Recent breakthroughs in regenerative medicine, in particular cell therapies based on autologous cultures of somatic stem cells, and the first pluripotent stem cell clinical trial recently proposed (NATURE NEWS BLOG Embryo-like stem cells enter first human trial. 14 Feb 2013; Clinical trials using induced pluripotent stem cells by next year TAMIL NADU, Updated Feb 03, 2012) , have generated enthusiasm and many efforts to explore new therapeutic potentials of both somatic and pluripotent stem cells [2, 3]. In recent decades, developmental biology elucidated many cellular and molecular mechanisms regulating stem cell dependent tissue homeostasis, informing on the molecular basis of many diseases.
Many human tissues and organs possess the ability to self renew and repair acute or chronic lesions. These processes rely on specific stem cells, which generate progenitors (often referred to as transient amplifying [TA] cells), which then generate terminally differentiated cells. Examples are provided by hematopoietic stem cells, which give rise to all blood cells through myeloid and lymphoid progenitors [2, 4], epidermal stem cells, which give rise to epidermis, hair follicle, and sebaceous glands [5, 6], and bone marrow-derived mesenchymal stem cells, able to generate all tissues found within a skeletal segment (bone, cartilage, adipocytes, fibroblasts, and hematopoiesis supporting stroma), hence better defined as skeletal stem cells [7, 8].
Lining epithelia play a crucial role in the homeostasis of the entire organism as they cover external and internal surfaces of the human body. Altogether, squamous epithelia represent approximately 20% of the body weight. They build an efficient and fail-safe barrier, which is able to separate the body from the external environment and maintain the internal milieu, preserving it from the constant attacks of microorganisms.
Squamous epithelia provide an ideal experimental system for studying adult stem cells. They are constantly renewed through the periodic proliferation of keratinocyte stem cells and an intricate balance between cell growth and differentiation . Rheinwald and Green  pioneering work made possible the cultivation of human keratinocytes using a feeder layer of lethally irradiated 3T3-J2 cells, introducing a somewhat irreplaceable condition for epithelial stem cell-mediated cell therapy. Cultures of autologous keratinocytes have been used to prepare grafts  that can permanently restore severe epidermal defects, such as life threatening massive full thickness burns, and can be used for gene therapy of genetic skin disorders [11–15]. Approximately 15 years ago, autologous cultures of limbal stem cells, which are the stem cells of the corneal epithelium, have been shown to fully restore a severely damaged corneal epithelium and to restore vision in patients with chemical burn dependent corneal destruction [16–20].
Stem cells were removed by enzymatic treatment, from a 1–2 mm2 biopsy of limbus, the narrow zone between conjunctiva and cornea. Limbus is the only corneal area with papillae-like invaginations termed palisades of Vogt and limbal epithelial crypts , containing very small basal cells lacking corneal-specific keratin 3 (K3). These limbal basal cells contain slow cycling cells  and holoclone-forming cells  whereas the corneal epithelium does not. Many markers were proposed to identify corneal stem cells; however, only few of them were proved to be associated with long-term corneal regeneration [23–26].
Since the initial report of successful clinical application of cultured limbal stem cells, several tens of similar protocols and clinical applications were proposed. Investigations on alternative methods and eligible pathologies have contributed to increase our knowledge in this field, while raising questions related to identification of causes of variability related to reagents for the manufacturing process, patient's selection, drugs, surgical, and postoperative management and their implication on success rate, safety, and reproducibility of clinical outcomes.
The cultural separation between different scientific fields makes it difficult to establish the multidisciplinary criteria that are necessary for a successful translational research. In this review, we try to address some of those criteria (specifically related to both clinical and biological parameters) in relation to translational medicine by means of cultured limbal stem cells.
The limbal stem cell pool of the corneal epithelium decreases as a result of hereditary or acquired injury leading to partial or total limbal stem cell deficiency (LSCD). This definition refers to a heterogeneous group of pathologies having deleterious effects on corneal integrity and wound healing . As a consequence of stem cell depletion, an invasion of peripheral and central cornea by conjunctival epithelium may occur (Fig. 1). Such tissue delocalization induces neovascularization of a normally avascular ocular area and corneal opacification, leading to decrease of visual acuity. Conjunctival invasion can be prevented by limbal restoration through transplantation of grafts generated by cultivation of autologous limbal cells taken from the uninjured eye [17–20, 28].
