$$\rightleftharpoonup{xx}$$
$$\longleftharp{xx}$$,
$$\longrightharp{xx}$$,
Using a rabbit model, a safe and reproducible method is presented for transvitreal delivery of cultured RPE on cell carriers into the subretinal space with a custom-designed shooter instrument. The described method offers a short/optimized surgical technique for easy learning, as it involves standard techniques in vitrectomy with subretinal maneuvers. Outcome is greatly facilitated by a clean vitreoretinal interface, and intraocular infusion that avoids fluid turbulence over the implantation site, inducing bleb retinal detachment (bRD) at low IOP, preventing retina and sclera damage through dryness, and appropriate positioning of the rabbit.
We caution however, as several intra-operative complications may occur any time, hindering the success of the implantation, for instance intra ocular bleedings, anesthesia fading off during vital steps such as the implantation, collapsing of bRD due to instrument manipulation or ocular hypotony, rabbit death due to excessive doses of anesthesia, low blood pressure during long operation causing hypoxic brain damage, or hyperthermia. Yet these complications decrease with time as they are quickly tackled and resolved by increasing experience of the surgical team.
Some complications could be reduced by following a few simple, yet crucial steps. Lubricant should be added every 5 - 10 min to prevent corneal, scleral and conjunctival damage during the operation, and to maintain a clear intraocular media, as dried/blackened sclera may be a cause of wound dehiscence, which in turn leads to ocular hypotony and/ or intraoperative leakage from sclerotomies. Heparin should be added to prevent the formation of a fibrin film that makes particularly subretinal implantation challenging and simultaneously adding epinephrine to reduce bleeding under heparin16. Too long heparin/ epinephrine exposure times (>1 hour) should be avoided to prevent corneal edema by endothelial decompensation33, hypertensive crisis or intraoperative fatality. Meticulous vitreous removal should be performed at instrument (entry) port to avoid retinal and/ or choroidal detachments. Intraocular instruments should be pointed towards posterior pole to avoid lens touch (causes iatrogenic cataract formation) or (entry site) retinal damage. An intraocular side-port infusion cannula should be used, as it attenuates the jet stream around the implantation area, thus preventing uncontrolled tearing of the retinotomy, and collapse of the bRD. bRD induction in the midline (vertical axis from optic nerve) or close to optic medullar fibers should be avoided to prevent extensive iatrogenic retinal detachments. Finally, last but not least bRD should be induced at low IOP, to avoid subretinal BSS injection using excessive flow rates which may lead to a retinal damage (e.g., by stretching).
Many study variables such as cell carrier variants, fetal, adult or stem cell derived RPE cell sources, choices for immunosuppressants, etc., can be explored 14,26,27,34. Further improvement such as serum-free RPE culture methods, characterization of xenoRPE in subretinal space, removal of the host RPE layer14 or strategies for implant anchorage are current work in progress.
To date the described techniques have been used on 5 different rabbit strains, including chinchilla bastard, Chinchilla bastard/KBL hybrids, New Zealand White/Red Cross, New Zealand White (albino) and Dutch belted. Both male and female rabbits were operated on, with rabbits at least 1.5 kg or 2 months of age (depending on species). Most surgeries were on pigmented rabbits (chinchilla bastard or chinchilla bastard hybrids) with weights between 2.5 - 3 kg.
All of the rabbit strains we have had the opportunity to work with seem to have some peculiarities. Given the exclusive availability of pigmented rabbits of the chinchilla bastard strain in Germany in 2009-13, we have collected the most experience with these animals. Unfortunately it is no longer available, since breeding has been discontinued, but compares very well to New Zealand White/Red Cross except for the more advantageous thicker sclera and larger eye volumes in the latter. Chinchilla bastard hybrids have significant intraoperative fibrin formation and require heparin/epinephrine use as outlined above to ensure successful subretinal maneuvers. This protocol has also been performed in non-pigmented albino rabbits (New Zealand White), however particularly bRD creation and subretinal implantation is more challenging given reduced contrast appreciation. The feasibility of inducing a posterior vitreous detachment did not seem rabbit strain dependent in our hands.
Transvitreal subretinal delivery is likely the future surgical strategy of choice given it is the most common route nowadays clinically to access the retina. As a result many other groups have presented such techniques for cultured RPE on carrier supports in animal work11,15,23,35. Aramant et al.36 have an instrument, which places rather than pushes their hydrogel-encapsulated soft implant to its subretinal target site. The design of Thumann et al. utilizes a hollow spatula, which releases the carrier-supported graft by floating it off through fluid injection 19. Both former strategies require subretinal insertion of the instrument, which in our view is more prone to complications, when compared to an epiretinally appositioned instrument. Montezuma et al.22 described a subretinal inserter instrument for the delivery of subretinal chip implants in pigs but no further work has been published since to the best of our knowledge. We have been able to extend the described technique with some modification to pig.
Our preferred cell carriers are 10 micron thick polyester terephthalate (PET) membranes. From a surgical perspective, this material has favorable stiffness and elasticity parameters, in addition to its broad versatility during cell culture experiments. We found similar experiences with expanded tetrafluoroethylene (ePTFE) 37 or nanofiber membranes electrospun from PET, poly-lactic/capronolactic acid (PLCL) or poly--lactic-co-glycolic acid (PLGA), as well as composite nanofiber (PLGA or PET) and ultrathin PET 26. When PET membranes are used with our metallic shooter instrument, they do have an occasional tendency to exhibit electrostatic charge, which challenges their ejection from the shooter 27. Ultrathin polyimide membranes could in our hands not be implanted in the subretinal space with the protocol outlined above (manuscript in preparation).
Marmor et al. have systematically studied spontaneous resorption of subretinal fluid in iatrogenic localized retinal detachments38-41. Even following manipulation in the subretinal space these were found to be reabsorbed by postoperative day 4 in uneventful surgeries. Laser retinopexy is not performed to secure the edges of the retinotomy. Although counterintuitive when compared to human surgery, air/ gas tamponade is not required. Unless meticulous removal of peripheral vitreous can be achieved, particularly in the superior quadrant, this may in fact result in giant retinal tears originating from the retinotomy site. It is only recommended to perform fluid air exchange with subsequent 20% SF6 gas tamponade to salvage intraoperative iatrogenic retinal detachments or in case a particular implant position needs to be secured.
Although the mechanically induced ablation of the neural retina can cause RPE and photoreceptor damage in rabbits42,43, its extent varies greatly (even with regular BSS) depending on factors such as IOP, syringe type used, injection volume with thereby induced retinal stretching, etc. We have also tested the often recommended Ca/Mg-free BSS facilitated detachment 42-44, but found that it causes intraoperative lens opacification (particularly with elevated temperature), and significantly delays or even impairs retinal re-attachment 27. Slow subretinal injection of 20 - 30 µl volume of regular BSS with a 100 µl syringe is therefore recommended; injection needle movements should be minimal so the retinotomy seals around it and prevent Bruch's membrane damage. Some of the iatrogenic damage may be resolved by RPE wound healing, and the observed relative preservation of ONL thickness after reattachment, suggests that the RPE/photoreceptor complex can tolerate this impairment, as also described by others45.
Cell-based therapeutics or retinal prosthetics require preclinical animal testing prior to regulatory approval and commencing human safety studies. The former vary from country to country. The rabbit model described here may serve as a cost-efficient and less challenging platform for establishing or even carrying out all requirements by regulatory authorities. Moreover, it may subsequently serve for training of surgeons in eventual multicenter clinical trials or further improvements of the technique along the way.