Molecular Vision 2011; 17:2956-2969 <http://www.molvis.org/molvis/v17/a319>
Received 18 December 2010 | Accepted 11 November 2011 | Published 16 November 2011

This Article

Download PDF

Citation (for Endnote)

Google Scholar

Articles by the Authors

Novel strategy for subretinal delivery in Xenopus

Federico Gonzalez-Fernandez,1,2,3,5 Cheryl A. Dann,4,5 Mary Alice Garlipp1,3,5

1Ross Eye Institute and Department of Ophthalmology, State University of New York, SUNY Eye Institute, Buffalo, NY; 2Pathology & Anatomic Sciences; State University of New York, Buffalo, NY; 3Graduate Program in Neurosciences; State University of New York, Buffalo, NY; 4Veterinary Medical Unit, Medical Research Service, Veterans Affairs Medical Center, Buffalo, New York, Buffalo, NY; 5Medical Research Service, and Departments of Ophthalmology and Pathology, Veterans Affairs Medical Center, Buffalo, NY

Correspondence to: Federico Gonzalez-Fernandez, Medical Research Service, Veterans Affairs Medical Center, Building 20, Room 219, 3495 Bailey Avenue, Buffalo, NY, 14215; Phone: (716) 863-2291; FAX: (716) 862-6526; email: fg23@buffalo.edu

Abstract

Purpose: The subretinal space, which borders the retinal pigment epithelium (RPE), photoreceptors, and Müller cells, is an ideal location to deliver genetic vectors, morpholino oligos, and nanopharmaceuticals. Unfortunately, materials injected into the space tend to stay localized, and degenerative changes secondary to retinal detachment limit its usefulness. Furthermore, such injection requires penetration of the sclera, RPE/choroid, or the retina itself. Here, we developed a strategy in Xenopus to utilize the continuity of the brain ventricle and optic vesicle lumen during embryogenesis as a means to access the subretinal space.

Methods: Wild-type and transgenic embryos expressing green fluorescent protein under the rod-opsin promoter were used for optic vesicle and brain ventricle injections. For injection directly into the optic vesicle, embryos were laid on one side in clay troughs. For brain ventricle injections, embryos were placed standing in foxholes cored from agarose dishes. Linear arrays with each embryo positioned dorsal side toward the micromanipulator facilitated high throughput injections. Twenty-five micrometer micropipettes, which were positioned with a micromanipulator or by hand, were used to pressure inject ~1.0 nl of test solution (brilliant blue, India ink, fluorescein isothiocyanate dextran, or 0.04 µm of latex polystyrene microspheres [FluoSpheres®]). FluroSpheres® were particularly useful in confirming successful injections in living embryos. Anesthetized embryos and tadpoles were fixed in 4% paraformaldehyde and cryoprotected for frozen sections, or dehydrated in ethanol and embedded in methacrylate resin compatible with the microspheres.

Results: Direct optic vesicle injections resulted in filling of the brain ventricle, contralateral optic vesicle, and central canal. Stages 24 and 25 gave the most consistent results. However, even with experience, the success rate was only ~25%. Targeting the vesicle was even more difficult beyond stage 26 due to the flattening of the lumen. In contrast, brain ventricle injections were easier to perform and had a ~90% success rate. The most consistent results were obtained in targeting the diencephalic ventricle, which is located along the midline, and protrudes anteriorly just under the frontal ectoderm and prosencephalon. An anterior midline approach conveniently accessed the ventricle without disturbing the optic vesicles. Beyond stage 30, optic vesicle filling did not occur, presumably due to closure of the connection between the ventricular system and the optic vesicles. Securing the embryos in an upright position in the agarose foxholes allowed convenient access to the frontal cephalic region. On methacrylate sections, the RPE-neural retina interphase was intact and labeled with the microspheres. As development continued, no distortion or malformation of the orbital structures was detected. In green fluorescent protein (GFP), transgenic embryos allowed to develop to stage 41, retinal FluoSpheres® labeling and photoreceptor GFP expression could be observed through the pupil. On cryosections, it was found that the FluoSpheres® extended from the diencephalon along the embryonic optic nerve to the ventral subretinal area. GFP expression was restricted to rod photoreceptors. The microspheres were restricted to the subretinal region, except focally at the lip of the optic cup, where they were present within the retina; this was presumably due to incomplete formation of the peripheral zonulae adherens. Embryos showed normal anatomic relationships, and formation of eye and lens appeared to take place normally with lamination of the retina into its ganglion cell and the inner and outer nuclear layers.

Conclusions: Diencephalic ventricular injection before stage 31 provides an efficient strategy to introduce molecules into the embryonic Xenopus subretinal space with minimal to the developing eye or retina.

