Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Comparative Study of Limbal Stem Cell Deficiency Diagnosis Methods: Detection of MUC5AC mRNA and Goblet Cells in Corneal Epithelium Iker Garcia, DSC,1 Jaime Etxebarria, MD,2 Ana Boto-de-Los-Bueis, MD,3 David Díaz-Valle, PhD, MD,4 Luis Rivas, PhD,5 Itziar Martínez-Soroa, MD,6 Nerea Saenz, MD,7 Carlos López, MD,8 Almudena Del-Hierro-Zarzuelo, MD,3 Rosa Méndez, MD,4 Javier Soria, PhD,1 Nerea González, DSC,1 Tatiana Suárez, PhD,1 Arantxa Acera, PhD1 Purpose: To evaluate a limbal stem cell deficiency (LSCD) diagnosis method based on the detection of the MUC5AC transcript by reverse transcription-polymerase chain reaction (RT-PCR) in comparison with the standard diagnostic method based on goblet cell detection by periodic acid-Schiff (PAS)– hematoxylin staining, using samples obtained from corneal epithelium impression cytology (IC). Design: Transversal, comparative case series. Participants: We studied 59 eyes from 43 patients clinically diagnosed with LSCD. Methods: Impression cytology was used to gather cells from corneal and conjunctival epithelium from the same eye. The presence of goblet cells in the cornea was determined by PAS-hematoxylin staining, whereas the presence of the MUC5AC transcript was detected by RT-PCR using a custom-designed primer pair. Main Outcome Measures: Goblet cells in the corneal epithelium were detected by light microscopy, and the MUC5AC transcript was detected as the corresponding PCR amplicon in agarose gels. Results: Our study included 59 corneal samples, together with their respective conjunctival samples for RT-PCR assays. Of these, 47 samples were also available for comparative PAS-hematoxylin staining. The MUC5AC amplicon was detected in 56 of 59 (94.9%) corneal epithelium samples. In contrast, conventional IC staining detected goblet cells in only 17 of 47 (36.2%) samples; these were not found in 27 of 47 (57.4%) samples (negative results), and 3 of 47 (6.4%) showed inconclusive results. Conclusions: The detection of the MUC5AC transcript in corneal epithelium is a more sensitive method to diagnose LSCD than the conventional PAS-hematoxylin method, although a minimum RNA concentration of 1.2 ng/l is required for negative results to be reliable. Moreover, RT-PCR is a highly specific and more objective technique. Overall, these findings indicate that molecular analysis facilitates a more precise clinical diagnosis of LSCD, thereby reducing the risk of surgical failure. Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references. Ophthalmology 2012;xx:xxx © 2012 by the American Academy of Ophthalmology. Limbal stem cells reside in the sclerocorneal limbus at the level of the palisades of Vogt. It is believed that these cells are responsible for the regenerative function that permits the maintenance of the corneal epithelium and for the so-called barrier function that prevents the migration of conjunctival cells over the cornea.1 Limbal stem cell deficiency (LSCD) involves the loss of these functions, which may be a consequence of the direct destruction of this cell population or its stromal microenvironment, or in the majority of cases, due to exogenous factors that destroy the limbal stem cells, including chemical or thermal burns, ultraviolet or ionizing radiation, Stevens–Johnson syndrome, ocular cicatricial pemphigoid, surgery, cryotherapy, contact lens use, or microbial infection.2 The partial or total destruction of the © 2012 by the American Academy of Ophthalmology Published by Elsevier Inc. limbal epithelium leads to abnormal wound healing of the corneal epithelium. This gives rise to a series of clinical signs, including the invasion of the conjunctival epithelium in the area of the cornea, vascularization, and chronic inflammation of the corneal stroma.3–5 Corneal conjunctivalization involves the loss of limbal epithelium, which normally acts as a barrier between the corneal and the conjunctival epithelia. Limbal stem cell deficiency is characterized at a histopathologic level by the presence of goblet cells on the cornea, vascularization, destruction of the corneal basal membrane, and chronic inflammation.2 The conjunctiva is responsible for the production of a variety of mucins, which are essential for tear stability and ISSN 0161-6420/12/$–see front matter doi:10.1016/j.ophtha.2011.10.031 1 Ophthalmology Volume xx, Number x, Month 2012 corneal transparency. Epithelial mucins are a heterogeneous group of large proteins (⬎200 kDa) and are components of all the mucous secretions present in the