Fig 1.
Novel transgenic models permit conditional and targeted cone photoreceptor ablation.
Pairs of transgenic zebrafish were engineered so that when they are bred together their progeny express nitroreductase in either UV or Blue cone photoreceptors, allowing for conditional photoreceptor-specific ablation upon addition of a prodrug. A. UAS/KalTA4 system of cell specific ablation in either UV or blue cones begins with KalTA4 (an optimized variant of Gal4) being driven by sws1 or sws2 promoters, respectively, in different lines of transgenic zebrafish. When these are bred to an additional line of transgenic zebrafish engineered herein, the KalTA4 binds the transgenic UAS promoter to induce expression of the bacterial enzyme nitroreductase (NTR) fused to the fluorescent reporter mCherry. This latter product also encodes KalTA4 protein that breaks away from the NTR-mCherry fusion protein due to inclusion of a 2A peptide; this excess KalTA4 creates a feed-forward loop (“Kaloop”) that maintains its own expression, improving longevity and penetrance of the transgene expression. B. Nitroreductase (NTR) converts the prodrug metronidazole (MTZ, applied as a bath treatment in the tank water) into a cell autonomous DNA crosslinking agent, initiating apoptosis in the cone photoreceptors of interest. Thus ablation occurs conditionally, only in the cells expressing the transgene and only in the presence of the prodrug MTZ. C. Removal of the prodrug MTZ stops the cone ablation and allows retinal recovery to begin. Left: A small region of the normal adult zebrafish retina is represented, including schematic of the consistent reiterated pattern of cone photoreceptors (cylinders, colours of which represent the four cone spectral subtypes underpinning colour vision) thus forming a predictable mosaic, with second order neurons beneath; Middle: Prodrug MTZ leads to ablation of the specific cone photoreceptor subtype (in this example ablation of blue cones is schematized); Right: Ablation ceases upon removal of the prodrug from the tank water, which we’ve previously shown to induce regeneration of the injured retina from innate retinal stem cells. To be therapeutically useful, the regenerated cones must rewire to the remaining retinal network, including not only other cones but also horizontal cells and bipolar cells (orange & black, respectively). D. Timeline of experiment. Larvae were assessed in OMR (optomotor response), assessed for changes in pigmentation in response to UV light, and sampled for histology immediately following application of the prodrug MTZ (at 8 days post-fertilization) and on each day after. 1-phenol-2-thiourea (PTU) was applied to transiently block synthesis of melanin pigment and allow identification of transgenic vs. sibling (wild type) larvae.
Table 1.
Transgenic fish engineered to ablate specific cone subtypes.
Fig 2.
Blue cone photoreceptors are efficiently ablated upon addition of prodrug.
(A-C) Prior to treatment with prodrug metronidazole (MTZ) most blue cones (cyan) express the nfsb-mCherry transgene (magenta); some exceptions are noted (dotted lined circles). Wholemount retinae labelled with antisense riboprobes against sws2 blue opsin and against mCherry viewed en face (tangential plane). 78±2% of blue cones express the nfsb-mCherry transgene, considering all 1430±60 blue cones across the entire retina of 7dfp larvae (n = 4), the age at which they begin to receive MTZ in our ablation paradigm. D-K. Retinal sections of larval zebrafish expressing the transgenes Tg[sws2:KalTA4; UAS:nfsb-mCherry-KalTA4] such that blue cones contain the nitroreductase-mCherry fusion protein (pseudocoloured magenta). MTZ was applied via bath treatment for 24 hours beginning at 7 days post-fertilization (dpf); the end of this 24 hour period is denoted in the text and figures as zero hours since the end of drug treatment. Top row: Blue cones are intact in retinal sections of transgenic larvae treated with the vehicle control (DMSO, applied at 7 dpf; time 0h = 8dpf). Bottom row: Blue cones are ablated, as shown in retinal sections of transgenic larvae treated with the prodrug. A decrease in the number of mCherry-positive blue cones (magenta) is apparent through 48 hours post ablation, followed by observation of mCherry-fluorescing blue cones at the ciliary marginal zone (panels G & H). Auto-fluorescence is included for spatial reference (green). L. Blue cones were reduced in abundance four-fold after prodrug MTZ treatment (red bars) relative to vehicle control (DMSO; grey bars). Abundance of cones in the peripheral retina is represented by lighter grey or red bars. Blue cones were gradually added to the periphery (light red bars) on the days following MTZ. Quantification considered cones in the peripheral retina and mature retina (⌘ indicates p<0.0001 when comparing blue cone abundance in mature retina treated with MTZ relative to DMSO, and p<0.05 when comparing peripheral retina treated with MTZ relative to DMSO; ❖ denotes significant difference (p<0.0001) when comparing blue cones in mature retina with MTZ relative to DMSO, but peripheral blue cone abundance is not significantly different between treatments. n≥5 fish for each treatment). Scale bars are 50 μm.
