and confocal laser scanning microscopy ... - Semantic Scholar

22 downloads 0 Views 4MB Size Report
Sep 8, 2009 - acideSchiff staining or Mucin 5AC staining) if there are no goblet cells in the sample. In addition, the keratin markers currently used are unable ...
Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Evaluation of ocular surface disorders: a new diagnostic tool based on impression cytology and confocal laser scanning microscopy Vanessa Barbaro, Stefano Ferrari, Adriano Fasolo, et al. Br J Ophthalmol 2010 94: 926-932 originally published online September 8, 2009

doi: 10.1136/bjo.2009.164152

Updated information and services can be found at: http://bjo.bmj.com/content/94/7/926.full.html

These include:

References

This article cites 23 articles, 10 of which can be accessed free at: http://bjo.bmj.com/content/94/7/926.full.html#ref-list-1

Email alerting service

Receive free email alerts when new articles cite this article. Sign up in the box at the top right corner of the online article.

Notes

To order reprints of this article go to: http://bjo.bmj.com/cgi/reprintform

To subscribe to British Journal of Ophthalmology go to: http://bjo.bmj.com/subscriptions

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science

Evaluation of ocular surface disorders: a new diagnostic tool based on impression cytology and confocal laser scanning microscopy Vanessa Barbaro,1 Stefano Ferrari,1 Adriano Fasolo,1 Emilio Pedrotti,2 Giorgio Marchini,2 Arianna Sbabo,2 Nicola Nettis,1 Diego Ponzin,1 Enzo Di Iorio1 1

The Veneto Eye Bank Foundation, Venice, Italy 2 Eye Clinic, Department of Neurological and Visual Sciences, University of Verona, Verona, Italy Correspondence to Dr Enzo Di Iorio, The Veneto Eye Bank Foundation, Via Paccagnella, 11 c/o Padiglione Rama, 30174 Zelarino, Venice, Italy; [email protected] Accepted 18 August 2009 Published Online First 8 September 2009

ABSTRACT Aim To provide a new tool for the evaluation of altered ocular surfaces by using a combination of impression cytology, laser scanning confocal microscopy and advanced image analysis. Methods The expression of keratin 3 (K3), keratin 12 (K12), keratin 19 (K19) and mucin 1 (MUC1) was analysed by immunofluorescence on both histological sections of nine corneoscleral buttons from normal donors comprising conjunctiva, limbus and cornea and impression cytology specimens from six healthy normal subjects (12 eyes) and 12 patients with chronic ocular surface disorders. Levels of fluorescence expression of the different markers were quantified through quantitative fluorescence immunohistochemistry (Q-FIHC). Results Impression cytology specimens from normal and diseased ocular surfaces showed distinct expression patterns for K12 and MUC1. Healthy corneas expressed only K12 (but not MUC1), while conjunctivalised corneas from patients with limbal stem cell deficiency (LSCD) were characterised by the presence of MUC1 and the disappearance of K12. Similar clear-cut results were not seen with the K3/K19 markers, which showed lack of specificity and overlapping signals in cornea and conjunctiva impression cytology specimens. Conclusions The ability of K12 and of the antibody against MUC1 to discriminate clearly between limbus/ cornea and conjunctiva in impression cytology specimens could become a valuable diagnostic tool for ophthalmologists in order to evaluate alterations of the ocular surface and the grading of LSCD.

INTRODUCTION Cornea, conjunctiva and lacrimal glands form the ocular surface of the eye and play an important role in preventing pathogen invasion, desiccation and injuries.1 2 Stem cells are essential for the integrity of the ocular surface, by promoting renewal in healthy states and re-epithelialisation in wound healing.3e7 Corneal epithelial stem cells are located in the basal layers of the limbus, while in conjunctiva stem cells are found in the fornical and bulbar zones.7e9 When limbal stem cells are damaged or destroyed (because of infections, chemical or physical burns, tumours or immunological diseases) the invasion of cornea by conjunctival epithelium, replete with goblet cells and vessels (ie, conjunctivalisation), occurs,3 resulting in neovascularisation, corneal opacity and loss of vision.10 A common feature to 926