Introduction

The vertebrate eye arises through a series of reciprocal inductive interactions between the neuroepithelium, surface ectoderm, and extraocular mesenchyme. Central to this choreography is the formation of the optic cup through the invagination of the optic vesicle. As the vesicle induces lens formation in the overlying competent surface ectoderm, its inner layer in turn gives rise to the neural retina, while the outer layer becomes the retinal pigment epithelium (RPE). As this is taking place, the optic stalk narrows, eventually separating the central nervous system (CNS) ventricles and subretinal space into unique compartments [1,2]. With the elongation of the outer segments, the interphotoreceptor matrix (IPM) accumulates within the expanding subretinal space. RPE zonula occludens prevent diffusion of matrix components sclera; while the zonula adherens prevents molecules with a Stokes’ radius >30 Å from leaving the subretinal space vitread between adjacent photoreceptor and Müller cells [3]. The matrix is thought to mediate RPE/retina interactions during development, including adhesion, sequestration of growth factors, and facilitating the exchange of retinoids between the RPE and neural retina in the visual cycle [4-8].

Because it borders the RPE, photoreceptors, and Müller cells, the subretinal space is an ideal location for delivering molecules to the outer retina. Subretinal injection can be easily performed in rodents [9-11], and has been useful for the introduction of viral vectors [12-14] and growth factors into the retina [15-18]. Many of these studies have been extended into clinical trials, further establishing the usefulness of the subretinal space in gene delivery [13,19-21].

Although the IPM is an ideal location to deliver molecules to the retina, several factors limit the usefulness of subretinal injections. In theory, material injected into the subretinal space should have access to the entire matrix, and therefore all of the RPE, photoreceptors, and Müller cells. However, injected material tends to stay localized at the injection site [9,10]. Furthermore, pathological changes secondary to the injection itself are not uncommon, as the method requires penetration of the sclera and RPE/choroid, or the retina itself. Finally, retinal detachment, an unavoidable consequence of subretinal injection, can alter ocular gene expression, and result in retinal and RPE degeneration [22,23].

A method allowing the less traumatic introduction of molecules into the subretinal space would be very useful for basic and translational research. In particular, Xenopus, a favorite of the experimental embryologist, develops exogenously, can be manipulated surgically, and is amendable for transgenic strategies. However, the inability to deliver test compounds to the retina and subretinal space remains a major obstacle. For example, it would be useful to be able to deliver DNA constructs, morpholino oligos, and morphogens into the subretinal space to study photoreceptor gene expression and function. Currently, morpholinos are commonly introduced into the embryo by blastula injection. However, when injected in this way, the morpholino is diluted out during organogenesis, limiting its ability to inhibit mRNA translation at late stages such as those needed to study photoreceptor development and function.

To circumvent a direct injection approach, we wanted to develop a less invasive strategy to deliver molecules to the subretinal space without disruption of the retina. Such a method should not disturb the retina anatomically. We considered that the normal developmental sequence of the CNS and ocular structures could provide an opportunity to accomplish this goal. Figure 1 illustrates the lumen of the optic vesicle is continuous with that of the brain ventricle early in development [24]. We anticipated that material introduced into the optic vesicle or brain ventricle would become sequestered within the subretinal space as the eye forms. Our studies suggest that optic vesicle or brain ventricle microinjection provides a nondisruptive strategy to deliver molecules into the subretinal space.

Methods

This research was performed under protocols approved by the Veterans Affairs Medical Center Institutional Animal Care and Use Committee. Adult pigmented Xenopus laevis were obtained from Xenopus 1 (Dexter, MI) or NASCO (Fort Atkinson, WI). Animals were maintained under a 12 h:12 h light-dark cycle in an XR1 aquarium system (Aquatic Habitats, Apopka, FL). Lines of transgenic animals expressing GFP under the rod-opsin promoter [25] were used in some experiments to identify the developing photoreceptors. These animals were generated by restriction enzyme mediated integration [26,27], using linearized plasmid pXO-GFP generously provided by Dr. Barry Knox (SUNY, Syracuse, NY) [28]. Methods for embryo care and Xenopus husbandry were followed as previously described [29]. Embryos were staged according to Nieuwkoop and Faber [30]. The following describes embryo handling, the subretinal delivery method, and useful photography and histological notes.

Embryo preparation

Embryos for injection can be obtained either by natural or in vitro fertilization. Although the latter has the advantage of generating embryos at nearly identical developmental stages, natural fertilization is often convenient, especially when different stages are desired. The preference therefore largely lies in what is routinely available in the laboratory. For either method, it is helpful to maintain groups of developing embryos at various temperatures ranging from 14 to 22 °C. This lengthens the time window when the desired stages are available. Khokha et al. [31] provides charts of the effect of temperature on Xenopus laevis and Xenopus tropicalis development (see also Xenopus development).

