R estoring the ocular surface in patients with severe ocular surface disease can be difficult to achieve although the last few decades have seen a surge in both the knowledge in this area as well as available treatments. Although the etiologies are diverse, cases of severe bilateral limbal stem cell deficiency (LSCD) with or without conjunctival involvement are particularly challenging to manage given the need to use allogeneic sources for limbal stem cell transplantation. This review will highlight the current practices (for diagnosis, staging, and treatment approaches), evolving techniques, and immunologic challenges in living and deceased donor ocular surface (ie. limbal with or without conjunctival) stem cell transplantation.

The limbal epithelial stem cells (LESCs) are found in the basal layer of the corneoscleral limbal epithelium. These cells function to regenerate the entire corneal epithelium approximately every 7 days. The basement membrane of the corneoscleral limbus is discontinuous and undulating, which forms the palisades of Vogt. Here, the LESCs are housed and provided a protective environment. The exact molecular pathways regulating the function of LESCs are complex, but these cells are clearly influenced by their microenvironment within the palisades of Vogt, which has been termed the limbal stem cell niche. Regulated by mesenchymal stem cells, LESCs divide either symmetrically into two identical cells or asymmetrically into another LESC and a transient amplifying cell (TAC). TACs then migrate centripetally, dividing into postmitotic cells (PMCs) and later into terminally differentiated cells (TDCs), which are eventually shed from the corneal surface.

LSCD occurs when there is an insufficient population of LESCs or when pathology affects the limbal niche. In the absence of the typical progression of LESC to TDC, the conjunctival epithelium is able to grow over the cornea. Therefore, the hallmark of LSCD is corneal conjunctivalization. The International Limbal Stem Cell Deficiency Working Group published the global consensus on the definition, classification, diagnosis, and staging of LSCD in 2019. This paper defined LSCD as an ocular surface disease caused by a decrease in the population and/or function of corneal epithelial stem cells characterized by conjunctivalization and/or other signs of epithelial dysfunction, such as persistent or recurrent epithelial defects with or without neovascularization, ocular surface inflammation, and scarring.

The pathophysiology of LSCD is inherently driven by conjunctivalization of the cornea. Conjunctival epithelium is biochemically and functionally different from corneal epithelium, and these differences manifest as the pathology of LSCD. ^, Conjunctival epithelium contains goblet cells and lacks the tight junctions of corneal epithelium, which results in its hazy, gray appearance. On examination, this allows fluorescein stain to slowly permeate the epithelium, manifesting clinically as late staining (≥10 minutes after instillation of fluorescein). Due to these loose connections and lack of antiangiogenic factors, there is resulting inflammation, epithelial erosions, corneal neovascularization.

Early LSCD may be asymptomatic, but as the disease progresses, patients develop foreign body sensation, conjunctival injection, tearing, blepharospasm, photophobia, and blurry vision. The diagnosis of LSCD can often be made by clinical examination upon observing the late fluorescein staining pattern, especially with other findings such as corneal neovascularization, conjunctivalization, haze/scarring, and persistent/recurrent epithelial defects. The past medical history is helpful for corroborating the diagnosis of LSCD, as patients diagnosed with an etiology known to result in LSCD (ie. aniridia, chemical injury, SJS, etc.) with clinical findings of LSCD often do not require a tissue diagnosis. Atypical cases, such as those without a clear etiology, may warrant other diagnostic modalities to support the diagnosis of LSCD. The histologic hallmark of LSCD is the presence of conjunctival goblet cells in the corneal epithelium, which can be identified with Alcian blue and/or periodic acid-Schiff (PAS). While impression cytology was the gold standard for LSCD diagnosis in the past, its poor sensitivity and association with sampling bias preclude it from being a reliable diagnostic test for LSCD. Impression cytology is performed by pressing nitrocellulose paper onto the anesthetized cornea, which removes some epithelial cells onto the paper. These epithelial cells can then be stained to identify goblet cells. Cell scraping or tissue biopsy may be performed if a deeper or more robust sample is needed. While the presence of goblet cells in the corneal epithelium rules in the diagnosis of LSCD, unfortunately, the absence of goblet cells cannot rule out the diagnosis of LSCD. Besides the sometimes patchy nature of LSCD, there are also certain stages of LSCD, such as end stage keratinization, that may result in complete absence of conjunctival goblet cells. Other tests such as in vivo confocal microscopy (IVCM), anterior segment optical coherence tomography (AS-OCT), epithelial mapping, immunohistochemistry (IHC), and reverse transcription polymerase chain reaction (RT-PCR) may be helpful when the diagnosis is uncertain.

