Donor and host photoreceptors engage in material transfer following transplantation of post-mitotic photoreceptor precursors

Abstract

Photoreceptor replacement by transplantation is proposed as a treatment for blindness. Transplantation of healthy photoreceptor precursor cells into diseased murine eyes leads to the presence of functional photoreceptors within host retinae that express an array of donor-specific proteins. The resulting improvement in visual function was understood to be due to donor cells integrating within host retinae. Here, however, we show that while integration occurs the majority of donor-reporter-labelled cells in the host arises as a result of material transfer between donor and host photoreceptors. Material transfer does not involve permanent donor-host nuclear or cell-cell fusion, or the uptake of free protein or nucleic acid from the extracellular environment. Instead, RNA and/or protein are exchanged between donor and host cells in vivo. These data require a re-evaluation of the mechanisms underlying rescue by photoreceptor transplantation and raise the possibility of material transfer as a strategy for the treatment of retinal disorders.

Figures

Figure 1. Real-time imaging of transplanted donor precursor cells migrating into host retinae.

NrlGFP post-mitotic photoreceptor precursor donor cells (green) were transplanted into 12 weeks old Prph2rd2/rd2 mice and explanted retinae were examined by real-time 2-photon fluorescence live imaging 3 days post transplantation (retinae, including donor and host cells, were acutely labelled with Mitotracker Orange CMTMRos (red) before the recording). (a), At the start of the time-lapse recordings, sub-retinally transplanted rod precursor cells were frequently observed at the level of the IPM or apical to the retinal tissue. A dashed line (white) was inserted at the presumptive level of the OLM. (b), Time course of rod precursor cell movement from the IPM into the ONL; penetration through the OLM (white dashed line) is indicated by white arrowheads (see also Supplementary Movie 1). (c,d), High magnification views of time frames shown in (b) depicting migration of the right (c) and left (d) rod precursor cell into the ONL at selected time points. Note that while GFP fluorescence gradually reduced over the imaging period, Mitotracker Orange CMTMRos labelling, which is present in both donor and host cells, persisted in its absence. Scale bars, 10 μm.

Figure 2. Host and donor photoreceptor exchange fluorescent reporter proteins as assessed by confocal imaging.

NrlGFP post-mitotic photoreceptor precursor donor cells (green) were transplanted into DsRed (red) hosts and examined by confocal microscopy 5-6 weeks post transplantation. Nuclei are labelled with Hoechst (blue). (a, b), histograms showing quantification of the proportion of cells analysed (n=157 cells; N=5 retinae) that expressed GFP alone (GFP+) or GFP and DsRed (GFP+/DsRed+) when measuring inner segments (a) or cell bodies (b). (c), confocal projection image of a representative host retina. Lines represent regions of interest (ROIs) through cells and inner segments shown in (d–g). (d), (e) single confocal sections and the respective line plots through two inner segments and (f,g), two cell bodies shown in (c). (h,i), scatter plot (mean±s.d.) and histogram showing the position of GFP+ and GFP+/DsRed+ photoreceptors with respect to the outer limiting membrane (OLM) of host retina. The relative positions of the two populations were significantly different (P=0.0486 two tailed t-test with Welch's correction; P<0.001 2-way ANOVA). (j), confocal projection image showing donor NrlGFP+ cells in SRS of DsRed+ host. Arrows, donor GFP+ cells that are DsRed+. Scale bars 5 μm (c), 3 μm (d) and 10 μm (g).

Figure 3. Host and donor photoreceptors exchange fluorescent reporter proteins as determined by flow cytometry.

NrlGFP post-mitotic photoreceptor precursor donor cells were transplanted into DsRed hosts and examined by flow cytometry 5-6 weeks post transplantation. (a), box (25–75% percentile) and whiskers (min/max) plot showing median (line) % of GFP+ only and GFP+/DsRed+ photoreceptors within each host retina (N=18 retinae). ***P<0.001, paired t-test. (b–d), representative flow cytometry plots for adult (b) wild-type (negative control), (c) DsRed (positive control) and (d) NrlGFP (positive control) retinae. Pink box shows gating for GFP+ cells. (e,f), representative plots from an example of a host retina showing (e) % of total retinal cells that were GFP+ (pink box) and (f) the proportion of these that were GFP+ only (left pink box) or GFP+/DsRed+ (right pink box). (g), plot showing the proportion of CD45+ cells within the GFP+ population shown in (e,f).

Figure 4. Material transfer is not the result of nuclear fusion.

(a–c), representative confocal images of retinal sections from Gnat1−/− and wild-type retinae following transplantation of NrlGFP post-mitotic photoreceptor precursors (green), showing expression of Lamin B (red). GFP+ cells (arrows) within the host ONL only ever presented a single nuclear envelope (red) and nucleus (labelled with Hoechst; blue). Right hand panels show nuclei and Lamin B staining only. Scale bars 10 μm (d–g), FISH for the male Y-chromosome. (d), Male and (e), female NrlGFP (green) retinal sections stained for the Y-chromosome (red), which appears as a red dot. Insert, Male GFP+ cells with Y-chromosome staining. No Y-chromosome staining was seen in the female sections (e). (f,g), examples of GFP+ cells within female host retinae 5-6 weeks post-transplantation of male NrlGFP donor cells, showing a GFP+ cell that was positive for Y-chromosome staining (f, insert) and another that was GFP+/Y-chromosome− (g, insert). Right hand panels show GFP and Y-chomosome staining only. Scale bars 25 μm

Figure 5. Material transfer permits robust exchange of an array of gene products including those genetically absent from host photoreceptors.