Embryos are secured for microinjections in modeling clay or agarose dishes, and for most routine work, visualized under Stemi 2000 stereomicroscopes equipped with zoom magnification (Carl Zeiss, Germany). Although embryos at the developmental stages used for injection are not free swimming, spontaneous movements can interfere with securing the animals and performing injections. Benzocaine (Sigma; St Louis, MO), at a final concentration of 0.005 to 0.01%, is effective in eliminating such movements (see comments under troubleshooting guide, Table 1). For injections directly into the optic vesicle, the embryo is laid on its side in a trough formed in modeling clay. Having tried several types of modeling clay, we found that the blue nonhardening/nontoxic type such as that from Sargent Art (Hazleton, PA) has the right malleability without any apparent toxicity to the embryos. The embryo is positioned in the clay as described below, and a flap of clay gently brought over to secure it in place.

For brain ventricle injection, we found it was best to orient the embryo in a vertical “standing position” in 1% agarose (BioRad; Hercules, CA) dishes as shown in Figure 2. We adapted a method previously described to orient zebrafish embryos for brain ventricle injections [32]. In adapting the method for Xenopus, we considered that the overall length of a stage 28 embryo is approximately 4 mm. For 60×15 mm plastic Petri dishes with a base diameter of 5.0 cm, 6.9 ml of agarose gives a thickness of ~3.5 mm allowing the head to protrude slightly out of the foxhole. The holes are punched out under water using a Pasteur pipette broken 1.0 cm from its tip (5 3/4” flint glass, Cat#13–678–6A; Fisher). This gives a final hole diameter of ~1.25 mm, which accommodates the embryo well. It is best to arrange the holes in linear arrays, and position each embryo with the dorsal side toward the micromanipulator. This facilitates performing numerous injections and keeping track of the injected embryos.

A generally useful trick for gently positioning embryos without directly handling them is to make use of their cement gland, an anterior-ventral pigmented structure. Embryos use this mucin-secreting organ to attach to stationary supports before being able to swim. The cement gland also appears to have a sensory function, stopping the embryo from moving once attached [33]. Embryos will stick to the forceps tip at the cement gland, and can then be dragged up and over the foxholes. The embryo is then pushed backward into the hole while guiding the tail end downward with a second forceps.

Subretinal delivery methods

Subretinal delivery can be accomplished by direct optic vesicle or brain ventricle injection. Securing the embryos for injection using either method is described above. For optic vesicle injection, embryos were tilted 30° to one side in clay dishes; for brain ventricle injections, the embryos were positioned vertically in agarose dish foxholes (Figure 2 and Figure 3). For both types of injection, glass micropipettes (1BBL no fill 1.0mm 4 in. cat# 1B100–4; World Precision Instruments, Sarasota, FL) were prepared on a micropipette puller. Tips were carefully broken by hand using jeweler’s forceps to a 20 to 30 µm outer diameter. The micropipettes were positioned with a Manual LH mechanical micromanipulator (Harvard Apparatus; Holliston, MA). We found that particularly for the brain ventricle injections, the micropipette could be positioned equally well by hand. This was more efficient when injecting numerous embryos (>25 embryos per dish). A PLI-100 microinjection system (Harvard Apparatus) was used to pressure inject approximately 10.0 nl of test solution to either the optic vesicle or brain ventricle. Typically, pressure injection per se was not necessary, and filling was accomplished simply by the system’s backpressure. Diluted blue food coloring (brilliant blue) was useful to help visualize the injection. The injected dye can be visualized accumulating in the optic vesicle lumen or brain ventricle under the ectoderm. In the last injection series of the included movie (Figure 4), the darkening of the optic vesicle represents such filling. Extension into the neural canal could often be observed. Materials injected included India ink, fluorescein isothiocyanate (FITC) dextran, and fluorescent 0.04 µm latex-polystyrene microspheres (FluoSpheres®, Life Technologies, Carslbad, CA). Cy3-FluoSpheres® were used to visualize injections into living embryos using a Stemi SV 11 epifluorescence microscope with an HBO103 mercury lamp and 29004 ET-Cy3 filter cube (ET545/25×, T565lpxr, ET605/70m; Chroma, Bellows Fall, VT). In contrast to the developing mammalian retina, Xenopus embryonic tissues, including the retina, contain autofluorescent cytoplasmic yoke granules. Cy3 and FITC dyes were selected for the injection markers, as they could be clearly distinguished from the tissue autofluorescence while allowing the possibility of using additional fluorochromes emitting further in the red for immunofluorescence double labeling. The reader is referred to additional helpful published protocols on the subject [34,35].