epithelia. They are highly glycosylated, so much so that a high percentage (⬎50%) of their molecular mass is made up of sugars, thus hindering their biochemical analysis.6 Epithelial mucins can be classified as being transmembrane or secreted. There are currently 11 known types of transmembrane mucins (MUC1, MUC3A, MUC3B, MUC4, MUC12, MUC13, MUC15, MUC16, MUC17, MUC20, and MUC21) and 7 types of secreted mucins, which can in turn be subdivided into soluble (MUC7 and MUC9) and gel-forming (MUC2, MUC5AC, MUC5B, MUC6, and MUC19) molecules.6 –13 Goblet cells are intercalated between non-secretory epithelial cells in the conjunctival epithelium. Their function is to synthesize and secrete the gel-forming MUC5AC mucin into the tear film.13 Thus, the MUC5AC mucin that is only found in these cells has been postulated to be a specific marker of goblet cells. At present, the most popular technique for LSCD diagnosis is impression cytology (IC) followed by periodic acid-Schiff (PAS)-hematoxylin staining because of the simplicity of the method, which is noninvasive and permits an analysis of the morphology and status of the ocular surface.2 However, this technique, which is useful for evaluating the anatomopathologic status of the conjunctival epithelium, is less sensitive for determining corneal conjunctivalization by means of evaluating the presence of goblet cells in the corneal epithelium. Another technique for detecting LSCD is the use of immunohistochemistry with monoclonal antibodies to evaluate the presence of CK19 in epithelial conjunctival cells14 –16 and the presence of mucin 5AC in goblet cells in the cornea. A number of studies have reported the detection of the MUC5AC transcript in the conjunctival epithelium in healthy individuals with the use of reverse transcriptionpolymerase chain reaction (RT-PCR),17,18 in different pathologies such as dry eye, and in contact lens users.19 However, none of these studies used the detection of the MUC5AC transcript in the cornea as a diagnostic tool or evaluated the efficiency and reproducibility of this molecular technique in comparison with conventional IC followed by PAS-hematoxylin staining. The objective of the present study was to compare the efficacy of 2 techniques to detect the presence of goblet cells in the cornea of patients with clinical signs of LSCD: IC coupled with PAS-hematoxylin staining and the specific detection of MUC5AC mRNA by RT-PCR. Materials and Methods Patients A transversal study was carried out in 6 hospitals pertaining to the Spanish National Health Service: Hospital de Cruces (Baracaldo, Vizcaya), Hospital de Donostia (San Sebastian, Guipúzcoa), Hospital La Paz (Madrid), Hospital Clínico San Carlos (Madrid), Hospital 12 de Octubre (Madrid), and Hospital de GaldakaoUsansolo (Galdakao, Vizcaya). A total of 43 patients diagnosed with LSCD by means of slit-lamp examination were enrolled 2 between February 2009 and February 2010. Impression cytology samples were obtained after patients had signed informed consent, following the Principles of the Declaration of Helsinki on Biomedical Research Involving Human Subjects. Independent IC samples were obtained on different days to comparatively analyze the 2 detection techniques by using cellulose acetate discs (HAWP304, Millipore, Bedford, MA). Samples were always obtained in the following order: first from the cornea and next from the conjunctiva, having applied topical anesthesia to the ocular surface using a mixture of oxybuprocaine chlorhydrate, tetracaine hydrochloride, and chlorobutanol (Colircusi double anesthetic, Alcon Cusí, Barcelona, Spain). The presence of goblet cells in the conjunctival samples of the same patient was evaluated as a positive control. In addition, samples from healthy control individuals were also tested as negative controls in the first stages of the study. The specificity of detection of the MUC5AC transcript was verified by using mRNA samples from the corneas of healthy individuals (MUC5AC negative). RNAse-free water was also used in RT-PCR instead of template as a negative control to demonstrate the absence of false positives by contamination. Periodic Acid-Schiff–Hematoxylin Staining Impression cytology samples for PAS-hematoxylin staining were obtained on 5⫻5-mm strips of cellulose acetate, immediately fixed in 96% ethanol, and subsequently stained with PAS-hematoxylin according to the Locquin and Langeron protocol, modified by Rivas et al.20 The samples were later examined under light microscopy. The presence of goblet cells in the cornea was