Fig 3.
UV cones are efficiently ablated using our model of cell specific ablation.
Prior to treatment with prodrug metronidazole (MTZ), most UV cones (green, labelled with antibody 10C9.1 against UV opsin) express the nfsb-mCherry transgene (magenta). A-C. Wholemount retinae viewed en face (tangential plane). 81±4% of UV cones express the nfsb-mCherry transgene, considering all 1622±96 UV cones across the entire retina of 7 days post-fertilization (dpf) larvae (n = 5), the age at which they would begin to receive MTZ in our ablation paradigm. Scale bar in panel J is 20 μm. D-K. Larval zebrafish expressing the transgenes Tg[sws1:KalTA4; UAS:nfsb-mCherry-KalTA4] in the UV cones, such that UV cones contain the nitroreductase-mCherry fusion protein (pseudocoloured magenta). Top row: UV cones are intact in retinal sections of transgenic larvae treated with the vehicle control (DMSO, applied at 7 dpf). Bottom row: UV cones are ablated, as shown in retinal sections of transgenic larvae treated with the prodrug MTZ (applied via bath treatment for 24 hours beginning at 7 dpf; the end of this 24 hours is labeled as zero hours since the end of drug treatment). A decrease in the number of mCherry-positive UV cones (magenta) is apparent through 48 hours post ablation, followed by observation of mCherry-fluorescing blue cones at the ciliary marginal zone (panel H). Auto-fluorescence is included for spatial reference (Green). Scale bars are 50 μm. L. UV cones were reduced in abundance four-fold after prodrug MTZ treatment (red bars) relative to vehicle control (DMSO; grey bars). Abundance of cones in the peripheral retina is represented by lighter grey or red bars. UV cones were gradually added to the periphery (light red bars) on the days following MTZ. Quantification considered UV cones in the peripheral retina and mature retina (⌘ indicates p<0.0001 when comparing UV cone abundance in mature retina treated with MTZ relative to DMSO, and p<0.05 when comparing peripheral retina treated with MTZ relative to DMSO; ❖ denotes significant difference (p<0.0001) when comparing UV cones in mature retina with MTZ relative to DMSO, but peripheral UV cone abundance is not significantly different between treatments. n≥5 fish for each treatment).
Fig 4.
Ablating blue cones does not cause any detectable off-target cell death.
Transgenic zebrafish engineered for blue cone ablation were labeled for apoptosis using TUNEL to examine death of target photoreceptors (blue cones) and any effects on adjacent photoreceptors. (A, B) Zebrafish larvae that had received either vehicle control (DMSO) or prodrug metronidazole (MTZ) treatment, respectively, sacrificed at 9 days post fertilization (dpf). Blue cones expressing mCherry are pseudocoloured in magenta, TUNEL labelling is represented in green, nuclei are in cyan. Intact blue cones are not detectable in the prodrug treated larvae. (C) Quantification of the number of TUNEL-positive cells per section within the ONL. The number of TUNEL+ mCherry- cells did not differ significantly between treatment groups, indicating no detectable off-target cell death. The number of TUNEL+ mCherry+ cells differed significantly between groups, as was expected since the prodrug treatment is designed to induce apoptosis in the mCherry+ blue cones of this transgenic model. Scale bars are 100μm. *** = p<0.001. n = 12 & 11 fish for the DMSO & MTZ treatment groups, respectively.