all these disorders is the depletion of stem cells from the limbus.11e13 Assessment of the presence of limbal stem cells is therefore essential for the diagnosis of limbal stem cell deficiency (LSCD) and to predict the outcome of ocular surface reconstruction. Diagnosis of LSCD relies on the confirmation of cornea conjunctivalisation, through either the presence of goblet cells6 or the altered expression of keratins14 15 in specimens obtained by impression cytology (IC). However, it is difficult to distinguish conjunctival epithelia from corneal epithelia by conventional cytology techniques (haematoxylineeosin, Periodic acideSchiff staining or Mucin 5AC staining) if there are no goblet cells in the sample. In addition, the keratin markers currently used are unable to distinguish clearly between limbus/cornea and conjunctiva.16 17 The aim of this study is to improve the quality of diagnostic tests based on IC through the selection of more reliable markers of the ocular surface epithelia, namely keratin 12 (K12) and mucin 1 (MUC1), and quantification of their expression by means of laser scanning confocal microscope.

MATERIALS AND METHODS Human corneoscleral button samples Nine corneoscleral buttons unsuitable for transplantation were collected from normal donors approximately 5e48 h after death. All specimens had intact conjunctival and limbo-corneal epithelia, and were chosen on the basis of characteristics previously described.18 Corneas were evaluated through slit-lamp examination, and those not showing epithelial defects, dehydration, oedema or inflammation were chosen for further analyses.

Patients Cytological specimens were collected by impression from six normal subjects (aged 60e74 years; table 1) and 12 patients (aged 2e78 years) with different degrees of conjunctivalisation. For each patient, details of age, cause of disease and previous treatments are indicated in table 2. The research study adhered to the tenets of the Declaration of Helsinki.

Cell collection and processing through IC Sixty IC specimens were analysed. For validation studies, specimens were obtained from both eyes of healthy subjects (n¼6). For each donor, three IC specimens were taken from each eye: one from the temporal bulbar conjunctiva (2e8 mm from the Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science Table 1

Quantitative fluorescence immunohistochemistry analysis of impression cytology specimens from healthy donors (n¼6)

Donor

Age (years)

1

64

2

62

3

74

4

71

5

60

6

71

Area of impression cytology

Keratin 3 (%)

Keratin 19 (%)

Merge* keratin 3/keratin 19

Keratin 12 (%)

Mucin 1 (%)

Merge* keratin 12/mucin 1

Conjunctiva Limbus Central cornea Conjunctiva Limbus Central cornea Conjunctiva Limbus Central cornea Conjunctiva Limbus Central cornea Conjunctiva Limbus Central cornea Conjunctiva Limbus Central cornea

70 87 99 31 81 80 84 85 95 39 81 96 97 45 79 65 45 92

63 61 1 75 44 66 59 80 64 93 81 92 70 87 47 88 67 46

0.99 0.98 0.01 0.97 0.95 0.91 0.89 1.00 0.89 0.79 1.00 0.02 0.97 0.86 0.79 0.74 0.82 0.78

1 42 99 8 56 99 1 34 99 1 48 95 1 31 99 1 21 99

99 58 1 92 44 1 99 66 1 99 52 5 99 69 1 99 79 1

0.09 0.09 0.07 0.01 0.08 0.07 0.01 0.05 0.08 0.09 0.09 0.02 0.09 0.09 0.07 0.01 0.01 0.07

*Colocalisation coefficient, value range¼0e1 (0, no colocalisation; 1, all pixels colocalise).

limbus), one from the limbus and one from the central cornea (for a total of n¼36 specimens). For patients with ocular surface disorders (n¼12), two membranes were used for each affected eye (for a total of n¼24 specimens). To perform the IC, subjects were anaesthetised topically before a sterile membrane (Millicell CM 0,4 mm, Millipore, Bedford, Massachusetts) was gently pressed onto the ocular surfaces for a few seconds. To increase the number of cells collected, the ocular surface was slightly dried up by keeping the eye open before sampling. With this procedure, approximately 50e70% of the epithelial cells of the ocular surface were collected onto the membranes. Membranes were labelled to match the presence of specific markers with the Table 2 (n¼12)

sector of the ocular surface analysed. Cells were fixed with KitoFix (Kaltek, Padua, Italy). Membranes were cut with a blade, laid down into wells of a multiwell plate, mounted using a mounting medium containing DAPI (Vectashield, Vector Laboratories, DBA, Milan, Italy) and analysed by indirect immunofluorescence. All samples were tested through double immunofluorescence for both K3/K19 and K12/MUC1 expression as below described.