The conjunctiva is important for the overall health of the ocular surface by extensively covering the ocular surface from the upper and lower lid margins to the limbus with mucous membrane. It has important roles in tear film production (aqueous and mucin components), immunology (lymphoid tissue, cytokines), and antimicrobial effects (immunoglobulins) to protect the ocular surface. Found within the conjunctival epithelium, goblet cells are the mucin-secreting glands that are the primary source of mucin MUC5AC, which along with the other mucins function to stabilize, protect, hydrate, and lubricate the ocular surface. ^, Deficiency of the conjunctiva can be brought about by conjunctival inflammatory disorders and/or trauma such as Stevens Johnson Syndrome (SJS), mucous membrane pemphigoid, severe chemical injuries, and severe atopic keratoconjunctivitis. This conjunctival inflammation can result in conjunctival scarring or fibrosis which can lead to forniceal foreshortening and symblepharon/ankyloblepharon; limbal stem cell deficiency; decreased mucin and aqueous tear production; and ultimately ocular surface keratinization. Stem cell transplantation has a worse prognosis in this setting due to the poor tear film, active inflammation, and abundance of immune mediators; therefore, the underlying conjunctival inflammation must be addressed first prior to any further ocular surface rehabilitation.

LSCD is a heterogeneous disease that may result from a variety of etiologies. This paper will focus on bilateral causes, which can be broadly classified into acquired and hereditary. Acquired etiologies of LSCD may be immune-mediated or non–immune-mediated. Acquired immune-mediated causes include Stevens–Johnson syndrome (SJS)/toxic epidermal necrolysis, mucous membrane pemphigoid (MMP), severe allergic ocular surface disease (vernal and atopic keratoconjunctivitis), and graft-vs-host disease. Acquired non–immune-mediated causes include chemical and thermal injuries, contact lens overwear, surgical trauma to the limbus, and medication toxicity. Hereditary causes of LSCD include congenital aniridia, ectrodactyly–ectodermal dysplasia–clefting (EEC) syndrome, autoimmune polyendocrine syndrome (APS), keratitis–ichthyosis–deafness (KID) syndrome, and epidermolysis bullosa, among others.

The Cornea Society’s global consensus staging system of LSCD is the most commonly used. The staging system divides LSCD into three stages based on the extent of corneal and limbal involvement. Stage I disease involves normal corneal epithelium within the central 5 mm zone, with substages based on the percentage of limbal involvement: IA (<50%), IB (≥50% but <100%), and IC (100% limbal involvement with the central cornea still spared). Stage II disease affects the central 5 mm zone of the cornea, with substages IIA (<50% limbal involvement) and IIB (≥50% but <100%). Stage III represents total LSCD, where the entire corneal surface is affected. This staging system was created to guide therapeutic decision-making, yet it overlooks the status of the conjunctiva (ie inflammation and possible resultant deficiency), which is often the most important prognostic factor when planning an ocular surface stem cell transplantation (OSST).

In 2002 (published originally in textbook format, peer-reviewed manuscript in 2010), Holland et al proposed a useful classification system for severe ocular surface disease that incorporated the condition of the conjunctiva alongside the degree of limbal stem cell loss, providing a more comprehensive picture of the ocular surface. ^, In this staging system, patients are first categorized by the extent of limbal stem cell depletion: Stage I if less than 50% of the stem cell population is depleted, and Stage II if more than 50% is lost. Patients are then substaged based on the conjunctival condition: Stage “a” for normal conjunctiva (eg, aniridia, iatrogenic causes), Stage “b” for previously inflamed but currently quiet conjunctiva (eg, remote chemical injury), and Stage “c” for actively inflamed conjunctiva (eg, active SJS or MMP, acute chemical injury). This classification system is particularly valuable because the status of the conjunctiva directly impacts surgical prognosis. Stage IIc patients (those with total LSCD and active conjunctival inflammation) have the worst natural disease course and the poorest prognosis for surgical rehabilitation, as the hostile inflammatory environment of the conjunctiva threatens transplanted tissue with both nonspecific inflammation and immunologic rejection. Patients who transition from Stage I/IIc to I/IIb (ie, conjunctival inflammation has quieted) have improved surgical outcomes, supporting the principle that OSST should be deferred until conjunctival inflammation has been adequately controlled. ^,