Fig 5.
Blue cone ablation does not cause alterations in the abundance of non-target cone subtypes.
(A-H) Transgenic zebrafish retinae labelled with antibodies against UV and double cones. Top Row: Treated with vehicle control DMSO; Bottom Row: treated with prodrug metronidazole (MTZ) to induce ablation of blue cones. Retinas are visualized as whole mounts, wherein UV cones are labeled with antibody 10C9.1, double cones are labeled with antibody zpr1, and blue cones are filled with mCherry (pseudocoloured to cyan) fused to nitroreductase. Note that in (G) blue cones are absent due to application of prodrug MTZ. (D, H) Merged image showing the larval mosaic. Scale bar is 50 μm. (I) The numbers of UV and double cone photoreceptors obtained from DMSO and MTZ treated zebrafish larvae are indistinguishable, i.e. ablating blue cones does not detectably disrupt adjacent cones. n = 12 fish for both treatments.
Fig 6.
Ablating UV cones does not cause any detectable off-target cell death.
Transgenic zebrafish engineered for UV cone ablation were labeled for apoptosis using TUNEL to examine death of target photoreceptors (UV cones) and any effects on adjacent photoreceptors. (A, B) Zebrafish larvae that had received either vehicle control (DMSO) or prodrug metronidazole (MTZ) treatment, respectively, sacrificed at 9 days post fertilization (dpf). UV cones expressing mCherry are pseudocoloured in magenta, TUNEL labelling is represented in green, nuclei are in cyan. Intact UV cones are not detectable in the prodrug treated larva. (C) Quantification of the number of TUNEL-positive cells per section within the ONL. The number of TUNEL+ mCherry- cells did not differ significantly between treatment groups, indicating no detectable off-target cell death. The number of TUNEL+ mCherry+ cells differed significantly between groups, as was expected since the prodrug treatment is designed to induce apoptosis in the mCherry+ UV cones of this transgenic model. * = p<0.05. n = 5 & 9 fish for the DMSO & MTZ treatment groups, respectively. Scale bar is 100μm.
Fig 7.
UV cone ablation does not cause alterations in the abundance of non-target cone subtypes.
(A-F) Transgenic zebrafish retina labelled with antibodies against double cones. Top Row: Treated with vehicle control DMSO; Bottom Row: treated with prodrug metronidazole (MTZ) to induce ablation of UV cones. Retinas are visualized as whole mounts, wherein double cones are labeled with antibody zpr1, and UV cones are filled with mCherry (pseudocoloured to cyan) fused to nitroreductase. Note that in (E) UV cones are greatly reduced in abundance due to application of prodrug MTZ. (C, F) Merged image showing the larval mosaic. Scale bar is 50 μm. (I) The numbers of UV and double cone photoreceptors obtained from DMSO and MTZ treated zebrafish larvae are indistinguishable, i.e. ablating blue cones does not detectably disrupt adjacent cones. n = 6 or 5 fish for DMSO and MTZ treatments, respectively.
Fig 8.
Validation of OMR stimulus as a visually mediated behaviour.