Immunofluorescence staining Human corneoscleral samples fixed in 3% paraformaldehyde (overnight at 48C) were embedded in OCTcompound, frozen and

Quantitative fluorescence immunohistochemistry analysis of impression cytology specimens from patients with oular surface disorders

Age Patient (years)

Cause of disease

Previous therapeutic treatments

1

52

2

62

3

40

4

61

5

2

6

53

7

78

8

60

9 10 11

69 49 52

12

31

Chemical burn < Amniotic membrane (32) Complete conjunctivalisation < Stem cell graft +neovascularisation (3608) Trauma Penetrating keratoplasty Conjunctivalisation + neovascularisation peripheral areas. Clear cornea centrally Chemical burn Amniotic membrane Complete conjunctivalisation +neovascularisation (3608) Chemical burn < Oral mucosa Neovascularisation but clear < Lamellar keratoplasty (32) comea centrally < Stem cell graft Likely infection Penetrating keratoplasty Complete conjunctivalisation before birth +neovascularisation (3608) Chemical burn Amniotic membrane Complete conjunctivalisation +neovascularisation (3608) Chemical burn None Neovascularisation in peripheral areas Chemical burn Treated for Complete conjunctivalisation symblepharon Chemical burn Amniotic membrane Complete conjunctivalisation NA NA Complete conjunctivalisation NA Amniotic membrane Symblepharon, partial conjunctivalisation and neovascularisation Chemical burn Stem cell graft Conjunctivalisation in different areas of the cornea

Clinical outcome

Keratin 3 (%)

Keratin 19 (%)

Merge* keratin 3/keratin 19

54

72

0.99

79

63

82

Keratin 12 (%)

Mucin 1 (%)

Merge* mucin 1/keratin 12

2

98

0.09

0.97

11

92

0.05

56

0.99

1

99

0.01

61

33

0.91

29

71

0.10

28

65

0.88

20

80

0.03

2

80

0.01

1

99

0.06

18

92

0.57

4

96

0.02

32

68

0.99

5

95

0.09

41 32 45

81 68 70

1.00 0.79 0.70

28 20 35

72 80 65

0.01 0.09 0.07

27

51

0.66

1

99

0.04

NA, information not available. *Colocalisation coefficient, value range 0e1 (0, no colocalisation; 1, all pixels colocalise).

Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152

927

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science cut into 5e7 mm sections. Corneoscleral sections and IC specimens were analysed through indirect immunofluorescence by using antibodies against K12 (sc-17099, goat polyclonal, 1:100, Santa Cruz Biotechnology, Santa Cruz, California), MUC1 (H-295) (sc-15333, rabbit polyclonal, 1:200, Santa Cruz Biotechnology), K3 (mouse AE5 clone, 1:100, MP Biomedicals, Solon, Ohio) and K19 (RB-9021, rabbit polyclonal, 1:200, NeoMarkers, Freemont, California) for 1 h at 378C. Rhodamine and FITC-conjugated secondary antibodies (1:100, Santa Cruz Biotechnology) were incubated for 1 h at room temperature. Specimens were analysed with LSM 510 Meta Confocal Microscope. Image and Z-stack analyses were performed with LSM 510 software (Zeiss, Milan, Italy) and quantification of fluorescent signals performed as previously described.19

RESULTS Criteria for selection of markers to be used for analysis of IC specimens The aim of this study is to select more reliable markers of the cornea and conjunctiva to use in IC specimens. Potential candidates should be expressed either in the cornea or in the conjunctiva, not secreted in the extracellular spaces and localised within the first apical layers of ocular surface epithelia. In fact, Z-stack analyses showed that the thickness of the specimens obtained through IC corresponds to that of just one cell layer

(the apical one) and only occasionally includes the subapical cell layers (data not shown). Only the most superficial cells of the ocular surface are therefore collected onto the IC membranes and analysed. K12 was chosen as previously reported to be restricted to the cornea16 17 and mRNAs found in primary human corneal cells, but not in conjunctiva (data not shown). For the conjunctiva, we used a polyclonal antibody (H-295) raised against the amino acids 961e1255 mapping at the C-terminus of the human MUC1.20 The antibody recognises all the MUC1 splice variants (evaluated through http://service.uniprot.org/clustalw), but not the secreted MUC1/SEC isoform.21 K12 and MUC1 were compared with K3 and K19, previously reported to be suitable for identification of corneal and conjunctival epithelial cells in IC specimens.14 15 Unlike K3 and K19, staining for K12 and MUC1 is restricted to cornea/limbus and conjunctiva, respectively To evaluate the expression of K12, MUC1, K3 and K19, sections from corneoscleral buttons comprising corneal, limbal, and conjunctival epithelia were analysed. As shown in figure 1, strong K3 expression was observed in the suprabasal layers of the conjunctiva (figure 1A), the limbus (figure 1B) and all the layers of the cornea (figure 1C), suggesting a lack of specificity of K3. On the contrary, K12 expression was restricted to the corneal epithelium (all layers) (figure 1F) and the suprabasal layers of the limbus (figure 1E) only (although

