Ex vivo cultivated oral mucosal epithelial cell transplantation for limbal stem cell deficiency: a review

Destruction or dysfunction of limbal epithelial stem cells (LESCs) leads to unilateral or bilateral limbal stem cell deficiency (LSCD). Fifteen years have passed since the first transplantation of ex vivo cultivated oral mucosal epithelial cells (COMET) in humans in 2004, which represents the first use of a cultured non-limbal autologous cell type to treat bilateral LSCD. This review summarizes clinical outcomes from COMET studies published from 2004 to 2019 and reviews results with emphasis on the culture methods by which grafted cell sheets were prepared.


Background
Damage to the limbus can lead to a decrease in limbal epithelial stem cells (LESCs) and dysfunctional homeostasis of the corneal epithelium. This failure, termed limbal stem cell deficiency (LSCD) [1][2][3], leads to disruption of the barrier function and invasion of conjunctival cells onto the corneal surface [4,5]. Conjunctivalization is followed by vascularization, chronic inflammation, photophobia, recurrent pain, and decreased vision [4,[6][7][8]. LSCD is classified as partial or total and may occur unilaterally or bilaterally [9].
Conjunctival limbal autograft (CLAU) and cultivated limbal epithelium transplantation (CLET) are procedures often used in the treatment of unilateral LSCD [10,11]. However, patients with bilateral total LSCD do not have limbal tissue available for use in either CLAU or CLET. Thus, options for a source of LESCs are limited to living-related or cadaveric donors and entail use of immunosuppression to prevent rejection [12].
In 2004, Nakamura and co-workers performed the first transplantation of autologous oral epithelial cells cultured ex vivo on human amniotic membrane (AM) to offer an alternative to use of allogenic tissue and avoid immunosuppression [13]. The treatment of LSCD using ex vivo cultivated oral mucosal epithelial cell transplantation (COMET) minimizes the risk of graft rejection and has the added advantage that it can be repeated if necessary. However, neo-angiogenesis following transplantation is a drawback associated with this procedure [13]. This review summarizes clinical outcomes from COMET case series from 2004 to 2019 and reviews the methods used in preparation of transplanted cell sheets.
A case report of one patient (one eye) was excluded from this review [37].

Diagnosis
Diagnosis of LSCD is based on the following clinical features: irregular corneal surface with loss of light reflex, corneal epithelial opacity, loss of limbal palisades of Vogt, fluorescein staining, epithelial thinning in a vortex pattern, corneal neovascularization, peripheral pannus, persistent epithelial defect (PED), corneal stroma scarring, and opacification [6,38].

Prognostic factors
The presence of pre-operative epithelial defects and/or poor tear production may affect successful outcome [23,44,45]. Thus, numerous studies included assessment of dry eye in the pre-operative evaluation (Table 1) [13-17, 22-24, 27, 31, 33, 34]. DeSousa et al. recommend that adnexal abnormalities, including the health and function of the eyelids, fornix, and tear film, be assessed and improved prior to surgery to ensure the best chance of epithelial healing [46]. Conjunctival swab has revealed the presence of pathogenic organisms, which is likely due to a poor ocular surface and absence of a tear film. Therefore, performing a conjunctival swab culture before COMET to ensure a receptive ocular surface is suggested [47]. A complete oral exam is also recommended as successful culture of oral mucosal epithelial cells (OMECs) sheets may be affected by poor oral hygiene and smoking [15,34].
Fetal bovine serum (FBS, or FCS when referred to as "fetal calf serum") was used in nine studies [13, 16, 19-21, 28, 34-36], five used autologous serum (AS) [17,22,23,27,31], and four used FBS and AS (Table 3) [15,29,30,32]. Only one was serum-free [33]. Use of AS eliminates exposure to xenogeneic compounds contained in animal serum. One study compared use of AS with FBS and found that cell sheet morphology and expression of structural proteins were similar in both groups [17]. Preliminary in vitro work has also shown that AS promotes similar expression of putative stem cells markers in cultured OMEC sheets compared to use of FBS [31]. The two patients receiving AS feeder-free cultured OMEC sheets in this study had significant improvement in corneal epithelium integrity, pain relief, and visual acuity (VA) [31].

Airlifting
Fifteen studies (258 eyes) used airlifting to promote formation of a stratified epithelium (Table 3) [13, 15-19, 23-25, 27, 29, 30, 32, 34, 36]. Airlifting produced more stratification with four to nine layers compared to two to five in non-air-lifted OMEC sheets. Stratification promotes cell-cell adhesion between superficial epithelial cells via tight junction formation, which helps to prevent loss of the transplant due to blinking [49]. On the other hand, highly differentiated air-lifted sheets have lower proliferative function, which is consistent with a decrease in p63α-expressing stem cells [50].