The response of zebrafish larvae to our OMR stimuli comprised of blue and dark red bars was used as a measure of visual function. We validated this conclusion in several ways. A. 24 hours following drug application on transgenic fish (= blue cone ablation, red bars), responses of zebrafish larvae were observed relative to control fish (grey bars) with various stimuli. Stimuli are detailed in S1 Fig, and consisted of typical OMR stimuli of black & white bars, or alternatively blue bars with intervening red bars of various saturation. All stimuli presented were able to distinguish the visually mediated behaviour of larval zebrafish with blue cones ablated relative to controls. The blue and dark red stimuli pair evoked responses that were least variable between individuals and showed the largest magnitude of change to both control treatments, and thus was selected as the stimulus for the majority of our work. B. Larvae acutely blinded by UV light, presumably bleaching their photopigments, responded less than their unblinded siblings (p<0.05). C. Larvae showed photoreceptor degeneration one day following blinding by intense UV light (S2 Fig) and responded less than their unblinded siblings (p<0.05). D. Blind fish have reduced responses to our OMR stimulus. gdf6a-/- mutants, which have previously been shown to exhibit reduced OMR response with typical stimuli, presumably due to overt microphthalmia and cone photoreceptor degeneration, responded less to our OMR stimulus relative to their normophthalmic sighted gdf6a+/- siblings (p<0.05). E. Larvae responding to our OMR stimuli moved through the arena more when stimuli were presented in a typical fashion (described in S1 Fig) compared to when stimuli were presented moving in the opposite direction (p<0.05 by t-test). The latter was predicted to stimulate larvae to remain in their initial position, thus reducing the distance between their initial and final positions, and this prediction was met. F. Alternative stimuli were applied to test for any potential movement deficits induced by combining our transgene and MTZ prodrug. A tapping stimulus delivered to the same computer monitor (that was otherwise used to deliver OMR stimuli) induced equivalent amounts of movement in Tg[sws2:nfsb-mCherry] larvae treated with MTZ (= “Blue MTZ” in later figures) as in their siblings treated with DMSO (p = 0.32). G. As per panel F, but potential movement deficits were assessed by Touch Evoked Escape Response (TEER). Transgenic larvae with blue cones ablated did not move less than larvae treated with vehicle DMSO (p = 0.34). H. Spontaneous Swim Assay records movement of 10 fish per trial for ten minutes (see S3 Fig). Transgenic Tg[sws2:nfsb-mCherry] larvae treated with MTZ did not move less compared to when treated with DMSO or compared to wild type treated with MTZ. Statistical comparisons were made via t-test. Sample sizes (n = number of larvae individually tested, except in panel H sample sizes = number of groups of ten larvae tested) are indicated.
Fig 9.
Visually mediated behaviour is reduced following ablation of UV or blue cones but recovers rapidly following blue cone ablation.
Thus the recovery of visually mediated behaviour has a time course that depends on the type of cone photoreceptor ablated. Ablation of UV or Blue cones leads to the immediate impairment of visual function (panels A & B, respectively; the first triplet of histograms in each shows behaviour severely decreased following cone ablation (red bars) compared to controls (grey bars)). Visually mediated behaviour regenerated following UV cone ablation over the course of several days (A). Unexpectedly, vision is rapidly restored following ablation of blue cones (B; e.g. red bar at 24 hours since the end of drug treatment). We designed and empirically optimized a visually evoked behaviour based on the optomotor response (OMR; Figs 8A and S1) that stimulates a quantifiable directional movement of freely swimming larval zebrafish. “WT MTZ” are wild type zebrafish receiving the prodrug metronidazole (MTZ) as a control treatment, presented in dark grey bars; “UV DMSO” or “Blue DMSO” in light grey bars are a second control, representing transgenic fish that express nitroreductase for ablation of UV or blue cones, respectively, and these fish received vehicle control (DMSO) only; “UV MTZ” or “Blue MTZ” in red bars are transgenic fish treated with prodrug MTZ and thus had their UV or Blue cones ablated, respectively, immediately prior to testing of visual ability at time zero. Sample sizes (n = number of larvae tested) are reported at the bottom of each bar, and are detailed in Table 2. Table 2 also reports the data prior to the normalization used to generate this Figure. ***p<0.001 experimental relative to controls, *p<0.05 experimental compared to the wild-type control. The visually mediated OMR behaviour had significantly recovered to be indistinguishable from controls 72 hours after UV cone ablation (A), and 24 hours following blue cone ablation (B).