Figure 1 Expression of epithelial cell markers in corneoscleral buttons. Sections from corneoscleral buttons comprising conjunctiva, limbus and cornea were analysed to elucidate the expression pattern of keratin 3 (K3) (purple, AeC), keratin 12 (K12) (red, DeF), keratin 19 (K19) (yellow, GeI) and mucin 1 (MUC1) (green, LeN). Unlike K3/K19, K12 and MUC1 were specifically localised in the limbus/cornea and conjunctiva epithelia, respectively. Scale bar¼50 mm. (O) Whole corneoscleral button stained with antibodies against MUC1 and K12 allows to appreciate better the specificity of the two markers. Scale bar¼500 mm. 928

Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science some weak staining was observed in the basal cells). As opposed to K3, K12 was never observed in conjunctiva (figure 1D), confirming that K12 is a specific marker of the cornea.16 K19 was found expressed in the basal and suprabasal layers of all three epithelia (figure 1G,H,I). although expression levels in conjunctiva were slightly higher (figure 1G). In contrast, the expression of MUC1 was restricted to the superficial layers of the conjunctival epithelium (figure 1L) and no, or below threshold levels, staining was observed in the limbus (figure 1M) and cornea (figure 1N). The ability of the H-295 antibody to recognise all the MUC1 splice variants together with the relatively higher abundance of MUC1 in conjunctiva21 might explain the specific staining for the conjunctiva, despite studies reporting the expression of MUC1 also in the cornea.22 23

K12/MUC1 compared with K3/K19 on IC specimens obtained from the ocular surfaces of healthy donors To evaluate whether the results observed in corneoscleral buttons could be duplicated in IC specimens, samples were obtained from the ocular surfaces of healthy donors (n¼6). For each donor, six IC specimens were taken, three from each eye: two from conjunctiva, two from the limbus and two from central cornea. Samples were evaluated using K12/MUC1 markers and the results compared with those obtained with the K3/K19 markers. Only one membrane was used for the double staining of each pair of markers. As shown in figure 2, a distinct expression pattern was observed when the new markers were used: MUC1 was only

expressed in conjunctiva (figure 2A) and K12 in cornea (figure 2C). Overlapping signals were seen when K3 and K19 were used, thus suggesting lack of specificity (figure 2D,F). Interestingly, the limbus, being in the middle, showed separate and clearly distinct stainings of K12 and MUC1 (figure 2B, figure 3). The percentage of cells expressing K3/K19 and K12/MUC1 was evaluated through quantitative fluorescence immunohistochemistry (Q-FIHC) (table 1). Q-FIHC is an assay based on the use of laser scanning confocal microscope and advanced image analysis for the quantification of fluorescent intensity (FI) in tissues and cells.19 K3 and K19 were not highly specific, as the percentage of expression was not very different among the three epithelia of the ocular surface (K3: 64.3611.4% vs 70.768.9% vs 90.263.8%; K19: 74.766.1% vs 70.067.1% vs 52.7613.6%, respectively, in conjunctiva, limbus and cornea, n¼6). On the contrary, K12 was mainly expressed in the cornea (98.360.7% vs 2.261.3% in the conjunctiva) and MUC1 in the conjunctiva (97.861.3% vs 1.760.7% in the cornea). Expression of both K12 (38.765.6%) and MUC1 (61.365.6%) was nearly halved in the limbus, As expected, expression of both K12 (38.765.6%) and MUC1 (61.365.6%) was nearly halved in the limbus, the border between the cornea and the conjunctiva. An additional feature of Q-FIHC is to characterise the degree of overlap between two different fluorescent-tagged labels, thus allowing one to evaluate whether different cellular ‘targets’ are localised/expressed within the same cell.19 All FI values (expressed as pixel intensity) can be plotted onto scatter diagrams and found within three different