NrlGFP post-mitotic photoreceptor precursor donor cells were transplanted into adult Gnat1−/− hosts and immunostained for rod α-transducin 5-6 weeks post-transplantation (a–c). Confocal projection images of GFP+ cells (green) in the ONL of host retinae stained for rod α-transducin (red). Nuclei were stained with Hoechst 33342 (blue). Note that in (b) the right hand panel shows the same image but with increased gain for the blue channel to show the nuclei of the RPE: No donor cells were present in the SRS overlying the GFP+/rod α-transducin+ cells. (d), Examples of GFP−/rod α-transducin+ host cells (arrow) were also seen. (e), Histogram showing the mean percentage (±s.d.) of GFP+ photoreceptors within the host ONL that were rod α-transducin+ (n=138 cells; N=4 retinae). f, rod α-transducin staining was seen in the outer segments of both GFP- and DsRed-reporter-labelled photoreceptors (arrow). Scale bars 10 μm.

Figure 6. Material transfer is specific to photoreceptor precursors and does not result from uptake of free proteins from environment.

(a–d), confocal projection images of adult retinal sections taken at 48 h, 1 week, 2weeks and 6 weeks post-injection of rEGFP and stained with anti-GFP antibody (green). Insert, Very few GFP+ cells were seen; those that were had normal rod photoreceptor morphology. (e,f) Scatter plots showing the mean (±s.d.) number of GFP+ cells within the host ONL at each time point in stained and unstained serial sections. (g), confocal projection images of retinal sections taken at 6 weeks post-injection of GFP+ fibroblasts. (h), scatter plot showing the number of GFP+ cells within the host ONL. (i), confocal projection images of retinal section taken at 6 weeks post-injection of E11.5 GFP+ retinal progenitor cells (RPCs). (j), scatter plot showing the number of GFP+ cells within the host ONL. For panels (a)-(d) and (g),(h), images from top to bottom show individual and/or combined channels for the same region of interest. Scale bars 50 μm.

Figure 7. Material transfer involves interaction between donor and host photoreceptors

NrlGFP+/CD73+ (green) and DsRed+/CD73+ (red) donor cells were transplanted into adult wild-type and Gnat1−/− recipients. (a–c), confocal projection images showing heterogeneous mixes of DsRed+ (closed arrow) and GFP+ (open arrow) cells within the host ONL. Occasionally, a photoreceptor within the host ONL bore both GFP and DsRed labels (asterisk). Nuclei are labelled with Hoechst (blue). (d,e), confocal projection images showing a similar finding for donor cells in the SRS. The majority expressed either GFP (open arrow) or DsRed (closed arrow), but some expressed both reporters (asterisk). For all panels, images from left to right show individual and/or combined channels for the same region of interest. Scale bars 10 μm.

Figure 8. Material transfer does not require sustained physical interaction between donor and host cells.

(a), confocal image of GFP+ cells within the host ONL immediately below the donor cell mass. Labelled user segments terminate close to the cell mass. (b), confocal image of GFP+ cells within the host ONL a significant distance away from the cell mass, with no donor cells present in the overlying SRS. (c,d), 3D reconstructions of GFP+ labelled cells in the ONL of (c) wild-type and (d) Gnat1−/− host retinae show no physical interaction with any other GFP+ cell. Scale bars 10 μm.

Figure 9. Transplantation of Cre+ photoreceptor precursors leads to widespread expression of floxed reporters in host cells.

(a), box (25–75% percentile) and whiskers (min/max) plot showing median (line) number of reporter-labelled photoreceptors in adult dTomatofloxed host retinae following transplantation. **P<0.01 2-way ANOVA. N.B. all controls (uninjected, PBS-injected and final wash) were highly significantly different from both Cre+/GFP+ and TdTom+ populations but are not denoted on graph for clarity). (b–d), confocal projection images showing retinal sections after sub-retinal injection of (b) ShH10.CMV.iCre (5 × 1012 viral particles per ml−1), (c) uninjected, (d) PBS-injected, or (e) mice injected with final wash from cell preparation. (f), confocal projection images of GFP+ (green) and tdTomato+ (red) cells within the floxed tdTomato host retinae following transplantation of Crx.GFP+/Cre+ donor cells. Note that there are significantly higher numbers of tdTomato+ cells than GFP+ cells. Inset, example of a GFP+, tdTomato+ cells, surrounded by GFP−/tdTomato+ cells. (g), Immunostaining for CRE (grey) shows robust protein expression in donor cells in SRS and some weaker expression in reporter-labelled cells within host retina. (h), some cells in the SRS cell mass were tdTomato+ (arrow), indicating the potential for bidirectional material transfer. For panels (f) and (g), images from top to bottom show individual and/or combined channels for the same region of interest. (i), tdTomato+ cells were present in areas where donor cell masses were present but no GFP+ cells within the underlying host ONL. (j), few, if any, GFP+ or tdTomato+ cells were seen in retinae where donor cell mass had been rejected (N=2 retinae). Scale bars 10 μm.