$$\rightleftharpoonup{xx}$$
$$\longleftharp{xx}$$,
$$\longrightharp{xx}$$,
There are two main approaches being evaluated for submacular RPE transplantation-the injection of an RPE suspension and the transplantation of a monolayer RPE graft. A detailed comparison between the two methods is beyond the scope of this manuscript. However, the transplantation of a monolayer RPE graft may be advantageous as the RPE cells are more organized in a monolayer than in a suspension. RPE cells in the graft are organized in a confluent monolayer, which resembles the organization of the physiological RPE cell layer and enables the transplanted RPE cells to perform their physiological functions. This enables more precise dosing parameters compared to cell suspensions, which is highly relevant for regulatory work and industrial scale-up.
Delivery of the RPE patch graft into the subretinal space requires careful manipulation of the macula and accurate insertion of the graft in the subretinal space. Technological advances in microsurgery, such as miOCT, and a better understanding of intraoperative retinal tissue dynamics have reduced the learning curve of this procedure. In this discussion, the rationales of the following aspects will be explained: i) pre-operative plasminogen injection; ii) the use of intraoperative miOCT; iii) the use of a custom 41 G dual-bore cannula, low IOP settings, and PFCL for subretinal bleb creation; iv) scraping of the native RPE cell layer before transplantation; v) the use of sirolimus, triamcinolone, doxycycline, and minocycline to reduce immunogenic graft rejection.
Preoperative plasminogen injections release parafoveal retinal adhesions
In the initial experiments, it was challenging to detach the fovea with a single fluid wave. On assessment with miOCT, the images revealed the presence of parafoveal outer retinal adhesions to the native RPE along with evidence of intraretinal trauma20. These adhesions may have led to a vertical expansion of the bleb rather than the subretinal fluid wave spreading across the retinal contour, resulting in foveal trauma. Plasminogen is the inactive precursor of plasmin, a protease targeting fibronectin and laminin. Ocriplasmin is a bioengineered variant of human plasmin, approved by the Food and Drug Administration (FDA) and European Medicines Agency (EMA) for the treatment of symptomatic vitreomacular traction with or without a concomitant macular hole. However, postapproval reports of cystoid macula edema development after ocriplasmin injection have suggested a more extensive effect of the enzyme on the retina23.
Although the exact mechanisms have not been identified, it was suggested that plasmin could weaken retinal adhesion through the degradation of the interphotoreceptor matrix elements responsible for photoreceptor-RPE adhesion24. In this protocol, NHP eyes were treated with intravitreal plasminogen 1 week before surgery to release the parafoveal outer retinal adhesions. Under the assumption that the photoreceptor-RPE adhesion is weakened, a lower force is required to detach the neurosensory retina, including the distal parafoveal ring, which typically resists the subretinal fluid wave20. Thus, the force administered during retinal bleb detachment results in the expansion of the bleb across the retinal contour rather than stretching the retina tangentially. This reduces the risk of foveal tears. However, it should be noted that the effect of plasminogen on long-term graft survival was not studied in this protocol. Future studies should attempt to determine this effect.
miOCT provides anatomical feedback to guide subretinal bleb creation, graft implantation, and subretinal fluid drainage
Intraoperative, atraumatic manipulation of the macula is key to achieving good transplantation outcomes. However, microstructural changes of the macula related to manipulation may not always be evident on the operating microscope. In such procedures, the miOCT is an important tool that provides real-time, three-dimensional, intraoperative feedback of the macular structure. miOCT is especially useful during the steps of foveal detachment, graft implantation, and drainage of the subretinal fluid using a fluid-air exchange. During foveal detachment, miOCT can determine the vertical and horizontal dimensions of the bleb. Foveal microtears, which may not be visualized clearly on the surgical microscope, can be confirmed by miOCT (Figure 3). During the graft implantation, miOCT images guide by showing the graft's location or proximity to the fovea, through the often less-transparent, detached retina. miOCT can also highlight possible areas of retinal adhesion during a difficult transplantation process25. Finally, in the subretinal fluid drainage process, miOCT can reliably guide subretinal fluid drainage until complete retinal-RPE graft contact is achieved.
The combination of a dual-bore cannula, low IOP settings, and PFCL vitreous tamponade synergistically reduces macular trauma during subretinal bleb creation
Tangential retinal stretching and fluid turbulence can occur during the subretinal BSS injection for foveal detachment leading to unwanted foveal tears. To counteract these phenomena, factors, such as the relative position and distance from the foveal center where the injection is initiated, injection volume and velocity, vitreous tamponade, choice of subretinal instrumentation, and IOP have all been shown to be relevant20,26,27. The subretinal bleb for foveal detachment should be situated at a location adequately distant from the fovea, as retinal stretching may be highest at the bleb initiation site27. IOP should also be kept low throughout the creation of the subretinal bleb. When the IOP of the eye is high, a higher vertical increase in bleb size rather than expansion along the contour of the retina is observed, whereas blebs are shallower at lower pressures20. Furthermore, although an intravitreal injection of 50 µL will effectively double the IOP in humans28, given the shorter eye length in NHPs, the IOP rise during subretinal injection will probably be higher and more rapid than in humans. While most vitrectomy machines adjust for IOP fluctuation, the adjustment is not a simultaneous but rather a reactive process that occurs as subretinal injection proceeds. Hence, the higher the IOP, the higher the risk of retinal overstretching and resultant foveal trauma. Thus, it is essential to maintain a stable low IOP during subretinal injection.