Histology and microscopy

For imaging the FluoSpheres® injection in vivo, living embryos were visualized with a Zeiss Stemi 11 fluorescence stereomicroscope. For histological evaluation, anesthetized embryos and tadpoles were fixed in 4% paraformaldehyde and cryoprotected through graded sucrose solutions (final, 20% sucrose), followed by infiltration with Optimal Cutting Temperature compound (Ted Pella, Redding, CA; see notes in Table 1). Alternatively, fixed embryos were dehydrated in ethanol and embedded in Lowicryl resin according to the manufacturer (London Resin Co.). Lowicryl sections have excellent morphological detail, and being a methacrylate-based hydrophilic resin, Lowicryl is compatible with the latex-polystyrene FluoSpheres®. Confocal microscopy was performed on an Axiovert 200 motorized microscope equipped with a Zeiss LSM 510 system (laser lines: GFP, ex 477 nm; Cy3, ex 565 nm). The multiphoton capability used to visualize 4',6-diamidino-2-phenylindole (DAPI) was provided through a Coherent Chameleon Ultra II laser. Data collection and image analysis were performed with LSM software version 4.2, and figures created in Adobe Creative Suite version 4.0 (San Jose, CA).

Results

We compared the effect of embryonic stage and injection route on the ability to target the subretinal space. In our initial experiments, we injected the optic vesicle directly at stages 24 through 26. The most consistent fills were obtained at stages 24 and 25, where the optic vesicle lumen appears to be at its largest volume. Targeting the vesicle is more difficult beyond stage 26 due to flattening of the lumen. We positioned the micropipette on the transverse plane near the apex of the protruding vesicle. Filling of the injected vesicle with FITC dextran was almost immediate, followed by the brain ventricle and contralateral optic vesicle (Figure 5). After one to two minutes, the label diffused into the central canal (white asterisk). Compared to dextran, the fluorescent microspheres showed less rapid canal diffusion, but readily filled the optic vesicles and brain ventricle (see below).

Our second strategy took advantage of the continuity of the brain ventricle with the lumen of the embryonic optic vesicle. Although the ventricle lumen is accessible from a variety of injection angles, we found that an anterior midline approach gave the most consistent fills. As shown in Figure 3A, the embryonic diencephalic ventricle is located along the midline, and protrudes anteriorly just under the frontal ectoderm. A midline anterior approach conveniently accesses the ventricle without disturbing the optic vesicles. Securing the embryos in an upright vertical position in agarose dishes allowed convenient access to the frontal cephalic region, and facilitated the injection of multiple embryos (Figure 2).

We anticipated that material introduced into the diencephalic ventricle would diffuse into the optic vesicles bilaterally; vesicle involution would sequester the material into embryonic subretinal space. The distribution of fluorescent microspheres after stage 26 injections are shown in a living embryo (Figure 6) ~15 min after the injection, and histologically (Figure 7) following optic cup formation (~stage 28). Figure 6A,C shows dorsal and lateral views, respectively, illustrating the optic vesicles, which normally appear to bulge from the head laterally (slanted arrows). Viewed dorsally under fluorescence, filling of both vesicles can be seen; accordingly, when viewed laterally, filling of the diencephalic ventricle can also be observed (vertical arrow). In Figure 7, formation of the optic cup is complete, and the outer layer of the cup has become pigmented, which is indicative of RPE differentiation (Figure 7A). The outer retina shows little evidence of differentiation such as retinal lamination or outer segment formation. The tissue in this photomicrograph was embedded in methacrylate, a hydrophilic resin, which provides higher resolution than paraffin or frozen sections, and is compatible with the FluoSpheres® (organic solvents such as xylenes required for wax infiltration compromise the integrity of the microspheres). When viewed under fluorescence (Figure 7B), microspheres injected in the contralateral vesicle are found distributed at the interphase between the RPE and neural retina. Whether the microspheres are present in the interphotoreceptor matrix or in the cytoplasm of cells bordering the subretinal space cannot be determined from this photomicrograph. As the embryo continues to develop into a free-swimming tadpole, we do not detect any distortion or malformation of the orbital structures (Figure 8A), and retinal labeling appears as fluorescence emanating through the pupil (arrow in Figure 8B). These findings support the conclusion that subretinal access via brain ventricle injection or injection of the contralateral optic vesicle does not disturb the normal formation of the ocular structures, nor does it prevent the normal involution of the vesicle and apposition of the RPE with the neural retina.

Access to the optic vesicle via the brain ventricle was dependent on the embryonic stage. Figure 9 compares embryos at stages 29 and 31 photographed immediately after injection. Here, diencephalic ventricular injection supported optic vesicle filling at stage 29, but not at stage 31 (Figure 9B). The two embryos in Figure 10 were injected either at stage 24 or 31, and allowed to mature to swimming tadpoles. Although the eyes of both tadpoles developed normally, fluorescence was detected only in the tadpole injected at stage 24. These results suggest that the connection between the brain ventricle and optic vesicle closes beyond stage 30.

