Table 3.

Non-limbal stem cells investigated for the ability to differentiate in the limbal niche.

Cell Source	Summary	Reference
Oral mucosa epithelium	Oral mucosa epithelium transplantation into the limbal area was first reported in 2004. In this work, autologous oral mucosal biopsy samples were obtained, and the submucosal connective tissue was manually removed. The harvested mucosa was divided into smaller sections and the oral mucosa cells were enzymatically separated. Isolated oral mucosa cells were then seeded onto a prepared amniotic membrane with a supporting layer of fibroblasts. After 2–3 weeks, the cultured oral mucosa epithelium on an amniotic membrane were confluent and viable for transplantation. Three eyes afflicted by SJS-induced LSCD and three eyes afflicted by chemical burn-induced LSCD received the prepared oral mucosal epithelium transplants. All cases had improved visual acuity at a mean follow-up time of 13.9 months. Mild peripheral neovascularization was observed in all eyes. Additional clinical trials and studies with similar methodology have been published since this initial report.	[58]
Hair follicle epithelial stem cells	Adult murine hair follicle epithelial stem cells were harvested and cultured in an in vitro environment mimicking the limbal niche. Three- to five-week-old mouse pups were sacrificed and the upper lip pads containing vibrissae were dissected. After enzymatic digestion of the collagen capsule, hair follicles were isolated and underwent further trypsin digestion to isolate individual cells. Hair follicle epithelial stem cells were isolated via FACS. Isolated hair follicle epithelial stem cells were then expanded on a supporting 3T3 fibroblast layer and subsequently introduced to culture conditions mimicking the limbal niche. Limbal-specific extracellular matrix proteins as well as conditioned media from human limbal and corneal fibroblasts were used to emulate the limbal niche. Microscopy, RT-PCR, immunocytochemistry, and western blotting for putative LSC markers confirmed that hair follicle stem cells transdifferentiated into corneal epithelium-like cells under conditions mimicking the limbal niche. A follow-up study by the same group demonstrated an 80% transdifferentiation success rate in an ex vivo mouse model of LSCD. While these methods provide an in vitro concept of transdifferentiation of hair follicle stem cells into corneal epithelium-like cells, this work has not been extended to humans.	[86]
Pluripotent stem cells	A xenogeneic- and supporting feeder cell-free protocol was developed to direct differentiation of human pluripotent stem cells into human LSCs and achieved >65% LSCs in 24 days using either embryonic or induced-pluripotent stem cells. Human pluripotent stem cells were obtained from human embryos and exposed to conditions intended to emulate the limbal niche. In this protocol, Hongisto et al. describe the culture mediums required to induce and differentiate pluripotent stem cells into limbal epithelial stem cells. Furthermore, the authors outline a protocol to cryopreserve and bank the human pluripotent stem cell-derived LSCs, thereby facilitating widespread adoption and dissemination of this technology. This research laid a foundation for subsequent derivation of LSCs from pluripotent stem cells with clear therapeutic implications.	[88]
Dental pulp	Monteiro et al. utilized a chemical-burn rabbit model of LSCD and performed superficial keratectomy 30 days post-injury. Experimental groups then received human immature dental pulp stem cell (hIDPSC) transplants while control groups received amniotic membrane. Immunohistochemistry and RT-PCR analyses showed that hIDPSCs expressed putative LSC markers 3 months after transplantation into the limbal niche. Transplanted hIDPSCs also successfully reconstituted the corneal surface epithelium. To prepare the hIDPSC transplants, the authors first isolated and expanded the stem cells. Three days prior to surgery, the hIDPSCs were lifted and seeded directly onto a temperature-responsive cell culture dish at a density of 2 × 106 per dish. On the day of the surgery, the confluent cell sheets were harvested via a change in temperature and this layer was placed directly on the site of superficial keratectomy and covered with acellular human amniotic membrane. In a similar study, Gomes et al. transplanted a sheet of tissue-engineered hIDPSCs covered by an amniotic membrane into the same rabbit model of LSCD and again demonstrated the ability of dental pulp cells to reconstitute the corneal epithelium when transplanted into the limbal niche. This stem cell source has not been applied in human patients.	[87,89]
Mesenchymal stem cells	In a 2019 clinical trial, Calonge et al. transplanted allogeneic human bone marrow-derived MSCs into the limbal niche and observed MSC transplantation success rate of 76.5–85.7% at 6-month and 12-month follow-ups. They reported no significant difference in the transplantation success rate between allogeneic MSCs and cultivated limbal epithelial cells. This trial demonstrated the viability of MSC transplantation into the limbal niche to treat LSCD.	[91]
Umbilical cord stem cells	In this work, bone marrow was harvested from the iliac crest of allogenic donors. Mesenchymal stem cells were isolated and cultured on human amniotic membrane until 90% confluent. Transplantation involved scraping the corneal-limbal pannus and placing the stem cell amniotic membrane graft cell side down. Transplants were sutured and covered with a bandage contact lens for 4 weeks.	[90]