A commercial 20/41 G (DORC) or a custom-made 25/41 G dual-bore subretinal cannula is recommended for subretinal injection. The cannula allows fluid to exit the vitreous cavity in exchange for BSS injected into the subretinal space. This ensures the 'simultaneous' regulation of IOP during the subretinal injection. A schematic of the dual-bore cannula is seen in Figure 2. Finally, PFCL is utilized to reduce the risk of foveal tears20,26,27. As PFCLs, such as octaline, have higher specific gravity, they exert a downward force on the retina during foveal detachment29. This further stabilizes the foveal detachment bleb creation process and enhances the expansion of the bleb along the retinal contour. This technique has been successfully used for the subretinal injection of rtPA in the setting of massive submacular hemorrhage due to nAMD30.
Pretransplantation removal of native RPE allows the restoration of RPE-photoreceptor complex
Host RPE should be removed before graft transplantation. This is because the restoration of the RPE-photoreceptor complex is required to enable the RPE transplant to perform its physiological functions of supporting the photoreceptors21. The host RPE, if not removed, may pose as a mechanical barrier, which prevents the restoration of this complex. It can be removed either through the administration of RPE-toxic chemicals or by using physical means of removal. Chemical removal methods include the systemic or subretinal administration of sodium iodate31,32. As sodium iodate causes widespread degeneration of photoreceptors, RPE cells, and Choriocapillaris when administered, its retinal and systemic toxicity precludes its use for human trials32,33. Hence, physical intraoperative techniques are preferred. Various physical methods have been conceptualized. When physical methods are utilized, it is crucial that the Bruch's membrane remains undamaged. Many in vitro studies have demonstrated the dependence of RPE graft survival on an intact Bruch's membrane34,35,36.
Attempts at hydraulic debridement were associated with breaks in Bruch's membrane, an increased rate of epiretinal membrane development, and proliferative vitreoretinopathy, resulting in tractional retinal detachment37. A diamond-dusted spatula proposed for RPE debridement also led to breaks in the Bruch's membrane, resulting in cellular proliferation from the choroid into the subretinal space38. Interestingly, a custom-made extendable loop instrument could remove the overlying RPE with preservation of Bruch's membrane in the eyes of rabbits and pigs11,39. The removal of the underlying RPE is also useful for establishing animal models with RPE and outer retinal atrophy, similar to the advanced atrophic form of AMD. When a focal area of RPE is removed from the macula, the RPE wound closes via the hypertrophy of the remaining RPE cells. However, this wound healing response is associated with atrophy of the outer nuclear layer40. While the creation of an animal model is beyond the scope of this manuscript, a similar procedure can create an animal model of an advanced atrophic AMD phenotype for the testing of RPE-derived cell therapeutics.
The use of sirolimus, triamcinolone, doxycycline, and minocycline to reduce immunogenic graft rejection
The subretinal space is thought to be an immune-privileged site, maintained by an intact blood-retinal barrier and other factors41. In many studies involving the subretinal transplantation of stem-cell derivatives with an intact blood-retinal barrier, immunosuppressive drugs play a negligible role in graft survival42. The outer blood-retinal barrier is thought to be formed by the native RPE layer and the tight junctions between the RPE cells. While native RPE removal allows better integration of the transplanted RPE and host photoreceptors, the blood-retinal barrier is disrupted in the process, increasing the likelihood of an immune rejection. Classically, T-cells are central to the process of transplant rejection of other organs such as the kidney and liver43. Hence, initial immunosuppressive regimens for retinal tissue transplantation were targeted towards reducing these adaptive immune responses.
Sirolimus, a mechanistic target of rapamycin inhibitor, and tacrolimus, a calcineurin inhibitor, are examples of immunosuppressive drugs targeting adaptive immune responses. However, despite adequate T-cell suppression, graft survival rates remain low. In addition, RPE cells are known to suppress T-cell activation through the release of inhibitory factors and promote the generation of regulatory T-cells44. Hence, it has become increasingly apparent that adaptive immunity may not be the only contributor to graft rejection42. Subretinal transplantation of cellular products can result in the accumulation and activation of microglia45.
Microglia are the macrophages of the retina. They consist of two main populations: 1) the perivascular microglia of the inner retinal vasculature and 2) the microglia within the retinal tissue parenchyma. As microglia are part of the innate immune response, intravitreal glucocorticoids, such as triamcinolone, can suppress cytokine-mediated proliferation46. Doxycycline and minocycline can also suppress microglial activation and should be considered47,48. Lastly, differences in immune rejection of RPE allografts versus xenografts are incompletely understood49. For instance, alloantibodies against induced pluripotent stem cell-derived RPE cells have been reported in the serum of in vivo immune rejection models. However, the role of these antibodies and the importance of antibody-mediated rejection in graft survival remains unknown50. Hence, a multidrug regimen utilizing sirolimus for the suppression of adaptive immunity and a combination of triamcinolone, doxycycline, and minocycline for innate immunity suppression is proposed. This regimen has been successfully used in rabbits with good graft survival outcomes and minimal systemic effects11.
Limitations of this surgical technique
This paper describes a possible surgical method to deliver an RPE graft sheet into the subretinal space of NHP; however, this does not mean this is the only optimized way. Different vitreo-retinal surgeons may have other preferences for instrumentation and technique. For example, this implantation device design can only deliver flat implants supported with a stiffer cell carrier and hence may not be suitable for relatively flexible (or rolled) implants. RPE suspension transplants can omit much of this technique. Accordingly, surgical details will require modification based on each delivery strategy.
As interest in cellular therapeutics for the treatment of degenerative retinal diseases continues to grow, the NHP animal model will be essential in preclinical studies for studying the factors affecting RPE graft survival. In this manuscript, strategies are proposed to enable the smoother delivery of a submacular monolayer RPE graft in the NHP eye. Methods for better visualization of intraoperative complications are also recommended. It is anticipated that these methods will continue to improve as the use of cellular therapeutics expands. Future method papers should also consider proposing a comprehensive list of investigations to assess various structural and functional aspects of the graft.