Table 2.
Visually mediated behaviour is lost following ablation of UV or Blue cones (top & bottom half of Table, respectively) but recovers rapidly following blue cone ablation.
This data, after normalization, is also plotted in Fig 9.
Fig 10.
Rapid functional recovery of visually evoked behavioural response is not dependent on blue cone generation or regeneration.
Blue cones were ablated in larvae, which were treated with the prodrug metronidazole (MTZ) for 48 hours beginning at 7 dpf, then behaviourally tested immediately following the removal of the prodrug. This treatment was intended to kill any blue cones that were potentially being generated during the 24–48 hours-post-ablation window, which was when behavioural recovery unexpectedly occurred in Fig 9. Dotted line & shading represent the mean ± SE of OMR response at 24 hours post ablation, i.e. is a benchmark of reduced visually mediated behaviour immediately following cone ablation derived from the red shaded bottom left histograms in Fig 9B. Despite this extended treatment of prodrug MTZ that prevented any putative regeneration/addition of blue cones, visually-mediated behaviour fully and rapidly recovered (compare red bar to grey controls, no significant difference) Figure legend as per Fig 9, such that red bars represent fish with blue cones ablated and grey bars represent control treatments. “WT MTZ” are wild type zebrafish receiving the prodrug metronidazole (MTZ) as a control treatment, presented in dark grey bars; “Blue DMSO” in light grey bars are a second control, representing transgenic fish that express nitroreductase for ablation of blue cones that received vehicle control (DMSO) only; “Blue MTZ” in red bars are transgenic fish treated with prodrug MTZ and thus had their Blue cones ablated. n = 9–10 fish per bar for the 48 hour treatment, as detailed in S2 Table that also reports the data prior to the normalization used to generate this Figure. The data are from fish comparable to those in the second set of histograms in Fig 9B, wherein visually-mediated behaviour had returned to control levels 24 hours after cone ablation; here, MTZ was kept on the larvae for an additional 24h (48h total) so that the rapid recovery of vision could be assessed while the addition/regeneration of blue cones was blocked. The recovery of vision to control levels in these conditions demonstrates that addition of blue cones (e.g. via regeneration or proliferation at the CMZ) is not required for the rapid recovery of vision reported in Fig 9B.
Fig 11.
Photoreception of UV light is reduced following ablation of UV cones as measured by changes in body pigmentation.
UV cones were ablated in larvae, and larval body pigmentation was measured in response to UV light; zebrafish larvae of this age are known to darken in response to UV light via melanin dispersion. A. Dorsal view of larval zebrafish. Binarization of images was used to quantify melanosome reaction to UV light. Larvae were imaged at multiple time points, beginning with dark-adapted larvae. B. The number of pixels in dark-adapted larvae was given a melanin index of 1. C. UV light exposure causes wild type fish to darken, increasing the melanin index. D. Shortly after drug treatment (0 hours), control larvae respond to UV light with the expected darkening of body pigmentation (grey lines), whereas larvae with UV cones ablated exhibit a significantly delayed response (red line). E. Larvae tested 24h after drug treatment reveal that UV cone ablation significantly impairs photoreception. F,G. Larvae tested 48 and 72 hours after drug treatment reveal a gradual recovery of sensitivity to UV light to levels indistinguishable from controls. ** = p<0.01 compared to UV+DMSO group *** = p<0.001 comparing UV+MTZ group to larvae in all control groups in the same timepoint, and these data are also significantly different (p<0.01) from the melanin index of the same UV+MTZ larvae in the dark-adapted state. Statistical significance was determined by Two-way ANOVA with a post-hoc Tukey test. The standard error surrounding each mean is reported in Table 3.
Table 3.
Larval zebrafish body pigmentation changes during exposure to UV light and is modulated by UV cone ablation 1.