In Figure 11, the diencephalic ventricle of a GFP-rod opsin transgenic embryo was injected through the anterior prosencephalon at stage 29. The embryo was then allowed to develop to stage 41. As in other embryos receiving such injections, the external appearance of the eye and pupil were unaffected (Figure 11A-C). FluoSphere® labeling of the retina and photoreceptor GFP expression could be observed through the pupil. The cryosection shown in Figure 11D,E uses DAPI as a fluorescent nuclear counterstain. GFP expression is restricted to rod photoreceptors, which are still undergoing outer segment elongation (arrowheads, Figure 11E). The FluoSpheres® labeled the entire diencephalon (asterisk, Figure 11D), presumably by direct diffusion from the ventricle lumen. In Figure 11D, the microspheres extend from the diencephalon to the eye along the embryonic optic nerve (arrowhead, Figure 11D). Higher magnification shows the FluoSpheres® in the ventral subretinal area (Figure 11E). Peripherally, the microspheres are focally present within the outer and inner nuclear layers (arrows, Figure 11E). Overall, the embryo shows normal anatomic relationships of the diencephalon, eye, and gastrointestinal lumen. Furthermore, formation of the eye and lens appears to have taken place normally with lamination of the retina into its ganglion cell, and inner and outer nuclear layers. The above findings are consistent with the developmental stage, and suggest that the procedure does not significantly affect the formation of the eye and retina.

Discussion

Direct access to the subretinal space in vivo is problematic due to the consequences of retinal detachment. However, in vitro systems do not always adequately model the complex interactions between the RPE, interphotoreceptor matrix, and neural retina. The motivation for the present study was to develop a strategy to introduce molecules into the subretinal space without disturbing the relationship between the RPE and neural retina. We reasoned that material injected into the common embryonic CNS lumen would diffuse to both optic vesicles. Subsequent involution of the vesicles would bring the material into the subretinal space of the optic cups. A particular advantage of this method is that the injected material has access to the entire RPE, photoreceptor layer, and Müller cell glia. In contrast, subretinal injections as typically performed generally allow access only to a limited portion of the interphotoreceptor matrix [9]. Previous studies in Xenopus using intraretinal injection and electroporation achieved only focal labeling of the retina [36]. Although the overall strategy is feasible, advantages and limitations of the method are presented so that it can be tailored for different experimental paradigms. With a little practice, the injections can be performed routinely even by those with little Xenopus experience.

We began by attempting to inject the optic vesicles directly. Direct injection into the optic vesicle could be advantageous when the opposite retina is needed as a control, or when brain injection is not desirable (for example, in optic nerve tracking studies). However, although it can be an effective way to deliver to the subretinal space of both retinas, the success rate was low (~25%) even with experience. Furthermore, one of the eyes was unavoidably subjected to the trauma of the injection itself. The main problem was accurately accessing the appropriate depth of penetration for the micropipette. It should be kept in mind that the vesicle lumen at most stages is a compressed flattened chamber, and therefore difficult to hit (note that the lateral protrusions of the optic vesicle are mainly due to the thickening of the developing inner neural retina). Ohnuma et al. [37] reported some experience using direct optic vesicle microinjections to transfect plasmid constructs in Xenopus laevis. Using stage 24 embryos, they found that the transfection efficiency for injections made “outside” the vesicle (lumen region) was only half that of injections made “inside” (between the neural retina and ectoderm). It is not known whether this regional difference reflects a biologic difference, or difficulties in targeting the optic vesicle lumen. Brain ventricle injection as an avenue to deliver material to the optic vesicles was not attempted in that study, or in any other study that we are aware of in the literature.

Brain ventricle injection had several advantages over direct optic vesicle injection. The most significant was that the diencephalic ventricle has a larger volume compared to that of the optic vesicle. This allows it to accept injections along a wider range of micropipette penetration depths. Ninety percent or more of the injections are successful. Furthermore, brain ventricle injection avoids subjecting an eye to the trauma of the injection. Using the brain ventricular injection, we have not observed any abnormality in the development of the ocular structures. The pupil, rate and extent of pigmentation, external appearance of the globe, histological lamination of the retina, and rod photoreceptor differentiation all appear to be indistinguishable from controls. Finally, the brain ventricle method lends itself well to a higher throughput compare to direct optic vesicle injections. Here, the use of agarose dish foxholes facilitates orienting numerous embryos for injection. Hundreds of embryos can be injected in a single morning. Finally, it should be pointed out that any experimental paradigm using the diencephalic ventricle injection or the direct optic vesicle route should include appropriate experimental controls. Such controls would be critical to distinguishing a genuine biologic phenotype from nonspecific effects of the injection, carrier, and/or the test molecule itself.