Figure 2 Impression cytology specimens taken from the conjunctiva, limbus and cornea of healthy donors (n¼6) and stained with fluorescentlabelled antibodies against keratin 3 (K3) (purple), keratin 19 (K19) (yellow), keratin 12 (K12) (red) and mucin 1 (MUC1) (green). An example is given for impression cytology specimens from one subject analysed for expression of K12/MUC1 (AeC) and K3/K19 (DeF). For each panel, staining is shown anticlockwise: K12 (or K3 in corresponding panels), MUC1 (or K19), merge and confocal microscope grid. K12 was expressed exclusively in the cornea and MUC1 in the conjunctiva. The limbus showed a distinct and separate staining of both markers. Overlapping and mixed signals were seen with K3 and K19 in all the three regions. Scale bar¼500 mm. Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152

929

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science Figure 3 Details of the limbus area from impression cytology samples of three different healthy donors demonstrating the high specificity of keratin 12 (K12) (red) and mucin 1 (MUC1) (green). Similar results could not be obtained with keratin 3 (K3) and keratin 19 (K19) (data not shown). Scale bar¼100 mm.

regions: in scatter region-1 or -2 if no coexpression is observed and in scatter region-3 when two markers are coexpressed. For K3/K19 markers, all pixels were found in scatter region-3, meaning that signals were colocalised within the same cell and/ or overlapping (figure 4A,D). As shown in table 1, all the subjects (n¼6) analysed with K3/K19 have ‘colocalisation coefficients’ close to 1, indicating their coincident expression within the same cell. In contrast for K12/MUC1 markers, FI values were shifted towards scatter region 1 and 2, thus meaning that cells are positive for either K12 or MUC1 (figure 4E,H). As shown in table 1, all the subjects (n¼6) analysed with K12/

MUC1 have ‘colocalisation coefficients’ close to 0, that is, no overlapping signals. These results support the conclusion that K12 and MUC1 are more reliable than K3 and K19 for the detection of cornea and conjunctiva in IC specimens.

K12/MUC1 are compared with K3/K19 on IC specimens obtained from patients with ocular surface disorders After validating our IC technique in healthy donors, we evaluated whether it was also suitable for patients with ocular surface

Figure 4 Evaluation of marker coexpression in impression cytology samples after quantitative fluorescence immunohistochemistry analysis. FI values (expressed as pixel intensity) from impression cytology samples stained for keratin 3 (K3) (B) and keratin 19 (K19) (C) or keratin 12 (K12) (F) and mucin 1 (MUC1) (G) were plotted onto scatter diagrams. For K3/K19, all pixels were found in scatter region 3 (A) meaning that markers were coexpressed within the same cell, and signals were overlapping, as clearly visible when signals for K3 and K19 are merged (D). For K12/MUC1, FI values were shifted towards scatter regions 1 and 2 (E), thus meaning no coexpression and higher specificity of the two markers, as is clearly visible when signals for K12 and MUC1 are merged (H). Scale bar¼100 mm. 930

Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science disorders. Patients (n¼12) with a diverse range of disorders (figure 5, table 2) were asked permission to retrieve specimens from their ocular surfaces through IC. As expected, patients with completely conjunctivalised corneas (figure 5A) showed IC specimens characterised by the presence of MUC1 and the disappearance of K12 (figure 5B). In other patients (figure 5D), cells positive for K12 were rare and only seen in the central part of the IC membrane (figure 5E) as observed when the conjunctiva invades the central cornea. Our IC technique is therefore particularly useful in those cases difficult to diagnose solely on the basis of ophthalmological evaluations or characterised by sectoral LSCD/conjunctivalisation (figure 5G,H). This would not have been feasible with K3/K19 markers. As shown in figure 5C,F,I, no clear-cut confirmation of conjunctivalisation could be obtained by IC solely. Signals were overlapping, and ‘colocalisation coefficients’ were close to 1 in all patients (table 2).

DISCUSSION IC is a minimally invasive technique, first reported in 1977,24 allowing ophthalmologists to evaluate rapidly the ‘health status’ of ocular surfaces. Assessment of this feature requires (1) specific markers of the ocular surface epithelia and (2) the expression of these markers in the uppermost layers of the ocular surfaces. The aim of our study was to select more reliable corneal and conjunctival markers to use in a new diagnostic test based on IC and quantification of fluorescent-labelled signals through Q-FIHC analysis.