considered to be indicative of LSCD. In addition, we measured the areas of the cytoplasm and nucleus of non-secretory cells, cytoplasmic alterations and staining, nuclear alterations, and ratio of the nuclear and cytoplasmic areas. RNA Isolation, Quality Assessment, and Reverse Transcription For RT-PCR assays, two 8-mm diameter cellulose acetate discs were used. Both sides of each disc were placed in contact with the epithelium using a sterile tweezers to obtain the highest possible number of cells. One disc was placed on the corneal epithelium, and subsequently the other disc was placed on the epithelium of the superior bulbar conjunctiva. Slight pressure was applied to the discs for a few seconds to improve the efficacy of the sample. The discs were immediately placed in an RNA-conserving buffer (RNAprotect Cell Reagent, Qiagen, Valencia, CA) and stored at 4°C until use. Total RNA was isolated using the RNeasy plus micro kit (Qiagen). The membrane containing epithelial cells was transferred to a clean 1.5-ml RNAse-free Eppendorf tube (Hamburg, Germany) and briefly centrifuged to completely eliminate the RNA-conserving buffer with a micropipette. Next, 350 l of lysis buffer were added to the tube containing the membrane, and after vortexing for 1 minute at maximum speed, the resulting lysate was transferred to the gDNA eliminator column. The rest of the steps were performed as described by the manufacturer, using 16 l of diethyl pyrocarbonate-treated H2O for final RNA elution. All steps were performed at room temperature. Quantification and quality assessment of total RNA were performed using a 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA) and RNA Pico kits (Agilent Technologies, Inc.), using 1 l of each sample. Only those samples with an RNA Integrity Number ⬎7 were included in the study. Although no limitation for RNA concentration was established for positive results, we set a cutoff of 1.2 ng/l for a reliable negative result, Garcia et al 䡠 LSCD Diagnosis by Detection of MUC5AC mRNA Table 1. Demographic Data Purification and Sequencing of Amplicons Variable Value (n) Eyes Patients Gender Male Female Mean age (yrs) ⫾ SD 59 43 24 19 59.2⫾14.6 SD ⫽ standard deviation. below which the result was considered to be inconclusive or unreliable. Reverse transcription of total RNA to cDNA was carried out using a Transcriptor First Strand cDNA Synthesis Kit (Roche, Mannheim, Germany), with a mixture of Oligo(dT) and random hexamers as primers. The reaction conditions included a denaturation step (60°C for 10 minutes and then immediately to 4°C) for the RNA and primer mixture, followed by high-temperature RT after adding the remaining reagents (25°C, 10 minutes; 55°C, 30 minutes; 85°C, 5 minutes; 4°C). Primer Design Because of the repetitive and palindromic structure of the MUC5AC gene and its transcript, together with its high homology with MUC5B, several analyses were performed to identify suitable regions for primer design (data not shown), including Nucleic Acid Dot Plots (available at: http://www.vivo.colostate.edu/molkit/ dnadot/, accessed February 5– 8, 2009), RNAfold from Vienna RNA Suite (available at: http://rna.tbi.univie.ac.at/cgi-bin/RNA fold.cgi, accessed February 10, 2009), Primer BLAST (available at: http://www.ncbi.nlm.nih.gov/tools/primer-blast/, accessed February 12, 2009), BLAST (available at: http://blast.ncbi.nlm.nih. gov/Blast.cgi, accessed February 13, 2009), and BLAT (available at: http://genome.ucsc.edu/cgi-bin/hgBlat?command⫽start, accessed February 13, 2009). Primers were tailor designed on the basis of theoretically optimal candidate regions that ensured specificity (low homology with other family transcripts) and efficacy (e.g., absence of RNA secondary structure domains). The corresponding amplicons were theoretically reevaluated again by means of Primer BLAST (March 5, 2009), BLAST (March 5, 2009), BLAT (March 5, 2009) for specificity, and RNAfold (March 6, 2009) for palindromic structure. A primer pair was thus designed and denominated MUC5AC-F (CCTGCAAGCCTCCAGGTAG) and MUC5AC-R (CTGCTCCACTGGCTTTGG). To confirm that PCR products represented the amplified gene transcript, amplicons obtained during initial assays were excised from agarose gels, purified using Illustra GFX PCR DNA and Gel Band Purification kit (GE-Healthcare, Little Chalfont, UK), and sequenced on a 3100 genetic analyzer using Big Dye 3.1 terminator chemistry (both analyzer and software from Applied Biosystems, Foster City, CA). Results A total of 59 eyes from 43 patients were included in this