There has been a progressive evolution of OSST practices over the last three decades. OSST techniques were first introduced as an autograft treatment in 1964 when Barraquer described taking epithelial limbus (conjunctivocorneal) from the contralateral eye for treatment of a unilateral superficial chemical injury. In 1984, allogeneic OSST started with keratoepithelioplasty by Thoft which transplanted lenticules of peripheral cornea (without limbal cells). There was a better understanding of the location and function of the limbal stem cells in the 1970’s and 1980’s. Turgeon et al (1990) and later Tsai and Tseng in (1994) described modified techniques that included limbal tissue that later evolved into what most OSST surgeons perform as keratolimbal allograft (KLAL). ^,

Due to the vascularity of the limbus and presence of antigen presenting cells, systemic immunosuppression (SI) is necessary for long-term successful allogeneic outcomes. There has been an increased understanding of the needed degree of immunosuppression for KLAL. Early studies initiated 1 or 2 agent SI regimens, often using cyclosporine (CsA) with a tapering systemic corticosteroid. While the first studies tapered the SI relatively quickly with only 6 months of CsA, ^, groups started using long-term CsA. ^, Embracing the principles and SI protocols from solid organ transplantation, Holland et al started introducing triple-agent SI regimens (CsA, azathioprine, and tapering corticosteroid). With advancements in SI medications in regards to increased immunosuppressive effects as well as improved safety profiles, there was a switch over to tacrolimus from cyclosporine and to mycophenolate mofetil (MMF) from azathioprine ( Figure 1 ). ^,, While most recent larger studies have utilized a triple-agent regimen, there are still groups that prefer 1 or 2 agents postoperatively. ^,,

Cincinnati systemic immunosuppression protocol. MMF = mycophenolate mofetil; PRA = panel reactive antibody; TMP-SMX = Trimethoprim/sulfamethoxazole; IV = intravenous; CMV = cytomegalovirus; MWF = Monday-Wednesday-Friday.

Following KLAL, Kenyon and Rapoza described a technique where they transplanted limbal tissue with a conjunctival carrier from a living-related donor which they called “limbal allograft transplantation.” This procedure has become known as living-related conjunctival limbal allograft (lr-CLAL) and transplants allogeneic stem cells along with a variable amount of conjunctiva typically from an ABO- and HLA-matched relative (first degree relative gives the highest chance for a compatible match). Benefits of lr-CLAL over KLAL include decreased rates of rejection due to donor-recipient matching and the ability to provide fresh conjunctival tissue with goblet cells which can treat conjunctival deficiency and a concurrent dry eye. Conversely, less stem cells (maximum 5 clock hours) are transferred during the lr-CLAL as to not risk inducing LSCD in the donor eye compared to the abundant number of stem cells from 1.5 donor corneas in KLAL.

Briefly, the KLAL procedure involves using the corneoscleral rims (often 1.5 corneas from paired mates are used—ie, 2 corneas from the same donor). After the excess sclera is trimmed and the posterior half of the KLAL crescentic tissue is removed by lamellar dissection, 3 crescentic KLAL segments are arranged around the prepared recipient bed (360° peritomy, release of symblephara, removal of corneal fibrovascular pannus) and secured with sutures and tissue glue ( Figure 2 ).

Intraoperative images of performing a keratolimbal allograft: preparing the recipient bed (A), preparing the keratolimbal allograft segments (B), securing the segments (C), and segments in place (D).

For the lr-CLAL procedure, conjunctival limbal grafts (about 2.5 clock hours in limbal circumference) are typically harvested from the 12- and 6-o’clock positions. One must ensure that the total tissue to be harvested is less than half the limbal circumference (ideally 5 clock hours or less). The conjunctiva is elevated from Tenons/sclera and dissected from about 5 to 7 mm posterior to the limbus superiorly (and 3-5 mm inferiorly) to the limbus. Dissection is carried anteriorly about 1 mm onto the cornea past the limbal vessels to ensure inclusion of the limbal stem cells. These are transferred into balance salt solution while the recipient eye is prepared (similar to KLAL above). The lr-CLAL grafts are then brought onto the field and secured at the 12- and 6-o’clock positions with sutures and tissue glue ( Figure 3 ).

Intraoperative images of performing a living-related conjunctival limbal allograft (lr-CLAL): harvesting the donor segments (A), preparing the recipient bed (B), securing the lr-CLAL segments (C), and segments in place (D).