The weakest point of IC-based techniques is the choice of inappropriate markers, thus leading to errors in diagnoses. Despite initial reports identifying K3/K19 pair as potential putative markers for corneal and conjunctival epithelia,15 these findings were no longer confirmed.16 17 Our study is further evidence of the inadequacy of these two markers, since neither K3 was specific for the cornea, nor was K19 for the conjunctiva. On the contrary, K12 and MUC1 were found to be more suitable for the evaluation of IC specimens, as overlapping was minimal, and expression was more specific. Restricted expression of K12 in corneal cells has already been reported,16 and the finding that K12 was more specific than K3 to identify corneal cells in IC specimens was therefore expected. For the conjunctiva, we chose a polyclonal antibody raised against the amino acids 961e1255 mapping at the C-terminus of the human MUC1 and able to recognise all the splice variants of MUC1. Expression was restricted to the conjunctiva in all our tests. Apparently, our results seem to be in contrast with other studies reporting the presence of MUC1 also in cornea.21e23 Our hypothesis is that, while this is true, MUC1 is relatively more abundant in the conjunctiva. When MUC1 was evaluated through real-time PCR analysis of mRNAs, expression in conjunctiva was always found to be higher than in cornea or lacrimal glands.21 In addition, when monoclonal antibodies (HMFG-1 and 139H2) directed against the core protein of MUC1 were tested,22 only HMFG-1 was found able to bind to the apical cells of the cornea. Mice null for MUC1 showed increased frequency of conjunctivitis and

Figure 5 Analysis of impression cytology specimens from patients with ocular surface disorders. Examples are shown for patients (Adpatient 3 from table 2) with completely conjunctivalised corneas, (Ddpatient 4 from table 2) with neovascularisation, but clear central cornea or (Gdpatient 2 from table 2) without apparently visible conjunctivalisation. Specimens were stained with antibodies against the keratin 12 (K12)/mucin 1 (MUC1) (B, E, H) or keratin 3 (K3)/keratin 19 (K19) (C, F, I) and signals quantified through quantitative fluorescence immunohistochemistry analysis. K3 (purple) and K19 (yellow) were not able to discriminate between cornea and conjunctiva. No overlapping signals were seen with K12 (red) and MUC1 (green). Scale bar¼500 mm. Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152

931

Downloaded from bjo.bmj.com on July 29, 2010 - Published by group.bmj.com

Laboratory science blepharitis, and no corneal alterations, suggesting a role of MUC1 in conjunctiva.25 For all these reasons, while the presence of MUC1 in the cornea is undeniable, its relative expression seems to be higher in the conjunctiva at both molecular and protein levels. The ability of the H-295 antibody to recognise all the MUC1 splice variants together with the relatively higher abundance of MUC1 in conjunctiva25 might have led to a greater specificity for the conjunctiva, as observed in our IC specimens. In conclusion, the use of K12 and H-295-againstMUC1 for the detection of cornea and conjunctiva has allowed us to set up what could be considered as a new diagnostic test to characterise the ‘normal cytology pattern’ of ocular surfaces. Our method, based on the use of K12/MUC1 markers and Q-FIHC analysis, has the following advantages: 1. It makes use of more reliable markers able to discriminate between corneal and conjunctival epithelium with no overlapping signals (unlike K3 and K19). 2. Double immunofluorescence labelling allows all information to be obtained from just one cytological specimen per eye, thus reducing the invasiveness of the technique. 3. Fluorescent signals can be quantitatively determined through a Q-FIHC analysis. It is evident how such a detailed IC report, together with a thorough ocular surface examination and medical anamnesis, could allow ophthalmologists to make accurate diagnoses. Our technique could become an important diagnostic tool for monitoring of patients following stem cell loss and transplantation, and for many disorders of the ocular surface.

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Acknowledgements This work was supported partly through a grant of the Italian Ministry of Health (Ricerca Finalizzata 2006 e “Colture organotipiche corneali e diagnosi nelle disabilita visive dovute a lesioni corneali profonde”) and of the Fondazione Cariverona (Bando Ricerca Scientifica 2008). Competing interests None.

20. 21.

Patients consent Obtained. Provenance and peer review Not commissioned; externally peer reviewed.

22. 23.