study from the outpatient clinics of the ophthalmology departments of each of the participating hospitals. The mean age of patients was 59.2⫾14.6 years (24 men, 19 women), as indicated in Table 1. Patients were clinically diagnosed with LSCD due to a variety of causes, including chemical burn, pemphigoid, aniridia, Stevens– Johnson syndrome, trachoma, contact lenses, rosacea, idiopathic disease, and others (Table 2). In Figures 1, 2, and 3, corresponding to patients 9, 15A, and 4, respectively, some representative cases of LSCD are shown, each from a different hospital. Of the 59 eyes included in this study, 47 IC samples were also available for PAS-hematoxylin staining. The results indicated that of these, 17 (36.2%) had conjunctival goblet cells in the cornea (as illustrated in Fig 4A, corresponding to patient 9), thus confirming LSCD, whereas 27 (57.5%) did not (Fig 4B, corresponding to patient 11). Although the sample was negative in this patient according to PAS-hematoxylin, the sample was found to be positive by PCR). Moreover, 3 samples (6.4%) showed inconclusive results because of the scarce number of cells in the corneal epithelium samples. The results are summarized in Table 3. In addition, the conjunctival epithelium from all patients exhibited squamous metaplasia; most of these were cases of grade 2 to 4, which correlated with clinical signs. Conjunctival epithelial cells presented a nuclear-cytoplasmic ratio between 1:10 and 1:15, with small and pyknotic nuclei, and a cell shape that was large and polygonal (Fig 4C). Figure 4C shows an example of a grade 3 squamous metaplasia. To verify the specificity of the RT-PCR assay, purified PCR amplicons were sequenced. The sequencing chromatograms and alignment of the resulting sequences with the MUC5AC cDNA reference sequence confirmed the specificity of the amplified fragment (data not shown). To ensure the specificity of MUC5AC transcript amplification, the primers were also tested with RNA from corneal epithelium samples from healthy volunteers, under exactly the same experimental conditions as the rest of samples analyzed in the study, to further demonstrate that no amplification was detectable in healthy Polymerase Chain Reaction BioTaq DNA polymerase, reaction buffer, magnesium chloride, and dNTPs (Bioline Inc., Tauton, MA) were used to perform the amplification reaction. Custom primers (synthesized by MWGBiotech, Inc., High Point, NC) were used to specifically amplify a 103– base-pair (bp) fragment of the reversely transcribed MUC5AC transcript (cDNA). Detection of amplicons was performed using 2.5% agarose–Tris-acetate-EDTA gels. Primers for the ACTB and GAPDH housekeeping genes (Applied Biosystems, Warrington, UK) were used as internal controls for the quality and quantity of the reversely transcribed RNA (cDNA). The amplification conditions were 95°C, 3 minutes/(95°C, 15 seconds; 60°C, 15 seconds; 72°C, 15 seconds) ⫻40/72°C, 5 minutes/15°C on hold. Table 2. Causes of Limbal Stem Cell Deficiency Pathology Eyes (n) Chemical burn Pemphigoid Aniridia Stevens–Johnson syndrome Trachoma Contact lenses Rosacea Idiopathic disease Others Total 8 7 6 5 3 2 2 8 18 59 3 Ophthalmology Volume xx, Number x, Month 2012 Figure 1. Eye of a 68-year-old man (patient 9) with Stevens–Johnson syndrome. Note 360-degree corneal vascularization. Figure 3. Eye of a 20-year-old man (patient 4). Corneal opacity due to chemical burn. corneas. For this purpose, 2 healthy volunteers were tested, and no amplification of MUC5AC transcript was detected in the cornea sample; nevertheless, amplification of the GAPDH and ACTB housekeeping genes in the corneal epithelium samples and adequate RNA concentration (⬎1.2 ng/l) confirmed the reliability of the MUC5AC negative result (Fig 5A, B, available at http:// aaojournal.org). In these healthy samples, MUC5AC transcript amplification was also tested in conjunctival epithelium as a positive control, as was carried out in all samples of the study. Of the 59 eyes analyzed by RT-PCR, 56 (94.9%) were MUC5AC ⫹ according to PCR, therefore confirming LSCD, whereas 3 (5.1%) were MUC5AC – (Table 3). The PCR products were resolved using 2.5% Tris-acetate-EDTA–agarose gel electrophoresis. The samples shown in Figure 5C and D (available at http://aaojournal.org) correspond to 2 eyes clinically diagnosed with LSCD. Three different bands are shown in each of these gels: MUC5AC (103 bp, lanes 1 and 3), ACTB (171 bp, multiplexed with MUC5AC, lanes 1 and 3), and GAPDH (122 bp, individual reaction, lanes 2 and 4). Two separate reactions