The proper selection and preparation of the receiving bed is of paramount importance for a positive clinical outcome of limbal cultures. Failures can indeed be due to the severity of the ocular damage and severity of the tear film impairment, level of inflammation, and postoperative complications. Chemical burns might damage the eyelids, the conjunctiva, and the lachrymal system. During such deterioration, the ocular surface is chronically inflamed and the resulting alteration of the microenvironment might hamper the engraftment of cultured stem cells. In the case of large damage of the entire ocular surface, conjunctiva reconstruction that allows eye bulb movements and physiological tear film distribution should be obtained before cell transplants. In the absence of negative systemic or genetic stimuli, the transplantation of autologous cultures of limbal cells can itself partially or totally restore the macroenvironment or microenvironment. This might explain why, in severely damaged chemical burn-dependent ocular surfaces, the second transplant of cultured epithelium gives a higher success rate than the first . The first transplantation of cultured cells might “normalize the environment,” through production of laminin 332, proteoglycans, and collagen (all of which might restore the extracellular matrix), autocrine and paracrine secretion of a balanced amount of growth factors, such as TGFalpha, ILs, PDGF, IGF1, TGFbeta(s), NGF, basic fibroblast growth factor [29–35], and epithelial-mesenchymal crosstalk. Despite many years of successful clinical application, we still have no sense for the number of stem cells that can engraft onto the wound bed.
Furthermore, ocular surface diseases attributed to LSCD might have a variety of aetiologies, not all of which were proven to be associated to a total depletion of limbal stem cells. Indeed, they might be characterized by a macroenvironment or microenvironment not permitting proper proliferation and differentiation of existing stem cells. The variety of LSCD might thus provide transplanted cells with “behavior instructions” driving different cell responses during or after engraftment.
For instance, aniridia (hereditary LSCD) is characterized by mutations in the PAX6 gene . PAX6 is a highly evolutionary conserved transcription factor controlling the morphogenesis of the entire eye. Missense mutations in the DNA binding domains of PAX6 might hamper recognition of binding sites in target genes, resulting in partial or complete loss of protein function . Immune diseases, such as Stevens Johnson syndrome (SJS), are characterized by cell apoptosis and necrosis resulting in epithelial detachment. SJS is due, mainly but not exclusively, to adverse drug reactions. Drug response is a multifactorial and multigenic process, dependent on a complex interaction between multiple proteins and the environment. The wound bed of patients can be deeply altered by inflammation, cell debris, tear film abnormalities, and abnormal synthesis of enzymes and extracellular matrix proteins. In addition, it has been suggested that early appearance of CD14+ CD16+ cells of monocyte lineage plays an important role in epithelial damage associated to SJS/TEN, most probably by enhancing the cytotoxicity of CD8+ T cells . Other pathologies of the ocular epithelium, as cicatricial pemphygoid, are of clear autoimmune origin. In these cases, any transplant of cultured autologous cells will suffer from autoimmune reaction, interfering with ocular surface restoration.
Table 1 summarizes main conditions that determine a LSCD. Most of them are known to produce to some extent, alterations in ocular surface environment such as dry eye (2007 Report of the International Dry Eye workshop, DEWS).