Several issues should be considered when planning experiments using the delivery method described here. A potential limitation of the method is that delivery may not be uniform throughout the subretinal compartment for all developmental stages. In the brain ventricle injection shown in Figure 11, labeling was restricted to the dorsal retinal space. This may represent a positional effect due to gravity, or a regional difference in the accessibility of the subretinal space. It should be pointed out that confinement of any material introduced into the subretinal space will depend on a variety of factors, including the rate at which the material leaves the space. Turnover of molecules introduced into the interphotoreceptor matrix could depend on the integrity of the external limiting membrane and the size of the molecule. The exclusion limit of the zonulae adherens in rabbits is between 30Å and 36Å [3]. For example, interphotoreceptor retinoid-binding protein (140 kDa), a normal component of the interphotoreceptor matrix, exceeds this exclusion limit, and is therefore confined to the subretinal space where it accumulates [4]. However, smaller molecules would not be excluded, and therefore should diffuse into the retina (this may be desired depending on the experimental objectives). The labeling of the peripheral retina (Figure 11E) may be due to incomplete formation of the zonulae adherens in this less-developed region of the retina. Thus, confinement to the subretinal space may be dependent on the stage at which the injection is performed. Ongoing experiments in our laboratory are addressing the above issues.

Brain ventricle injection as a way to deliver molecules or microspheres to the subretinal space can be applied only when the optic vesicle lumen is continuous with the CNS ventricular space. There is little information in the literature about the timing of the closure of the passageway. Detailed published anatomic drawings suggest that the passage between the compartments is lost from stage 26 onward [24]. In contrast, we found that FluoSpheres® can diffuse from the diencephalic ventricle to the optic vesicle through stages 29/30. From stage 31 onward, the connection between the brain ventricle and subretinal compartment does not allow passage of the 0.04 µm microspheres. This does not rule out that the passage is patent to materials < 0.04 µm in diameter.

The subretinal delivery strategy described here may be applicable to some other systems where the subretinal space forms through the involution of the optic vesicle. In its present form, the method could probably not be applied to most invertebrates due to the absence of a common shared chamber similar to the vertebrate subretinal space. Although our experience to date is limited to Xenopus laevis, given the similarity of their retinas, the method should be applicable to Xenopus tropicalis. The fact that development occurs in the extrauterine environment is an obvious advantage of amphibians over rodents for this method. Subretinal delivery via brain ventricle injection can probably be applied to other vertebrates such as avian embryos where access to the brain ventricle is feasible. A strategy has been described for injection into the brain ventricular system of zebrafish [32,38]; however, to our knowledge, brain ventricle injection has not been performed at the early developmental stages where labeling of the retina might be expected. Nevertheless, the utility of the method in this species may be limited, as the zebrafish optic cup does not appear to form from a typical optic vesicle involution sequence as in other vertebrates [39].

In summary, the technique described here provides a strategy to introduce molecules into the subretinal space without disturbing eye development. The approach, which takes advantage of the common lumen between the brain ventricle and optic vesicles during development, may provide a practical approach to routinely deliver molecules to the subretinal space without disturbing the normal anatomic relationships between the RPE and retina. This could allow the introduction of molecules, including tagged proteins, antibodies, DNA constructs and morpholino oligonucleotides, for cellular and genetic manipulations. The method can probably be adapted to other experimental systems where the optic vesicle and/or embryonic brain ventricle can be accessed.

Acknowledgments

The authors thank Dr. Robert Grainger, Dr. Carla Green for helpful advice, and Guoqing Huang and Dugrgesh Kumar for technical assistance. The work was supported by Merit Review Award I01BX007080 from the Biomedical Laboratory Research & Development Service of the Veterans Affairs Office of Research and Development (F.G.-F.), NIH RO1 EY09412 (F.G.-F), R24 EY 016662 core instrumentation grant, American Association for Laboratory Animal Science Travel Award from Lab Products, Inc. (C.A.D.), and an Unrestricted Research Grant from Research to Prevent Blindness to the Department of Ophthalmology at SUNY at Buffalo. Dr. Gonzalez-Fernandez holds the Ira Gile Ross & Elizabeth Pierce Olmsted Ross Endowed Chair in Ophthalmic Pathology.