REFERENCES 1. 2.

932

Holly FJ, Lemp MA. Tear physiology and dry eye. Surv Ophthalmol 1977;22:69e87. Gipson IK. Anatomy of the conjunctiva, cornea, and limbus. In: Smolin G, Thoft RA, eds. The cornea. Boston: Little, Brown, 1994;3e24.

24. 25.

Akpek EK, Foster CS. Limbal stem cells transplantation. Int Ophthalmol Clin 1999;39:71e81. Kruse FE, Chen JJY, Tsai RJ, et al. Conjunctival transdifferentiation is due to the incomplete rmoval of limbal basal epithelium. Invest Ophthalmol Visual Sci 1990;31:1903e13. Huang AJW, Tseng SCG. Corneal epithelial wound healing in the absence of limbal epithelium. Invest Ophthalmol Visual Sci 1991;32:96e105. Dua HS, Azuara-Blanco A. Allo-limbal transplantation in patients with limbal stem cell deficiency. Br J Ophthalmol 1999;83:414e19. Haamann P, Jensem OM, Schimidt P. Limbal autograft transplantation. Acta Ophthalmol Scand 1998;76:117e18. 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;103:49e62. Dua HS. The conjuntival in corneal epithelial wound healing. Br J Ophthalmol 1998;82:1407e11. Tseng SC. Concept and application of limbal stem cells. Eye 1989;3:141e57. Tseng SCG, Sun TT. Stem cells: ocular surface maintenance. In: Corneal surgery: theory, technique, and brightbill FS tissue. St Louis, Missouri: Mosby-Year Book, Inc, 2000:9e18. Kenyon KR, Tseng SCG. Limbal autograft transplanatation for ocular surface disorders. Ophthalmology 1989;96:709e23. Pellegrini G, Traverso CE, Franzi AT, et al. Long-term restoration of damaged corneal surfaces with autologous cultivated corneal epithelium. Lancet 1997;349:990e3. Rama P, Bonini S, Lambiase A, et al. Autologous fibrin-cultured limbal stem cells permanently restore the corneal surface of patients with total limbal stem cell deficiency. Transplantation 2001;72:1478e85. Donisi PM, Rama P, Fasolo A, et al. Analysis of limbal stem cell deficiency by corneal impression cytology. Cornea 2003;22:533e8. Yoshida S, Shimmura S, Kawakita T, et al. Cytokeratin 15 can be used to identify the limbal phenotype in normal and diseased ocular surface. Invest Ophthalmol Vis Sci 2006;47:4780e6. Chen WY, Mui MM, Kao WW, et al. Conjunctival epithelial cells do not transdifferentiate in organotypic cultures: expression of K12 keratin is restricted to corneal epithelium. Curr Eye Res 1994;13:765e78. Di Iorio E, Barbaro V, Ruzza A, et al. Isoforms of DNp63 and the migration of ocular limbal cells in human corneal regeneration. Proc Natl Acad Sci 2005;102:9523e8. Di Iorio E, Barbaro V, Ferrari S, et al. Quantification of fluorescent immunohistochemistry to analyse p63 isoforms and cell cycle phases in human limbal stem cells. Microsc Res Tech 2006;69:983e91. Engelmann K, Kinlough CL, Mu¨ller S, et al. Transmembrane and secreted MUC1 probes show trafficking-dependent changes in O-glycan core profiles. Glycobiology 2005;15:1111e24. Imbert Y, Darling DS, Jumblatt MM, et al. MUC1 splice variants in human ocular surface tissues: possible differences between dry eye patients and normal controls. Exp Eye Res 2006;83:493e501. Inatomi T, SpurrpMichaud SJ, Tisdale AS, et al. Human corneal and conjunctival epithelia express MUC1 mucin. Invest Ophthalmol Vis Sci 1995;36:1818e27. Gipson IK, Argu¨eso P. Role of mucins in the function of the corneal and conjunctival epithelia. Int Rev Cytol 2003;231:1e49. Egbert PR, Laube S, Maurice DM. A simple conjunctival biopsy. Am J Ophthalmol 1977;84:798e801. Kardon R, Price RE, Julian J, et al. Bacterial conjunctivitis in Muc1 null mice. Invest Ophthalmol Vis Sci 1999;40:1328e35.

Br J Ophthalmol 2010;94:926e932. doi:10.1136/bjo.2009.164152