were performed to detect 3 amplicons in each tissue because of the difficulty of multiplexing more than 2 PCRs and the close vicinity of the 3 amplicons that could lead to confusing results. Lanes 1 and 2 correspond to the corneal samples, whereas lanes 3 and 4 correspond to the conjunctival samples used as a positive control for the presence of the MUC5AC transcript. Lanes 5 and 6 correspond to negative controls (lane 5, omission of reverse transcriptase in the RT reaction; lane 6, omission of cDNA in the PCR). Amplification of ACTB and GAPDH served as quality and quantity controls of the cDNA of each tissue. Examples of 2 cases diagnosed with LSCD are presented. Figure 5C (available at http://aaojournal.org) shows the case of a patient with LSCD (patient 9) for whom MUC5AC mRNA was detected as a 103-bp amplicon in both cornea (lane 1) and conjunctiva (lane 3). In contrast, Figure 5D (available at http://aaojournal.org) shows the case of a patient with LSCD (15A) with no MUC5AC mRNA being detected in the cornea. Amplification of ACTB (171 bp, multiplexed with MUC5AC, lane 1) and GAPDH (122 bp, individual reaction, lane 2) in the corneal mRNA sample indicates that RNA quantity and quality were satisfactory, thus confirming the negative result for MUC5AC. Comparative analysis of the 2 techniques indicated that 36.2% of analyzed samples (17/47) were positive according to both techniques. However, 57.5% of the total samples (27/47) were negative according to the PAS-hematoxylin technique. Finally, 6.4% corresponding to 3 samples were inconclusive according to PAS-hematoxylin because of the scarcity of the sample (Table 3). Discussion Figure 2. Right eye of a 68-year-old woman (patient 15A) with a history of deep myopic right amblyopia and childhood keratitis in both eyes. Right eye corneal biomicroscopy revealed micropannus (arrow), central corneal opacity, and ghost vessels. The periodic acid-Schiff– hematoxylin staining and reverse transcription-polymerase chain reaction test results were negative. This eye was the donor for autologous expanded limbal stem cell transplantation to the severely damaged left eye. 4 In the present work, we present a method for the diagnosis of LSCD based on the highly specific detection of MUC5AC mRNA in the cornea. The sensitivity of this molecular technique was compared with that of conventional PAShematoxylin staining of IC samples and clinical diagnosis. Numerous cases arise in clinical practice in which despite the presence of a clinical suspicion of LSCD, there is Garcia et al 䡠 LSCD Diagnosis by Detection of MUC5AC mRNA Figure 4. A, Microphotograph of corneal impression cytology (IC) with periodic acid-Schiff– hematoxylin stain from patient 9. Large epithelial cells that have lost their rounded morphology can be observed. Intercellular junctions are often lost, and spaces appear between the cells. Goblet cells can be seen in this sample (arrow), confirming the diagnosis of limbal stem cell deficiency. Magnification, 40⫻. B, In this corneal epithelium sample from patient 11, no goblet cells were observed (20⫻). C, Representative IC of conjunctiva. In this sample, the conjunctival epithelium exhibited grade 3 squamous metaplasia. Conjunctival epithelial cells presented a nuclear-cytoplasmic ratio of 1:15, with small and pyknotic nuclei, and a large and polygonal cell shape (20⫻). no cytologic evidence of this pathology according to conventional IC. This discrepancy may be due to a number of reasons: Damage of limbal corneal cells may not be sufficiently severe to destroy the entire cell population, LSCD may still be subclinical and worsen later, IC might not be sufficiently sensitive, and some patients simply may not manifest signs of LSCD.21 For this reason, a more sensitive and specific method is needed to diagnose LSCD, particularly in a patient who has only mild LSCD, and to more rigorously classify or select patients with corneal pathology who could undergo keratoplasty with a reasonable probability of success to minimize the risk of failure of this surgical technique. In addition, this molecular method can be used together with conventional IC to verify the restoration of the corneal phenotype and the regression of goblet cells.22,23 The current study has shown that RT-PCR– based analysis is more sensitive and specific than PAS-hematoxylin staining because RT-PCR detected more positive cases: 56 of 59 eyes clinically diagnosed with LSCD, equivalent to a diagnostic efficacy of approximately 95% (Table 3). The specificity of the RT-PCR assay was ensured