References
  1. Fuhrmann S. Eye morphogenesis and patterning of the optic vesicle. Curr Top Dev Biol. 2010; 93:61-84. [PMID: 20959163]
  2. Gonzalez-Fernandez F. Development of the Retina. In: Reynolds JD, Olitsky, S.E., editors. Pediatric Retina. New York: Springer; 2010. p. 1–38.
  3. Bunt-Milam AH, Saari JC, Klock IB, Garwin GG. Zonulae adherentes pore size in the external limiting membrane of the rabbit retina. Invest Ophthalmol Vis Sci. 1985; 26:1377-80. [PMID: 4044165]
  4. Gonzalez-Fernandez F. Interphotoreceptor retinoid-binding protein–an old gene for new eyes. Vision Res. 2003; 43:3021-36. [PMID: 14611938]
  5. Hollyfield JG. Hyaluronan and the functional organization of the interphotoreceptor matrix. Invest Ophthalmol Vis Sci. 1999; 40:2767-9. [PMID: 10549633]
  6. Hageman GS, Kuehn MH. Biology of the interphotoreceptor matrix-retinal interface. In: Marmor MF, Wolfensberber TJ, editors. The Retinal Pigment Epithelium. New York: Oxford University Press; 1998. p. 361 −91.
  7. Mieziewska K. The interphotoreceptor matrix, a space in sight. Microsc Res Tech. 1996; 35:463-71. [PMID: 9016449]
  8. Gonzalez-Fernandez F, Ghosh D. Focus on Molecules: interphotoreceptor retinoid-binding protein (IRBP). Exp Eye Res. 2008; 86:169-70. [PMID: 17222825]
  9. Dureau P, Legat L, Neuner-Jehle M, Bonnel S, Pecqueur S, Abitbol M, Dufier JL. Quantitative analysis of subretinal injections in the rat. Graefes Arch Clin Exp Ophthalmol. 2000; 238:608-14. [PMID: 10955663]
  10. Timmers AM, Zhang H, Squitieri A, Gonzalez-Pola C. Subretinal injections in rodent eyes: effects on electrophysiology and histology of rat retina. Mol Vis. 2001; 7:131-7. [PMID: 11435999]
  11. Johnson CJ, Berglin L, Chrenek MA, Redmond TM, Boatright JH, Nickerson JM. Technical brief: subretinal injection and electroporation into adult mouse eyes. Mol Vis. 2008; 14:2211-26. [PMID: 19057658]
  12. Bok D, Yasumura D, Matthes MT, Ruiz A, Duncan JL, Chappelow AV, Zolutukhin S, Hauswirth W, LaVail MM. Effects of adeno-associated virus-vectored ciliary neurotrophic factor on retinal structure and function in mice with a P216L rds/peripherin mutation. Exp Eye Res. 2002; 74:719-35. [PMID: 12126945]
  13. Lai CM, Yu MJ, Brankov M, Barnett NL, Zhou X, Redmond TM, Narfstrom K, Rakoczy PE. Recombinant adeno-associated virus type 2-mediated gene delivery into the Rpe65−/− knockout mouse eye results in limited rescue. Genet Vaccines Ther. 2004; 2:3 [PMID: 15109394]
  14. Vollrath D, Feng W, Duncan JL, Yasumura D, D'Cruz PM, Chappelow A, Matthes MT, Kay MA, LaVail MM. Correction of the retinal dystrophy phenotype of the RCS rat by viral gene transfer of Mertk. Proc Natl Acad Sci USA. 2001; 98:12584-9. [PMID: 11592982]
  15. Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVail MM. Photoreceptor degeneration in inherited retinal dystrophy delayed by basic fibroblast growth factor. Nature. 1990; 347:83-6. [PMID: 2168521]
  16. Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVail MM. Basic fibroblast growth factor and local injury protect photoreceptors from light damage in the rat. J Neurosci. 1992; 12:3554-67. [PMID: 1527595]
  17. Perry J, Du J, Kjeldbye H, Gouras P. The effects of bFGF on RCS rat eyes. Curr Eye Res. 1995; 14:585-92. [PMID: 7587305]
  18. Yu MJ, Shen WY, Lai MC, Constable IJ, Papadimitriou JM, Rakoczy PE. The role of vascular endothelial growth factor (VEGF) in abnormal vascular changes in the adult rat eye. Growth Factors. 2000; 17:301-12. [PMID: 10801078]
  19. Maguire AM, Simonelli F, Pierce EA, Pugh EN, , Jr Mingozzi F, Bennicelli J, Banfi S, Marshall KA, Testa F, Surace EM, Rossi S, Lyubarsky A, Arruda VR, Konkle B, Stone E, Sun J, Jacobs J, Dell'Osso L, Hertle R, Ma JX, Redmond TM, Zhu X, Hauck B, Zelenaia O, Shindler KS, Maguire MG, Wright JF, Volpe NJ, McDonnell JW, Auricchio A, High KA, Bennett J. Safety and efficacy of gene transfer for Leber's congenital amaurosis. N Engl J Med. 2008; 358:2240-8. [PMID: 18441370]