Based on the work from the solid organ transplantation literature, the Cincinnati group developed a protocol for screening potential recipients for OSST as well as standardized donor/recipient matching to identify the best available donor for each recipient to maximize successful outcomes ( Figure 4 ). The most appropriate OSST procedure is determined based on several factors such as disease etiology, laterality, conjunctival involvement, available family donors, and general patient health. In the appropriate setting of unilateral acquired LSCD, conjunctival limbal autograft (CLAU) is the treatment of choice following careful screening of the unaffected eye (eg, occult late staining, prior contact lens use, significant ocular surgeries). For bilateral causes of LSCD, OSST procedures such as lr-CLAL, KLAL, or a combination procedure are considered with subsequent keratoplasty if needed for stromal scarring.

Cincinnati protocol for donor/recipient matching and donor selection.

Certain LSCD etiologies predominantly affect the limbus with minimal to no conjunctival involvement (eg, aniridia, contact lens wear), while other conditions such as mucous membrane pemphigoid (MMP), Stevens-Johnson Syndrome (SJS), and severe chemical injuries can result in combined pathology LSCD and conjunctival inflammation/deficiency. ^, When lr-CLAL/CLAU donors are unavailable, KLAL may be suitable for diseases such as aniridia, contact lens wear-related LSCD, and iatrogenic LSCD when there is minimal to no conjunctival involvement. KLAL allows transplantation of a large number of stem cells but without healthy conjunctiva while lr-CLAL and CLAU provide both fresh limbal stem cells and conjunctiva. Severe ocular surface cases with conjunctival involvement may decrease KLAL success secondary to chronic conjunctival inflammation and scarring; decreased mucin and aqueous tear deficiency; and increased potential for keratinization. Here, prior treatment with mucous membrane grafting can reform the fornix to make space for the OSST and provide additional mucous producing cells to treat the associated dry eye. ^, Besides longer survival rates and decreased rejection rates, lr-CLAL tissue is ideal for these situations with conjunctival deficiency as the transplanted conjunctiva can help treat mild to moderate symblepharon and provide goblet cells. Additionally, glaucoma, lid malposition/exposure, and ocular surface inflammation should be adequately addressed prior to an OSST.

Patients who are candidates for OSST surgery must be willing to adhere with long-term regular follow-up. Nonadherence with the immunosuppression protocol has been shown to increase the risk of immune rejection and surface failure. Potential OSST candidates must be relatively healthy and be able to tolerate the SI medications for this elective procedure. Relative contraindications for SI include a history of malignancy <5 years, nonadherence with clinical/laboratory follow-up or nonadherence to medications, and significant comorbidities (ie. uncontrolled hypertension, uncontrolled diabetes, congestive heart failure, other organ failure, and advanced age).

To determine the best donor for eyes requiring an allogeneic OSST, patients are asked to identify potential first-degree donors. Siblings have the potential to be an HLA-identical match, whereas parents and children often are at least HLA-haplo-identical (50% identical). Laboratory testing is used to determine the compatibility between potential donor-recipient pairs. First, ABO-typing is performed; if incompatible, another donor is considered (when available) or KLAL may be necessary. If they are ABO-compatible, then HLA typing, donor specific antibody (DSA), and virtual crossmatch testing are performed; the recipient’s Panel Reactive Antibody (PRA) level is also checked for preoperative planning and prognostic value.

HLA antigens are genetically determined molecules (typically proteins) found on the cell surface that can stimulate the production of antibodies. With transplanted tissue/organs, a recipient’s body recognizes the HLA antigens of the transplant as foreign and forms specific antibodies against those particular HLA antigens. If a recipient has developed antibodies against a specific HLA antigen from prior sensitization (eg, pregnancy, blood transfusion, prior transplant, etc.), these specific antigens (eg, A1, B5) are deemed unacceptable antigens (or mismatches). For a specific recipient, unacceptable antigens are identified by the lab as antibodies detected above a particular mean fluorescence intensity (MFI) threshold (eg, 1500 MFI). These unacceptable antigens are used to determine the calculated PRA (cPRA) using the United Network of Organ Sharing (UNOS) cPRA calculator.