by means of the custom design of primers for the amplification of the MUC5AC transcript. During the first stages of our study, several published primer pairs for this gene were bioinformatically evaluated.24 –27 However, our analyses showed that most of these primers were designed on the basis of non-validated sequences available at that time in public databases, such as predicted sequences and partial clones, or expected to simultaneously amplify MUC5B Table 3. Summary of Results for the Different Diagnostic Methods Clinical ⴙ — Inconclusive Total PAS RT-PCR n % n % n % 59 0 0 59 100 — — 17 27 3 47 36.2 57.5 6.4 56 3 0 59 94.9 5.1 — PAS ⫽ periodic acid Schiff; PASS ⫽ PAS-hematoxylin staining; RTPCR⫽reverese transcription-polymerase chain reaction. (primers tested by means of the National Center for Biotechnology Information’s PrimerBlast). This is possible because of the high homology between MUC5AC and MUC5B, both of which are derived from the same genomic locus with different transcription initiation sites. The possibility of obtaining nonspecific results discarded the use of these published primer sequences as a suitable method for diagnosis and encouraged us to design the primers ourselves. Thus, multiple bioinformatic analyses were performed to identify the most suitable regions to be targeted by specific primers. After selecting a candidate primer pair, it was bioinformatically tested for specificity. Subsequently, the specificity of PCR detection was demonstrated by sequencing the PCR products obtained in our first assays; all the sequenced fragments (obtained from 3 different samples) were found to be part of the targeted MUC5AC transcript and showed no significant nonspecific targets throughout the human transcriptome or genome. Of note, no homology with MUC5B was observed. According to our results, of the 59 eyes analyzed by RT-PCR, 56 were found to present detectable levels of MUC5AC, confirming the LSCD diagnosis, whereas 3 did not. Negative results were found for patients 15A, 16B, and 17. The reliability of the negative results was evaluated with strict RNA quantity and quality controls using both the Bioanalyzer and housekeeping genes. On the other hand, positive PCR controls consisting of RNA from conjunctival epithelium were also simultaneously performed to evaluate whether the absence of amplification was due to RT or PCR inhibitors. Revision of the clinical history of patients who were LSCD negative by RT-PCR revealed that patient 16B had been clinically diagnosed with left eye superior limbic keratoconjunctivitis and LSCD on the basis of positive IC in 2005 and 2008; patient 16B’s condition improved with treatment, and she was included in this study to analyze the status of her limbus 1 year later. The fact that the MUC5AC transcript was not detected in her cornea suggests that the treatment had been effective and that LSCD had been corrected. 5 Ophthalmology Volume xx, Number x, Month 2012 Patient 15A reported a history of childhood keratitis. Her right eye exhibited micropannus, central corneal opacity, and ghost vessels (Fig 2). She presented with severe painful LSCD, whose cause was associated with the protracted use of glaucoma eyedrops plus corneal erosions. Both PAS and RT-PCR test results of her right eye were negative, and therefore this eye was used as a donor for the autologous transplantation to the severely damaged left eye of limbal stem cells expanded in vitro. Patient 17 presented corneal conjunctivalization that was restricted to the superior zone. It is conceivable that the disc used to obtain the sample during IC might not have been placed in contact with the affected zone, thereby explaining the absence of goblet cells in the cornea and the negative RT-PCR result for the MUC5AC transcript. The method we have presented represents a good tool for clinical practice because it is highly specific, sensitive, rapid, and noninvasive. In addition, it facilitates the selection of patients who can undergo keratoplasty with a higher probability of success, with a view to reducing the risk of failure of the said technique. This method can also facilitate the selection or rejection of a cornea as an adequate source of stem cells for autologous limbal transplantation. The large number of cases that were RT-PCR positive and PAS negative highlights the significantly improved sensitivity afforded by the exquisitely sensitive PCR technique. Another clear advantage is that the RT-PCR– based method produces objective results, which are not subject to subjective