Carrier
Most studies employed AM as a culture substrate and OMECs were transferred directly on the same substrate NA not available Area of an a 8-, b 6-, or c 3-mm diameter biopsy punch (Table 3). Two studies using temperature-responsive cell-culture inserts transferred cells on a supporter [14] or polyvinylidene fluoride membrane rings [26], which were removed after transfer to the cornea. A filter paper ring was used to transfer cell sheets grown on a fibrin substrate [27]. A support mesh was used in one study employing substrate-free culture [35]. For at least 4 days after the confluence f For at least 5 days after the confluence and then air-lifted for 1 to 2 days *Supporter **Polyvinylidene fluoride (PVDF) ring ***Filter paper ****Support mesh
It has been shown that the presence of at least 3% stem cells (defined as ΔNp63α-positive cells) is associated with clinical success in the treatment of LSCD using CLET [57]. It is likely that the percentage of stem cells in grafted OMEC sheets also influences COMET success. Nishida et al. showed p63 expression in the basal layer of OMEC cultures used in the successful treatment of four patients with LSCD (Table 3) [14]. Analysis of putative stem cell markers (ΔNp63α, ABCG2, and C/EBPδ) in transplants have shown that OMEC and limbal cells have similar expression levels [31]. Four studies employed the colony-forming efficiency (CFE) assay to show the presence of stem cells in OMEC sheets (Table 3) [14,22,26,31]. To date, any correlation between stem cell content in OMEC sheets before transplantation and clinical success using COMET remains to be investigated.

Success rate
Clinical success was most consistently defined in terms of a stable ocular surface. Secondary objectives reported were improved VA and best-corrected VA (BCVA). Post-graft investigations rarely included IVCM [16,21] or IC [19]. Satake et al. used IC to show that in 2/4 eyes, the oral mucosa phenotype persisted for up to 16 months post-operatively, and in some cases the assessed epithelium displayed a mixture of oral mucosal and conjunctival cells [19].
In total, 70.8% (172/243) of eyes receiving COMET achieved a stable ocular surface and were defined as successful (Table 4; see Fig. 2 for detailed results per etiology). This percentage is lower compared to transplantation of cultured limbal epithelial cells (LECs) (75%) [58]. Moreover, one study directly compared COMET to transplantation of allogeneic cultured limbal epithelial transplantation (ACLET) and reported 71.4% (30/42) eyes in the ACLET group achieved a stable ocular surface, versus 52.9% (18/34) eyes in the COMET group. The authors attributed the significantly higher success using ACLET to the lower incidence of post-operative PED, superior LEC proliferation and differentiation, and the ability of LECs to more readily form a stable corneal epithelium [36].

Visual improvement
VA improvement was reported in all but two of the studies ( Fig. 2 and Table 4), and 225/331 (68.2%) eyes had some improvement. An improvement in the BCVA of at least two lines was noted in 172/271 (63.5%) eyes (data from 20 studies). The absent or incomplete description of methodology for VA/BCVA measurement prevented an accurate comparison of results between studies. VA inconsistently measured either before or after subsequent surgeries, such as PKP, was another major confounding factor.

Survival of oral mucosal epithelial cells after grafting
Nakamura et al. have shown that post-COMET specimens exhibit a decrease in the number of epithelial layers from 5 to 6 in successful grafts to 2 to 5 disorganized epithelial layers in unsuccessful grafts [18]. The phenotype of COMET grafts (assessed from corneal buttons retrieved after a secondary procedure, mostly PKP) was also investigated in order to characterize the differences between successful (four samples) and unsuccessful (two samples) graft phenotypes [18]. Successful cases showed the presence of K3, a marker common to oral and corneal epithelium, in all specimens; K12, a corneal-specific keratin, presented only occasional staining in one case. K4 and K13, markers of oral mucosal epithelium, were present in both successful and failed samples. In failed specimens, one presented occasional staining for K3, but both were negative for K12. MUC5AC, a conjunctival goblet cell marker [59], was present only in both failed cases and found absent in successful cases [18].
However, without clear detection of cell origin (donor/ host) [60][61][62] it is difficult to determine clearly whether cultivated OMECs were transdifferentiated into the corneal lineage or whether the presence of corneal epithelial cells represents expansion and migration of remaining corneal cells. In vivo study on rats has shown that transplanted oral mucosal cell sheets were able to survive and retain stem/progenitor cells for at least 8 weeks postoperatively, which results in the long-term success of transplantation of cultured OMECs in LSCD patients [63]. It has been suggested that restoration of a noninflammatory environment post-operatively may be sufficient to allow repopulation of any remaining corneal cells to the ocular surface and/or resumption of normal homeostatic function by residual limbal stem cells [64].
Success of stem cell transplantation and the long-term survival of the graft in ocular surface therapy not only depends on the features of transplanted cells, but also on the surrounding microenvironment, as it provides the necessary signals required for cell maintenance and growth [48,65]. Huang et al. speculate that transplanted OMECs might be regulated by signals originating from healthy stroma and differentiate toward the corneal phenotype, while simultaneously maintaining the oral phenotype [56]. However, identification of the key factors necessary to promote transdifferentiation of OMECs to the corneal phenotype still requires further research.