As they are the most pertinent for HLA matching, only broad antigen (class I: HLA-A, HLA-B; class II: HLA-DR, HLA-DQ) mismatches are considered. For each HLA class, the least number of mismatches, ideally zero to two, for a donor-recipient pair identifies the closest match. Through a virtual crossmatch for class I and class II HLA, the laboratory assesses the immunologic compatibility based on a recipient’s alloantibody profile compared to a donor’s HLA antigen typing to predict the results of a physical crossmatch (ie, serologic/specimen testing). Alternatively, a direct crossmatch can be performed with samples from the donor and recipient. A negative result for each HLA class (ie, negative/negative) identifies a compatible donor-recipient pair. DSAs are recipient anti-HLA antibodies specifically generated against donor cells.

PRA serves as a measure of sensitization that is derived from determining whether or not a patient has any specific preexisting HLA antibodies. By testing the patient’s blood against lymphocytes obtained from a 100-donor panel, PRA is reported as a percentage (0%-100%) and can be used as a screening test for a range of known unacceptable antigens in the population. The percentage PRA is calculated based on antigen frequency regardless of donor-directed antibodies. Those recipients with higher PRA levels would be expected to potentially react to or reject a greater proportion of the population, whereas recipients with a low PRA level are considered low-risk. If a recipient has a PRA ≥50%, they are considered high risk and will undergo augmented induction therapy (ie. intravenous basiliximab). Basiliximab is an immunosuppressant monoclonal antibody that blocks the IL-2 receptor on activated T-lymphocytes; in renal transplantation, it is used to prevent acute kidney transplant rejection. PRA can help stratify donor-recipient compatibility risk which aids in identifying the best match while also eliminating unneeded HLA typing. If there is a recipient PRA of 0% without any unacceptable antigens and an ABO-compatible donor, full HLA typing and virtual crossmatch testing are not necessary. If a specific unacceptable antigen (ie, DSA) was uncovered in a recipient with 0% PRA due to low frequency antigen, only HLA typing for that specific HLA locus is necessary. With no DSAs, a PRA of 0%, and an ABO-compatible parent or child (at least HLA-haplo-identical), a full HLA-typing is not always performed by the laboratory.

Following identification of the best-matched available living-related donor, serologic testing for HIV I/II, hepatitis A/B/C viruses, and cytomegalovirus is conducted on all potential donors. The living-relative donor eyes are screened for pertinent ocular history (eg. prior long-term contact lens wear or prior eye surgery) and are evaluated preoperatively with a full ophthalmologic examination to assess for occult ocular disease, specifically looking for late staining. With proper screening and conservative harvest of the limbus (≤5 clock hours), there is low risk to donor eyes for inducing LSCD. ^, In addition to the same viral screening tests, a complete blood count, comprehensive metabolic panel, and lipid profile are checked for all recipients prior to starting SI.

If a lr-CLAL donor is not identified and KLAL is planned, KLAL quality tissue is requested from the eye bank and donor specifications should be communicated. The following donor tissue selection guidelines should be considered to provide high-quality donor KLAL tissue: ^,

• 1.

  Donors of younger age, preferably with upper cut-off at 60 years

• 2.

  Minimal time from donor death to tissue preservation

• 3.

  Implantation of KLAL tissue 9 days from death of donor

• 4.

  Maintaining conjunctiva (>3-4 mm skirt) and more sclera (4-5 mm) allows for transplantation of goblet cells and minimization of stem cells damage

• 5.

  Corneoscleral rims from both eyes of the same donor limits antigenic exposure to recipient

• 6.

  No active infection and no prior ventilator exposure in the donor

• 7.

  No prior history of melanoma or any metastatic cancer in the donor

• 8.

  Donor-recipient ABO match if possible

A recent retrospective review studied KLAL epithelialization outcomes based on donor tissue parameters. They found no specific donor tissue factors that affected total time to epithelialization; specifically, when looking at donor tissue preservation time, it appears that limbal stem cells can be safely stored for up to 9 days. Donor death-to-preservation time ranged from 3 to 24 hours in this study, with maximum death-to-cooling time at 14 hours. While the mean donor age was 50.4 years, the range encompassed 4 to 77 years.

Successful allogeneic OSST is centered on adopting the principles of solid organ transplantation and embracing SI. This requires a team approach, including an organ transplant specialist and multiple ophthalmic subspecialists. Based on solid organ transplantation protocols, 3-agent systemic immunosuppression protocols (eg, tacrolimus, mycophenolate mofetil, and tapering prednisone) can provide the highest rates of long-term allograft survival. ^,,,,,,,