|Limbal stem cell deficiency|
|Non immune-mediated||Chemical injury, thermal injury, radiation, contact lens misuse, surgical interventions, drug-induced, trachoma, limbal tumors|
|Immuno-mediated||Steven-Johnson and Lyell syndromes, mucous membrane pemphigoid?, peripheral ulcerative keratitis?, Allergic keratitis, graft versus host|
|Hereditary||Aniridia, dyskeratosis congenita, multiple endocrine deficiency, xeroderma pigmentosum?, ectodermal dysplasia?|
A precise diagnosis and grading of LSCD are thus needed in order to choose whether an appropriate limbal stem cell therapy can be performed and draw meaningful conclusions on its efficacy. Altogether, data generated during nearly 3 decades of clinical application of epidermal and limbal cultures showed that a specific protocol cannot be used in different pathologies, but should be tailored on the specific pathology and type of wound bed onto which the culture has to be transplanted [39, 40]. Currently, autologous cultures of limbal stem cells have been proven to be truly successful in chemical/thermal burn dependent corneal destruction characterized by a complete LSCD. In this clinical setting, the mechanism of action of the cultures is the long-term replacement of lost limbal stem cells, as formally proven by several authors [20, 41]. During the corneal repair process, the transplanted stem cells multiply, migrate, and differentiate to regenerate the corneal epithelium. The engrafted stem cells must relocalize in the limbal niche and maintain their self-renewal capacity, as formally proven by their ability to regenerate a second time a normal corneal epithelium after the keratoplasty performed 12–24 months after grafting to remove the stromal scarring and to restore a full visual acuity.
The proliferative compartment of human squamous epithelia contains three types of clonogenic keratinocytes [16, 42], referred to as holoclones, meroclones, and paraclones  (Fig. 2). The holoclone-forming cell is the stem cell of all squamous epithelia . Cultured epithelial sheets containing an appropriate number of holoclones permanently restore [13, 19] massive epithelial defects, and genetically modified epidermal holoclones generate a transgenic epithelium able to permanently restore a normal, fully functional epidermis in patients with genetic skin adhesion disorders . The holoclone produces meroclone and paraclone-forming cells, which have properties expected of TA progenitors. Clinical data obtained from eye suffering for chemical burns provided conclusive evidence that the most important (if not the only) biological criterion to assess graft quality and the likelihood of a successful outcome is a rather precise evaluation of the number of stem cells, detected as p63 bright holoclones, contained in the culture . The number of clonogenic cells, colony size, and cell growth rate are conditions necessary but not sufficient to predict the performance of the graft. Successful and unsuccessful cultures contain a virtually identical number of clonogenic cells, confirming that the vast majority (approximately 95%) of clonogenic keratinocytes (meroclones and paraclones) possess limited proliferative/regenerative potential and behave as TA progenitors .
Furthermore, several methods have been proposed for the cultivation of limbal stem cells (Fig. 3). The composition of culture medium plays an important role in preserving limbal stem cells, and several materials/reagents or mixtures of them have been proposed. However, applying current good manufacturing practice (GMPs) to biological reagents destined to manufacture living biological drugs is not straightforward. Notwithstanding, details on criteria used for selection of suitable materials/reagents were not consistently provided and it was never reported how they were considered suitable to support limbal stem cells. In the absence of these data, it is difficult to evaluate success rate or safety of each culture system developed and whether differences were related to the culture technique or to single reagent batch selection. Obviously, comparison of different culture conditions cannot highlight differences when stem cells are not maintained in all test items.
In the past years, a number of studies have proposed different combination of growth factors and hormones within the culture medium. Those media might have induced stimulation of different cellular pathways with a variety of cellular responses either in culture or after transplantation; for instance, some reports use only insulin and epidermal growth factor , whereas some others add also triiodothyronin, hydrocortisone, and cholera toxin [14, 20], all of which stimulate different signaling pathways and might affect the balance between cell proliferation and differentiation. Finally, some authors suggested the use of airlifted cultures, in order to increase differentiation in vitro [45, 46], but airlifting could induce cell differentiation at the expense of stem cell self-renewals, resulting in stem cell loss [45, 47]. To be successful, a limbal culture must contain a sufficient number of the keratinocyte stem cells essential for long-term corneal renewal rather than a well-organized stratified epithelium . Clinical outcomes obtained with the different techniques might contribute to identify key information on appropriateness of the cultivation process.