  20. Dinculescu A, Glushakova L, Min SH, Hauswirth WW. Adeno-associated virus-vectored gene therapy for retinal disease. Hum Gene Ther. 2005; 16:649-63. [PMID: 15960597]
  21. Bainbridge JW, Smith AJ, Barker SS, Robbie S, Henderson R, Balaggan K, Viswanathan A, Holder GE, Stockman A, Tyler N, Petersen-Jones S, Bhattacharya SS, Thrasher AJ, Fitzke FW, Carter BJ, Rubin GS, Moore AT, Ali RR. Effect of gene therapy on visual function in Leber's congenital amaurosis. N Engl J Med. 2008; 358:2231-9. [PMID: 18441371]
  22. Iribarne M, Canto-Soler MV, Torbidoni V, Suburo AM. Controlling retinal pigment epithelium injury after experimental detachment of the retina. Invest Ophthalmol Vis Sci. 2007; 48:1348-54. [PMID: 17325183]
  23. Ivert L, Kjeldbye H, Gouras P. Long-term effects of short-term retinal bleb detachments in rabbits. Graefes Arch Clin Exp Ophthalmol. 2002; 240:232-7. [PMID: 11935282]
  24. Hausen P, Riebesel M. The Early Development of Xenopus Laevis. New York: Springer-Verlag; 1991.
  25. Batni S, Scalzetti L, Moody SA, Knox BE. Characterization of the Xenopus rhodopsin gene. J Biol Chem. 1996; 271:3179-86. [PMID: 8621718]
  26. Ishibashi S, Kroll KL, Amaya E. A method for generating transgenic frog embryos. Methods Mol Biol. 2008; 461:447-66. [PMID: 19030816]
  27. Hutcheson DA, Vetter ML. Transgenic approaches to retinal development and function in Xenopus laevis. Methods. 2002; 28:402-10. [PMID: 12507458]
  28. Knox BE, Schlueter C, Sanger BM, Green CB, Besharse JC. Transgene expression in Xenopus rods. FEBS Lett. 1998; 423:117-21. [PMID: 9512341]
  29. Sive HL, Grainger RM, Harland RM. Early Development of Xenopus laevis - a laboratory manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press; 2000.
  30. Nieuwkoop PD, Faber J. Normal Table of Xenopus laevis. 2 ed. Amsterdam: North Holland Publishing Company; 1967.
  31. Khokha MK, Chung C, Bustamante EL, Gaw LW, Trott KA, Yeh J, Lim N, Lin JC, Taverner N, Amaya E, Papalopulu N, Smith JC, Zorn AM, Harland RM, Grammer TC. Techniques and probes for the study of Xenopus tropicalis development. Dev Dyn. 2002; 225:499-510. [PMID: 12454926]
  32. Gutzman JH, Sive H. Zebrafish brain ventricle injection. J Vis Exp. 2009; pii:1218 [PMID: 19352312]
  33. Sive H, Bradley L. A sticky problem: the Xenopus cement gland as a paradigm for anteroposterior patterning. Dev Dyn. 1996; 205:265-80. [PMID: 8850563]
  34. de Mazière AM, Hage WJ, Ubbels GA. A method for staining of cell nuclei in Xenopus laevis embryos with cyanine dyes for whole-mount confocal laser scanning microscopy. J Histochem Cytochem. 1996; 44:399-402. [PMID: 8601700]
  35. Lee C, Kieserman E, Gray RS, Park TJ, Wallingford J. Whole-mount fluorescence immunocytochemistry on Xenopus embryos. CSH Protoc. 2008
  36. Falk J, Drinjakovic J, Leung KM, Dwivedy A, Regan AG, Piper M, Holt CE. Electroporation of cDNA/Morpholinos to targeted areas of embryonic CNS in Xenopus. BMC Dev Biol. 2007; 7:107 [PMID: 17900342]
  37. Ohnuma S, Hopper S, Wang KC, Philpott A, Harris WA. Co-ordinating retinal histogenesis: early cell cycle exit enhances early cell fate determination in the Xenopus retina. Development. 2002; 129:2435-46. [PMID: 11973275]
  38. Lowery LA, De Rienzo G, Gutzman JH, Sive H. Characterization and classification of zebrafish brain morphology mutants. Anat Rec (Hoboken). 2009; 292:94-106. [PMID: 19051268]
  39. Schmitt EA, Dowling JE. Early eye morphogenesis in the zebrafish, Brachydanio rerio. J Comp Neurol. 1994; 344:532-42. [PMID: 7929890]
  40. Rapp LM, Basinger SF. The effects of local anaesthetics on retinal function. Vision Res. 1982; 22:1097-103. [PMID: 6983179]
  41. Bernstein PS, Digre KB, Creel DJ. Retinal toxicity associated with occupational exposure to the fish anesthetic MS-222. Am J Ophthalmol. 1997; 124:843-4. [PMID: 9402834]