interpretation. Nevertheless, there are admittedly a number of inconveniences associated with this molecular technique, such as the obtaining of the sample. Thus, it is critically important to obtain an optimal number of corneal epithelial cells, because few cells lead to lower performance and falsenegative results. For this reason, we suggest a minimal RNA concentration of 1.2 ng/L, below which negative results should be considered as unreliable, and the inclusion of housekeeping genes as an additional quantity and quality control. More important, it is vital that samples are taken exclusively from the cornea; false-positive results can arise if the sampling disc touches part of the conjunctival epithelium, because the PCR technique is extremely sensitive. In addition, given that this method is based on RNA isolation and expression analysis, the lability of RNA must be considered, and working under RNAse-free conditions becomes essential. Finally, this technique does not provide spatial information about the distribution of goblet cells in the cornea, making full versus partial LSCD indistinguishable. Several reports about the use of confocal microscopy for LSCD diagnosis have been published in recent years.28 –30 This technique could circumvent some of the limitations of the PCR-based technique, especially when trying to distinguish between full and partial LSCD. Although the use of this technology is promising, it has its own drawbacks, such as the high costs of the equipment, the limited number of laboratories with a confocal microscope, and the need for specialized expertise when manipulating the apparatus and interpreting the data. In contrast, PCR-based methods are widespread now and can be performed in any laboratory at 6 a low cost by routine technicians. The PCR technique also produces objective and easily interpretable results. Moreover, the exquisite sensitivity of RT-PCR facilitates the early detection of LSCD, which can be more difficult to detect with confocal microscopy. In this regard, the possible absence of cytologic evidence of LSCD during the initial stages of the pathology can easily produce false-negative results with confocal microscopy, a limitation that can be significantly reduced when using PCR-based technologies. In conclusion, the method described constitutes a robust system for the early detection of mild cases of LSCD and the corroboration of uncertain clinical cases. Moreover, this methodology has multiple potential applications beyond diagnosis, including the monitoring of treated patients, the evaluation of the evolution of the corneal epithelium after keratoplasty or amniotic membrane transplant, and the examination of the limbal condition of donor eyes before autologous stem cell transplantation. Finally, the use of this technique in clinical practice can have important economic repercussions, because it can advise against a corneal transplant in some patients diagnosed with LSCD. References 1. Schermer A, Galvin S, Sun TT. Differentiation-related expression of a major 64K corneal keratin in vivo and in culture suggests limbal location of corneal epithelial stem cells. J Cell Biol 1986;103:49 – 62. 2. Puangsricharern V, Tseng SC. Cytologic evidence of corneal diseases with limbal stem cell deficiency. Ophthalmology 1995;102:1476 – 85. 3. Chen JJ, Tseng SC. Corneal epithelial wound healing in partial limbal deficiency. Invest Ophthalmol Vis Sci 1990; 31:1301–14. 4. Chen JJ, Tseng SC. Abnormal corneal epithelial wound healing in partial thickness removal of limbal epithelium. Invest Ophthalmol Vis Sci 1991;32:2219 –33. 5. Kruse FE, Chen JJ, Tsai RJ, Tseng SC. Conjunctival transdifferentiation is due to the incomplete removal of limbal basal epithelium. Invest Ophthalmol Vis Sci 1990;31:1903–13. 6. Gipson IK, Inatomi T. Mucin genes expressed by the ocular surface epithelium. Prog Retin Eye Res 1997;16:81–98. 7. Gum JR Jr, Ho JJ, Pratt WS, et al. MUC3 human intestinal mucin: analysis of gene structure, the carboxyl terminus and a novel upstream repetitive region. J Biol Chem 1997;272: 26678 – 86. 8. Moniaux N, Nollet S, Porchet N, et al. Complete sequence of the human mucin MUC4: a putative cell membrane-associated mucin. Biochem J 1999;338:325–33. 9. Mehrotra R, Thornton DJ, Sheehan JK. Isolation and physical characterization of the MUC7 (MG2) mucin from saliva: evidence for self-association. Biochem J 1998;334:415–22. 10. Meezaman D, Charles P, Daskal E, et al. Cloning and analysis of cDNA encoding a major airway glycoprotein, human tracheobronchial mucin (MUC5). J Biol Chem 1994;269: 12932–9. 