The Issue of Xenogenic Components
Fetal calf serum (FCS) is required for limbal stem cell preservation. In an effort to prevent putative xenogenic contaminants, human autologous serum has been proposed as a potential substitute for FCS . However, variability of hormones and growth factors content due to individual genetic background could be detrimental for the reproducibility of the culture. It might have an impact on stem cell preservation, reduce the reliability of the in process controls, and hamper the definition of well-defined specific quality criteria for the culture. FCS is usually prepared from pools of sera, thus minimizing such variability. Obviously, FCS must be extensively tested for absence of pathogens in compliance with directives of Heath Authorities. The use of pools of human sera might circumvent individual variability. However, the risk of contamination by viral-, nonviral infectious agents, and prions also exists when using human derived reagents , particularly in the absence of interspecies barriers. Alternatively, synthetic media have been proposed, which, however, are [50, 51] also depleted from a huge number of physiological cell stimulators, part of them are well known and may be added exogenously. Conversely, most of them are still unknown and purification (under GMP conditions) of all hormones and growth factors needed would produce an unaffordable cost for culture media.
Preservation of limbal stem cells also requires an appropriate feeder layer of lethally irradiated fibroblasts . Feeder cells produce cytokines, growth factors, and extracellular matrix proteins, and provide optimal biomechanical properties for maintaining the balance between cell growth and differentiation. For instance, epithelial cells cultured on a feeder layer perceive a softer substrate than those on plastic, which might impact on mechano-transduction pathways. A cell on cell layering influences differentiation processes, as shown in myoblasts, where expression of striation of myotubes was seen only in the upper layered cells  (Fig. 4).
Lethally irradiated mouse 3T3 fibroblasts have been used for 3 decades to cultivate human limbal and epidermal keratinocytes [11, 13, 14, 28], and no adverse reactions were reported by [39, 53] groups performing large case series (reviewed in). As with autologous serum, multiple sequential transplantations of epidermal and limbal cultures proved the absence of an immune response against residual 3T3 “ghosts” present in the culture. To date, preservation of holoclone-forming cells has been demonstrated only in the presence of a feeder layer made of 3T3 cells (particularly the J2 clone) [13, 14]. Yet the use of a feeder layer made of autologous or allogeneic human fibroblasts has [54–56] been proposed, on the assumption of their safer profile. Concerning the potential transmission of pathogens, the same concepts expressed for FCS versus donor serum apply to murine cells versus allogeneic human fibroblasts.
A feeder-layer made of autologous fibroblasts suffers from the same intrinsic variability of autologous serum, due to individual genetic background. Furthermore, aging might have important additional effects on fibroblasts. Aged fibroblasts have a reduced ability to [57, 58] express matrix proteins or growth factors as compared to fibroblasts obtained from young patients and over express metalloproteinase, which may explain the age-related atrophy of extracellular matrix architecture . Thus, the confusion introduced by these criteria has major consequences when translational implications are considered. The overall quality of the culture cannot be predicted based on variable, hence unreliable, in vitro criteria.
The use of a clinical grade, GMP certified cell line, tested for known viruses, bacteria, yeasts, and fungi, and even unknown microbial/viral components (e.g., using Electron Microscopy and reverse transcriptase analysis) appears safer than proposed alternatives. Such a GMP produced Master Cell Bank made from a clone of well-characterized 3T3 cells not only would accomplish strict safety requirements but also would provide the manufacture process with precise, highly reproducible quality controls, product specifications, and release parameters, together with long-term clinical testing.
It is known that microchimerism, cell fusion, and tumor formation all require viable cells, which are able to replicate. Lethal irradiation acts as a preventative measure to ensure the cells pose no health risk in terms of cell proliferation, although, to our knowledge, cases of a human tumor originating from murine cells have not been reported. The procedure for lethal irradiation should also be GMP validated. Indeed, the case of microchimerism reported in the literature occurred after application of epidermal autografts cultured on 3T3 cells not irreversibly growth arrested . As a matter of fact, in most papers, it is not specified whether 3T3 came from a validated master cell bank.
