Patient and control group selection

Sixty-four patients with OSSN (44 SCC, 20 CIN) and 15 controls (conjunctival epithelialectomy tissues from limbal stem cell deficiency patients) from the years 2008–10 were enrolled in this study and approval was obtained from the institute ethics committee. After obtaining written consent fresh tissues from cases and controls were stored at −70°C for molecular studies, and the remaining tissue was fixed in 10% neutral buffered formalin and embedded in paraffin wax for histopathological examination and immunohistochemical analysis.

Clinical features including hours of exposure to solar ultraviolet irradiation and the presence of predisposing conditions such as xeroderma pigmentosum were recorded. Radiological details were also recorded.

Haematoxylin-eosin stained sections were examined to confirm the diagnosis of OSSN and the cases were classified as CIN or SCC. The CIN cases were further classified as mild, moderate or severe dysplasia based on the extent of epithelial involvement, and SCC cases were classified as well or poorly differentiated depending on the extent of keratinisation.16

Detection of HPV by multiplex PCR

DNA extraction from all 64 fresh tissue specimens was performed using the QIAamp DNA mini kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. The extracted DNA was quantified using a Nanodrop spectrophotometer (Nanodrop Technologies, Wilmington, Delaware, USA). The quality of extracted DNA was determined by β-globin-specific PCR using GH20/GH21 primer. Briefly, a 40 ng/µl aliquot of genomic DNA was added as a template to a PCR master mixture containing 1× PCR buffer (100 Mm Tris-HCl, 500 Mm KCl, pH 8.3), 200 µM each deoxynucleoside triphosphate, 200 nM primer, 1.5 mM MgCl₂ and 2.5 U of Ampli Taq Gold DNA polymerase. PCR was performed using the following conditions: 10 min at 95°C, followed by 30 s at 95°C, 1 min at 53°C, 1 min at 72°C and 10 min at 72°C for 35 cycles. DNA extracted from all 64 samples and 15 controls was amplified using β-globin, which showed a product of 450 bp, confirming the quality of DNA to be suitable for PCR.

The HPV-L1 capsid gene-specific consensus primer, PGMY09/11, which is an improvement for quality assurance was used for HPV detection. This primer detects all the major HPV types. The PCR assay was performed using a 40 ng/µl aliquot of DNA in master mixture containing 1× PCR buffer, 200 µM each deoxynucleoside triphosphate, 200 nM primer, 4.0 mM MgCl₂ and 7.5 U of Ampli Taq Gold DNA polymerase. Forty cycles of amplification were performed using the following conditions: 9 min at 94°C, followed by 1 min at 94°C, 1 min at 55°C and 1 min at 72°C.

For each reaction, DNA extracted from CaSki (HPV16) and HeLa (HPV18) cell lines served as positive and C33A (HPV negative) cell line and deionised water instead of the DNA template served as a negative control for HPV-specific PCR assay.

HPV genotyping by linear array

Genotyping of positive samples was done using the HPV linear array kit (Roche Molecular System, Branchburg, New Jersey, USA) following the manufacturer's instructions. Briefly, DNA was isolated and for each reaction, 20 µl of extracted DNA was subjected to PCR amplification using the HPV consensus PCR (Roche Molecular Systems Inc., Alameda, California, USA).This method utilises biotinylated primers (PGMYO9/PGMY11) that amplify a 450 bp fragment of the L1 open reading frame of 37 HPV genotypes. A control set of biotinylated primers simultaneously amplifies a 268 bp fragment of the human β-globin gene in each reaction. PCR products were then denatured and hybridised to separate HPV genotyping strips containing an immobilised array of 37 different HPV probes. Hybridised PCR products were visualised by using a streptavidin-horseradish peroxidase detection system. Negative and positive controls (provided in the kit) were run with each batch of specimens tested.

P16^INK4a immunohistochemistry

Immunohistochemistry for p16^INK4a was performed on 64 cases and 15 control conjunctival tissues. Unstained sections of 4–5 μm thickness were cut on polylisine-coated slides from the formalin-fixed paraffin-embedded tissues and subjected to immunohistochemical analysis using the avidin-biotin indirect method. Briefly, sections were deparaffinised in xylene, dehydrated and subjected to heat antigen retrieval in citrate buffer (pH 6.0) for 20 min. Endogenous peroxidase activity was blocked by 3% H₂O₂. The slides were incubated with primary mouse IgG2b against p16^INK4a (clone 6H12: prokaryotic recombinant fusion protein corresponding to the entire human p16^INK4a molecule) obtained from Novocastra Laboratories Ltd. (UK) at a dilution of 1 : 50 in a humidified chamber at 4°C overnight, followed by incubations with biotinylated-linked secondary antibody and peroxidase labelled streptavidin (Dako Cytomation LSAB+System-HRP kit). The sections were visualised using 3, 3′-diaminobenzidene peroxidase substrate for 3–4 min, then counterstained with haematoxylin and examined by light microscopy. p16^INK4a-positive cervical SCC served as a positive control. Primary antibody was replaced by Tris-buffered saline in the negative control. Immunoreactivity was considered positive when more than 10% of the cells showed nuclear as well as cytoplasmic positivity.17 ,18

Statistical analysis

Statistical analysis was performed with Windows SPSS V.11.0. The relationship between HPV, p16^INK4a status and clinicopathological variables such as age, sex, laterality, surgical intervention, histopathological differentiation, and tumour size, orbital and intraocular invasion was evaluated using a contingency table, and further analysed using Fisher's exact test.

p16^INK4a Immunostaining was evaluated in terms of sensitivity, specificity and predictive value of the test for detecting HPV. The κ statistics was used to assess the degree of agreement between p16^INK4a and HPV tests. κ Values between 0.40 and 0.75 are regarded as fair to good agreement beyond chance, values less than 0.40 as poor agreement beyond chance, and values greater than 0.75 as excellent agreement beyond chance. Disease-free survival curves were constructed using the Kaplan–Meier method and compared using the log-rank test.