Post-operative complications
The most common complications described following COMET were epithelial defects (52.8%; 36.1% PED), increased intraocular pressure or glaucoma (15%), stromal melting or perforation (9.4%), and infection (7.2%) (Fig. 3). For comparison, a review summarizing transplantation of cultured LECs (889 eyes) reported that the most common complications post-surgery were bleeding (8.7%), inflammation (7.5%), and blepharitis and epitheliopathy (4%) [58]. Epithelial defects making up more than half of the complications could reflect the often more serious nature of the bilateral LSCD diagnosis that demands an alternative treatment such as COMET.
Of note, there was no consensus on the definition of PED. For instance, Nakamura et al. considered epithelial defects to be persistent if they lasted for more than 4 weeks [24], while Hirayama et al. [27] defined PED occurring after 1 week following failure of conventional therapy. In a retrospective comparative study (76 eyes) a higher incidence of post-operative PED was reported in eyes receiving COMET (9/34 eyes) compared to those receiving ACLET (3/42 eyes) [36]. Several studies pointed to an association between incidence of postoperative with pre-operative PED [15,23,36]. It has also Others: advanced pterygium, Behcet's disease, contact lens hypoxia, contact lens hypoxia + congenital aniridia, cystinosis, drug toxicity, gelatinous drop-like dystrophy, graft-versus-host disease, hepatitis C, idiophatic, Lyell syndrome, multiple eye surgery, neuroparalytic keratitis, ocular trauma, post keratitis, radiation keratopathy, rosacea keratitis, Salzmann's corneal degeneration, severe Groenouw dystrophy, squamous cell carcinoma, and trachoma been shown that the transplanted epithelium exhibits decreased barrier function following COMET [19].
Baradaran-Rafii et al. suggest that PKP is inevitable in most cases of LSCD involving chemical burns due to the presence of significant corneal opacification [34]. Patients receiving PKP had improved visual function and the authors recommended performing PKP several months post-COMET to achieve the best chance of success [34].
Although most studies noted neovascularization posttransplantation [13-15, 17, 20, 21, 24, 26-28, 32-34, 36], they did not define this as a complication of the procedure. Corneal peripheral neovascularization occurred slowly in most cases, during the first postoperative year [33]. However, the central corneal area was usually spared, and neovascularization usually ceased to progress after 1 year, remaining stable thereafter [14,33] or gradually abating with time [15,24]. In the one retrospective study comparing ACLET and COMET, the incidence of neovascularization, corneal conjunctivalization, and improvement in symblepharon was similar between the two groups [36].
Nishida et al. pointed out that stromal vascularization observed beneath COMET transplants on the periphery of the cornea should be differentiated from subepithelial neovascularization that accompanies conjunctival ingrowth, which occurs several months post-transplantation [14]. The peripheral neovascularization seen after COMET may be caused by the lack of antiangiogenic factors, such as the soluble vascular endothelial growth factor (VEGF) receptor, fms-like tyrosine kinase-1 (sFlt-1), tissue inhibitor of metalloproteinase-3 (TIMP-3) and thrombospondin-1 (TSP-1) [28,66,67] or by an increase in fibroblast growth factor-2 (FGF-2) [68]. Initial in vitro work suggests that OMEC sheets produced in a culture system where 3T3 fibroblast cells are replaced with limbal niche cells as a feeder layer are less likely to induce postsurgical neovascularization [69].

Effect of preparation method on clinical success
We found that OMEC sheet preparation was relatively standardized; most studies used buccal tissue biopsy, DMEM/F12 culture medium, AM as a substrate and air lifting during culture. Several studies compared OMEC culture methods. The two elements that were directly compared were use of AS versus FBS in the culture medium [17] and use of substrate-free culture versus AM as a substrate [27]. Both AS and substrate-free culture have the advantage of minimizing patient exposure to potential contaminants. Clinical results so far suggest comparative or improved corneal epithelial integrity and VA with use of AS and substrate-free culture compared to the use of FBS and AM. However, larger defined comparative studies are necessary before conclusions can be drawn.
Hirayama et al. reported improved success (10/16; 62.5%) in patients receiving substrate-free OMEC sheets compared to those receiving OMEC cultured on AM (6/ 16; 37.5%) ( Table 4) [27]. Improvement in BCVA was also superior in the substrate-free group with 11/16 (68.8%) showing improvement compared to 7/16 (43.8%). Both methods resulted in a stable ocular surface. However, graft survival was significantly improved in the carrier-free group. This may be attributed to direct contact of transplanted OMECs with stromal keratocytes and promotion of proliferation and differentiation of cells in the transplant [70].

Conclusions
OMECs are to date the most common choice of nonlimbal autologous cells in the treatment of LSCD. COMET is a promising treatment modality for LSCD, Variation in methodologies (LSCD diagnosis, cellculture protocols, transplantation technique, postoperative management, and measurement of VA) among the studies did not allow a precise comparative analysis of results. The use of unified tools for characterization of pre-operative status, as well as standardized assessment of outcomes would allow better comparison of studies.