Fig 2: Effects of PEDF/SERPINF1 and SFRP2 on dermal papilla cells (DPCs) and hair follicles (HFs). (a) The mRNA expression levels of PEDF/SERPINF1 and SFRP2 were significantly downregulated after corresponding small interfering (si)RNA transfection in cultured DPCs. (b) xCELLigence system detection showed that siPEDF/SERPINF1 or siSFRP2 treatment significantly increased the proliferation of DPCs compared with the negative control (NC) group DPCs. (c) Recombinant protein PEDF or SFRP2 treatment decreased the proliferation of DPCs. (d) Immunofluorescence staining assay identified decreased protein levels of PEDF and SFRP2 after siPEDF or siSFRP2 treatment of HFs. (e) PEDF/SERPINF1 knockdown increased the length of hair shafts, whereas PEDF recombinant protein significantly inhibited HF growth compared with the NC after 2 days of culture. (f) Treatment with siSFRP2 or recombinant SFRP2 protein showed no significant effect on HF growth. (g) Macroscopic images of cultured hair follicles in different groups were obtained every other day. (h, i) Quantification of hair cycle stage from macroscopic images of cultured hair follicles on day 6. A greater percentage of anagen HFs and a lower percentage of catagen HFs remained in the PEDF/SERPINF1 siRNA‐treated group, and PEDF recombinant protein facilitated catagen transition and shortened the anagen stage. By contrast, siSFRP2 and recombinant SFRP2 showed no significant effect on the HF hair cycle. Data are expressed as the mean (SD) of each group. P‐values were calculated using unpaired‐sample t‐tests. **P < 0·01, *P < 0·05. [Colour figure can be viewed at wileyonlinelibrary.com]

Fig 3: Validating Vegfa silencing and PEDF expression from the therapeutic cassette(A) Diagram of optimized Vegfa targeting miR-agshRNA construct combined with PEDF expression.(B) Bar plot (mean ± SD) shows the knockdown activity of the 13-12AB-PEDF construct depicted in A (blue bar) using the co-transfection of dual-luciferase reporter plasmid in HEK293 cells with the full-length murine Vegfa sequence. Rluc/Fluc ratios are normalized to pcDNA3-CMV-intron (n = 3). Statistical comparisons were performed using one-way ANOVA and Tukey’s multiple comparisons test.(C) Validation of PEDF expression in HEK293 cells transfected with 13-12AB-PEDF plasmid by immunostaining. Scale bars = 50 μm.(D) Western blot analysis for PEDF in transfected HEK293 cells in lysate (left-hand side) and media (right-hand side). M denotes a kilodalton size marker.

Fig 4: Subretinal delivery of AAV5-based dual Vegfa targeting miR-agshRNA unit with downstream protein expression(A) Diagram of the AAV-based vectors. In the upper panel, the transcribed region of S1-S2AB-AsRed or 13-12AB-AsRed is placed downstream of the RPE cell-specific promoter VMD2 and placed back-to-back with the PGK-eGFP-syn.pA (PES) reporter cassette. In the lower panel corresponding PEDF expression vector is shown (S1-S2AB-PEDF or 13-12AB-PEDF). Full vector names and abbreviations: pAAV-VMD2-S1-S2AB-AsRed+PES, S1-S2AB-AsRed; pAAV-VMD2-13-12AB-AsRed+PES, 13-12AB-AsRed; pAAV-VMD2-S1-S2AB-PEDF+PES, S1-S2AB-PEDF; pAAV-VMD2-13-12AB-PEDF+PES, 13-12AB-PEDF.(B) Representative examples of RPE/choroidal flat mounts collected on day 35 after the subretinal injection of 1 × 109 vg of AAV5 particles show the RPE expression of AsRed (red). GFP signal is presented in green. Scale bars = 20 μm.(C) Timeline for in vivo Vegfa knockdown efficacy study. Mice were subretinally injected at day 0 with 1 × 109 vg of AAV5/S1-S2AB-PEDF or AAV5/13-12AB-PEDF. FFI was performed at 18 dpi, where mice used for the quantification of Vegfa levels in crude RPE cells were sacrificed. At 28 dpi, mice used for the isolation of GFP+ RPE cells by FACS were sacrificed.(D) Bar plot (mean ± SD) shows Vegfa mRNA expression relative to Actb and normalized to the mean of the S1-S2AB-PEDF group. Each data point represents a pool of FACS-sorted GFP+ murine RPE cells.(E) Bar plot (mean ± SD) shows Vegfa mRNA expression relative to Actb and normalized to the mean of the S1-S2AB-PEDF group. Each data point represents RPE cells isolated from a whole eyecup.(F) Fundus fluorescence imaging of GFP from the eyes used in (E).(G and H) Quantification of area (in pixels) and the mean gray value in the eyes presented in (F) shows no difference in transduction between the two groups.(I) Timeline for the in vivo assessment of PEDF expression. At day 0, mice were subretinally injected with 1 × 109 vg of AAV5/S1-S2AB-AsRed (n = 3 mice, 6 eyes) or AAV5/S1-S2AB-PEDF (n = 4 mice, 8 eyes). FFI was performed at 29 dpi, whereafter mice were sacrificed.(J and K) Western blot analysis of PEDF levels in the murine retina/RPE/choroid tissue. Bar plot (mean ± SD) shows PEDF protein levels relative to vinculin and normalized to the mean of the S1-S2AB-AsRed group (set to 1). Each data point represents an eye. Statistical comparison in (D), (E), (G), (H), and (K) was performed using the Student’s t test, and in (K) with Welch’s correction. AAV, adeno-associated viral vector; dpi, days post injection; eGFP, enhanced green fluorescent protein; ITR, inverted terminal repeat; FACS; fluorescence-activated cell sorting; FFI, fundus fluorescence imaging; PEDF, pigment epithelium-derived factor; PES, PGK-eGFP-Syn-pA.; PGK, phosphoglycerate kinase 1 promotor; poly(A), polyadenylation signal; RPE, retinal pigment epithelium; Syn-pA, synthetic polyadenylation signal; vg, vector genomes; VMD2, vitelliform macular dystrophy 2 promoter.

Fig 5: Subretinal delivery of 13-12AB-PEDF reduces CNV formation(A) Timeline for CNV study. Mice were subretinally injected at day 0 with 1 × 109 vg of AAV5/S1-S2AB-AsRed, AAV5/13-12AB-AsRed, AAV5/S1-S2AB-PEDF, or AAV5/13-12AB-PEDF. Fundus fluorescence imaging (FFI) and subsequent laser-induced CNV were performed at 28 dpi. At 35 dpi, OCT of all lesions were performed followed by sacrifice of the mice.(B) Representative examples of CNVs on RPE/choroidal flat mounts from eyes injected with the four different AAV5 vectors. GFP expression (green), AsRed (red), CD31 (magenta), and GS-IB4 (cyan) is shown on the merge. Scale bars = 50 μm.(C and D) Quantification of (C) GS-IB4 and (D) CD31 CNV area from injected eyes. n = 11–12 mice per group. Violin plots of individual CNV measurements are presented. Crossbar represents the median. Statistical comparisons were performed using a linear mixed effects model with vector type as a fixed effect and eye nested under the animal factor as a random effect. AAV, adeno-associated viral vector; CD31, cluster of differentiation 31; CNV, choroidal neovascularization; dpi, days post injection; GS-IB4, Griffonia simplicifolia type I isolectin; OCT, optical coherence tomography; PEDF, pigment epithelium-derived factor; vg, vector genomes.