11. Lapensee L, Paquette Y, Bleau G. Allelic polymorphism and chromosomal localization of the human oviductin gene (MUC9). Fertil Steril 1997;68:702– 8. 12. Inatomi T, Spurr-Michaud S, Tisdale AS, Gipson IK. Human corneal and conjunctival epithelia express MUC1 mucin. Invest Ophthalmol Vis Sci 1995;36:1818 –27. Garcia et al 䡠 LSCD Diagnosis by Detection of MUC5AC mRNA 13. Inatomi T, Spurr-Michaud S, Tisdale AS, et al. Expression of secretory mucin genes by human conjunctival epithelia. Invest Ophthalmol Vis Sci 1996;37:1684 –92. 14. Ang LP, Tan DT, Beuerman RW, Lavker RM. Development of a conjunctival epithelial equivalent with improved proliferative properties using a multistep serum-free culture system. Invest Ophthalmol Vis Sci 2004;45:1789 –95. 15. Kasper M, Moll R, Stosiek P, Karsten U. Patterns of cytokeratin and vimentin expression in the human eye. Histochemistry 1988;89:369 –77. 16. Liu S, Li J, Tan DT, Beuerman RW. The eyelid margin: a transitional zone for 2 epithelial phenotypes. Arch Ophthalmol 2007;125:523–32. 17. Corrales RM, Chaves FJ, Herreras JM, et al. Normal human conjunctival epithelium expresses MUC13, MUC15, MUC16 and MUC17 mucin genes [in Spanish]. Arch Soc Esp Oftalmol 2003;78:375– 81. 18. Argüeso P, Balaram M, Spurr-Michaud S, et al. Decreased levels of the goblet cell mucin MUC5AC in tears of patients with Sjögren syndrome. Invest Ophthalmol Vis Sci 2002;43: 1004 –11. 19. Corrales RM, Galarreta D, Herreras JM, et al. Conjunctival mucin mRNA expression in contact lens wear. Optom Vis Sci 2009;86:1051– 8. 20. Rivas L, Oroza MA, Perez-Esteban A, Murube-del-Castillo J. Topographical distribution of ocular surface cells by the use of impression cytology. Acta Ophthalmol (Copenh) 1991;69: 371– 6. 21. Daya SM, Bell RW, Habib NE, et al. Clinical and pathologic findings in human keratolimbal allograft rejection. Cornea 2000;19:443–50. 22. Kenyon KR, Tseng SC. Limbal autograft transplantation for ocular surface disorders. Ophthalmology 1989;96:709 –23. 23. Prabhasawat P, Tseng SC. Impression cytology study of epithelial phenotype of ocular surface reconstructed by preserved human amniotic membrane. Arch Ophthalmol 1997;115: 1360 –7. 24. Gipson IK, Spurr-Michaud S, Argüeso P, et al. Mucin gene expression in immortalized human corneal-limbal and conjunctival epithelial cell lines. Invest Ophthalmol Vis Sci 2003; 44:2496 –506. 25. Sabo E, Meitner PA, Tavares R, et al. Expression analysis of Barrett’s esophagus-associated high-grade dysplasia in laser capture microdissected archival tissue. Clin Cancer Res 2008 2008;14:6440 – 8. 26. Ou XM, Wang BD, Wen FQ, et al. Simvastatin attenuates lipopolysaccharide-induced airway mucus hypersecretion in rats. Chin Med J (Engl) 2008;121:1680 –7. 27. Kraft M, Adler KB, Ingram JL, et al. Mycoplasma pneumoniae induces airway epithelial cell expression of MUC5AC in asthma. Eur Respir J 2008;31:43– 6. 28. Cho BJ, Djalilian AR, Holland EJ. Tandem scanning confocal microscopic analysis of differences between epithelial healing in limbal stem cell deficiency and normal corneal reepithelialization in rabbits. Cornea 1998;17:68 –73. 29. Kobayashi A, Sugiyama K. In vivo corneal confocal microscopic findings of palisades of Vogt and its underlying limbal stroma. Cornea 2005;24:435–7. 30. Dua HS, Miri A, Alomar T, et al. The role of limbal stem cells in corneal epithelial maintenance: testing the dogma. Ophthalmology 2009;116:856 – 63. Footnotes and Financial Disclosures Originally received: May 19, 2011. Final revision: October 24, 2011. Accepted: October 24, 2011. Available online: ●●●. Manuscript no. 2011-755. 1 Bioftalmik, Parque Tecnológico de Vizcaya, Vizcaya, Spain. 2 Hospital de Cruces, Vizcaya, Spain. 3 Hospital La Paz, Madrid, Spain. 4 Hospital Clínico San Carlos, Madrid, Spain. 5 Hospital Ramón y Cajal, Madrid, Spain. 6 Hospital de Donostia, Guipúzcoa, Spain. 7 Hospital 12 de Octubre, Madrid, Spain. 8 Hospital de Galdakao-Usansolo, Vizcaya, Spain. Financial Disclosure(s): The author(s) have made the following disclosure(s): The authors declare that the results reported in this article, including the patented MUC5AC primer sequences, may be of commercial interest to Bioftalmik, Vizcaya, Spain. Financial Support: The Centre for the Development of Industrial Technology partially supported this work through its NEOTEC Program, Grant IDI-20080118. Correspondence: Arantxa Acera, PhD, Bioftalmik, Parque Tecnológico de Vizcaya, Ed. 800, 2a. Planta, E-48160 Derio, Vizcaya, Spain. E-mail: arantxa.acera@ bioftalmik.com. 7