Patterning of Surfaces and Biomechanical Signals
A typical strategy in therapeutics based on stem cells consists in engineered tissues made of cells cultured on suitable biomaterials to mimic the in vivo biochemical and biophysical microenvironment. Many different materials were proposed for cell culturing, such as fibrin, amniotic membrane, plastic, or polymers [17, 20, 60, 61]. The interplay between stem cells, the surrounding microenvironment, and external forces (representing the niche) needs to be elucidated before stem cell-based therapies are applied in clinics. Cell behavior is not only governed by pure chemical signals . Tissue architecture and mechanical forces are overarching signals that govern cell decisions: cell matrix and cell–cell adhesion, organization of the cytoskeleton, and tensional forces that keep individual cells and the whole tissue in a certain shape, all represent architectural signals [62–64]. Mechanical connections between the matrix and the cytoskeleton allow cells to exert traction forces that are transmitted to the cell nucleus and the resulting force triggers signals, which modulate adhesion, spreading, migration, proliferation, and differentiation. More generally, mechano-transduction envisages a combination of signals, including interfacial presentation of molecular, topographic, and mechanical cues . To study those interactions, specifically engineered surfaces displaying selected biofunctional groups or micrometer scale patterns have been used [65, 66]. They highlighted that substrate stiffness, surface nanotopography, microgeometry, extracellular forces, and even shear flow can all have significant influences on regulating activities of several cell types [63, 64]. For instance, embryonic stem cells increase their spreading as a function of substrate stiffness and, on softer substrates, become able to self renew even in the absence of leukemia inhibitory factor (LIF) ; in vivo, formation of cutaneous scars triggers TGFβ/Smad, integrin, and calcium signaling ; mitogen-activated protein kinase (MAPK) is involved in activation of responses to shear sensitive genes, together regulating vascular function and disease. When laid on wide lanes of soft hydrogel, myoblasts decrease their proliferation rate and increase their fusion index [52, 66]. In mammalian hair cells, mechanical stimuli cause stereocilia to move, triggering the mechano-transduction channel to open and, thus, activating NO, Ca2+, and Maps signals. Recently, cues from different cell sites were found to converge on the regulation of the downstream effectors of YAP and TAZ pathways [62, 65].
Drugs in Surgery and Follow-Up
Managing post-traumatic or postoperative problems relies on the extensive use of drugs. In the case of LSCD treatment, the acute corneal cytotoxicity and the chronic toxicity of eye drop components are not adequately considered for their impact on regenerative medicine protocols. However, a large number of experimental and clinical studies have shown that long-term use of topical drugs may induce ocular surface changes, causing ocular discomfort, tear film instability, loss of goblet cells, inflammation, conjunctival squamous metaplasia, epithelial apoptosis, and subconjunctival fibrosis. Prolonged use of local anesthetics is associated with delay of corneal re-epithelialization after wounding , altered lubrication and tear film , corneal swelling, and disruption of epithelial motility [70, 71]. Lidocain, one of the most commonly used anesthetics, already when used at concentrations of 250 mg/ml (i.e., below the clinical dosage), impairs normal wound healing . Vasoconstrictors, which are generally used to increase of duration of local anesthesia, can produce a cytotoxic effect and pigment deposition, as reported for epinephrine and some commonly used antiglaucoma drugs .
Depending on dosage and mode of administration, ophthalmic corticosteroids that are widely used, both systemically and topically, in the treatment of ocular inflammatory conditions, have side effects including impaired corneal wound healing. Nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit prostaglandin synthesis and exert an equivalent anti-inflammatory effect. They are increasingly used because of their concomitant analgesic effect, and their effectiveness on allergic ocular disorders. However, comparative analysis of corneal toxicity showed cell damage, altered cell viability, proliferation, and migration after short-term exposure of human corneal cells to some NSAIDs . Other commonly used drugs, such as prostaglandin analogs, antiallergic eye drops, or multipurpose solutions (particularly in presence of preservatives) were reported to have similar toxic effects . In particular, data have suggested an irreversible damage of the epithelial proliferative compartment, including stem cells; effects on cell viability (up to 40% decrease) and proliferation, corneal thickness, endothelium permeability, DNA integrity, and eye lubrication due to common preservatives, such as benzalkonium chloride, were described, especially in case of repeated administration or persistence of such drugs on damaged ocular surface [76–78]. It is clear that all these toxic effects are amplified when drugs are used before the full engraftment of a pure population of cultured cells. In conclusion, no drugs appear to be free of toxicity to some extent, therefore a careful selection of drugs on cell cultures and harmonization of dosage and mode of administration should be considered to increase the benefit/risk ratio.