulum (ER) marker anti-BiP (red) antibodies. Cells infected with LNT.CTRP5 or LNT. CTRP5-Mut both showed intracellular staining surrounding the nucleus (a, d) suggesting that both proteins enter the secretory pathway. While the staining on LNT-CTRP5 infected cells is rather diffuse (a), the LNT-CTRP5-Mut showed punctate staining (d) that co-localised with the ER marker BiP (e, f, shown by white arrow heads) suggesting that the mutant form of the protein is retained in the cell and cannot exit the ER correctly. Controls showed that there was no contribution to the fluorescent signals from the respective anti-rabbit secondary antibodies (g, h, i, j, k, l). Scale bar = 20 µm. (B) Haematoxylin and eosin stained retinal sections from C57BL/6 mouse eyes at 5 weeks after subretinal injection of either LNT.C1QTNF5-WT, LNT.C1QTNF5-Mut, LNT.GFP or PBS. No obvious differences could be observed between the four groups despite some injection related trauma (not shown here). (C) RPE and outer nuclear layer (ONL) disease scoring was assessed on randomized images and showed no significant difference between the any of the four treated groups, n = 4 per group (One-way ANOVA, all p-values p>0.05). Scale bar = 100 µm.

A) Schematic representation of the murine Mfrp/C1qtnf5 genes. Boxes represent exons, the solid line represents intronic sequence (not drawn to scale). B) The targeting construct showing the long (6.8 kb) homology arm (LA), short (1.4 kb) homology arm (SA) and the central fragment with the Ser163Arg mutation labelled with a *. FRT: Flippase Recognition Target sites, Neo: the neomycin selection cassette. LoxP: sites flanking the introduced mutation and Neo gene, allowing subsequent Cre-recombinase-mediated deletion to generate a knockout mouse. C) Southern blot performed using genomic DNA from two heterozygous mice with a 3′ C1qtnf5 probe showing wild-type genomic DNA digested by AvrII, resulting in a 11.3-kb band, while genomic DNA containing the targeted Ser163Arg mutant showed the expected 6.6-kb band following Flp-mediated excision of the neo cassette. D) Validation of the Ser163Arg point mutation in heterozygous mice by DNA sequencing. The wild-type codon is AGC (serine), the mutant is AGG (arginine), the heterozygous mice show both alleles, highlighted in blue. E) Genotyping of tail biopsy DNA from wild-type and mutant (C1qtnf5 Ser163Arg) mice by PCR amplification of the native wild-type and mutant fragments. The primers anneal close to the FRT-sites flanking the neo cassette (see Figure 1 and Materials and Methods). The wild-type allele corresponds to the lower 432 bp band and the mutant allele to the upper 548 bp band. The gel shows genotypes in a mixture of C1qtnf5 Ser163Arg homozygous mutant (n = 9), heterozygous mutant (n = 4) and wild-type mice (n = 2). A DNA molecular weight marker V (8–587 bp; HaeIII digested pBR322 (Roche)) is shown in the right hand lane.

A) Retinal sections from mice of the C1qtnf5 wild-type, Ser163 Arg heterozygous (Het) and homozygous (Mut) genotypes at the ages indicated, stained and examined by light microscopy. B) Ultrastructures of the basal site of the RPE and Bruch's membrane of retinas from wild-type (Wt) and C1qtnf5 Ser163Arg KI (Het, Mut) mice at the ages of 6 months and 20 months, respectively. C) The thickness of Bruch's membrane increased with age (Pearson correlation Wt: n = 6, p = 0.103, r² = 0.526; Het: n = 22, p = 0.0005, r² = 0.459; Mut: n = 21, p = 0.0039, r² = 0.362) but no significant difference between wild-type and C1qtnf5 Ser163Arg KI mice was found.

A) Representative fundus images of inner and outer retina obtained by AF-SLO imaging from wild-type, heterozygous and homozygous mutant C1qtnf5 S163 KI mice at 16–18 months of age. No obvious difference in clearly demarcated autofluorescent areas or punctuated pattterns in the fundus images were observed in any of the three gentotypes. Wt: wild-type mice, Het: heterozygous KI mice, Mut: homozygous KI mice. B,C) Semiquantitative analysis of the inner (B) and outer (C) retinal background autofluorescence. The mean background autofluorescence (AF-pixel above threshold) for the right and left eye per animal is shown for each genotype (n = 3 animals per genotype). No significant difference in background autofluorescence in the inner retina (B, Kruskal-Wallis with Dunn's multiple comparison test, p = 0.393) or in the outer retina (C, Kruskal-Wallis with Dunn's multiple comparison test, p = 0.067) was observed between any of the three gentoytpes.

A, Fundus images from the early and late phase of in vivo fundus fluorescein angiography at 1 and 2 weeks after photocoagulation/laser induced choroidal neovascularization (CNV) in 15–16 month old wild-type and C1qtnf5 KI mice. Wt: wild-type mice, Het: heterozygous KI mice, Mut: homozygous KI mice. B, The microglia in laser lesions on RPE flat mounts resulting from CNV laser treated C1qtnf5 KI mice after 2 weeks were labelled with anti-Iba1 (green) and co-labelled with anti-lectin (blood vessels) antibodies. Scale bar, 250 µm. C,D, Quantitative analysis of area of hyperfluorescence in the fundus images from 1 week old animals (C) and 2 week old animals (D) as a measure for CNV lesion size. No significant difference between the hyperfluorescent areas were observed at 1 week (Kruskal-Wallis with Dunn's mutliple comparison test, p = 0.911) or 2 weeks (Kruskal-Wallis with Dunn's mutliple comparison test, p = 0.638). The number of animals for both time points were n (wt) = 4, n (het) = 12, n(mut) = 10.