2.1

Specimens

Fifty-one formalin-fixed and paraffin-embedded oral biopsies were selected from the Department of Pathology, School of Medicine of Ribeirão Preto, University of São Paulo, archives between 1992 and 2005; and their use was approved by the local Institutional Ethical Committee on human experimentation (protocol 10108/2004). Tumor grades were evaluated histologically by experienced pathologists in tissues stained with hematoxylin and eosin. All samples were graded according to the oral lesion criteria recently proposed by van der Waal and Axéll (2002).

Patients were stratified into 3 groups according to the type of lesion. The first group consisted of 18 of 51, (35.2%) biopsies of oral benign lesions. The second group consisted of 16 of 51, (31.3%) biopsies of oral premalignant lesions and 17 of 51 (33.3%) biopsies of oral squamous cell carcinoma (OSCC), malignant lesions. Five-micrometer sections were cut, placed on organosilane-pretreated slides, and submitted to immunohistochemical assays (HLA-G). In addition, a 10- μ m section was cut for DNA extraction and HPV typing.

2.2

Immunohistochemistry (HLA-G)

A total of 51 formalin-fixed and paraffin-embedded biopsies of oral lesions were analyzed. Immunohistochemistry using the streptavidin-biotin system (EP-USA/500, Signet) was performed for the detection of HLA-G antigens. The tissue specimens were dewaxed in xylene, rehydrated in graded alcohol, and rinsed in water. For antigen retrieval, the sections were immersed in 10 mmol/L sodium citrate buffer (pH 6.0). Endogenous peroxidase blocking was performed using immersions in a hydrogen peroxide bath in absolute methanol (15 minutes each change), and nonspecific binding was performed with 3% low-fat dried milk diluted 1:100 in phosphate-buffered saline (PBS). Slides were incubated with the primary monoclonal antibody for HLA-G (diluted 1:500; clone 87G) in a humidified chamber at 4°C overnight and then incubated with the streptavidin-peroxidase complex at 37°C for 30 minutes. The sections were then incubated in a solution containing 5 mg of diaminobenzidine (GIBCO, Gaithersburg, MD) dissolved in 5 mL of PBS and 100 μ L of fresh peroxidase solution (450 μ L of PBS and 50 μ L of hydrogen peroxide) for 10 minutes, lightly counterstained with Carrazzi hematoxylin without acid for 60 seconds, exhaustively rewashed with tap water, air dried, and mounted with Permount mounting medium (MERCK, Darmstadt, Germany).

All scoring and qualitative interpretations of the immunohistochemical results were carried out by an experienced pathologist; and the results were classified as negative when the cytoplasm was immunolabelling (≤25%), and positive (immunolabeling <25% until 100%). The mAb 87G (HLA-G1 and HLA-G5 antigen-specific HLA-G) was kindly provided by Dr Edgardo D Carosella from the Service de Recherche en Hémato-Immunologie, Hôpital Saint Louis, Institut Universitaire d’Hématologie, Paris, France. This specific clone has served as a probe in several studies. It is able to recognize both the membrane-bound and soluble forms of the HLA-G antigen . The positive control was a section of trophoblastic tissue from a third-trimester human placenta. The negative control was the same human tissue used as positive control, in which the primary antibody was omitted from the assay and replaced with PBS.

2.3

Image acquisition and quantification of HLA-G by image analysis

Positive cytoplasms were automatically quantified with a computer-assisted system (Image-Pro Plus; Cybernetics, Bethesda, MD) consisting of a microscope, a digital camera, and a software package. A mean of 10 random microscope fields was selected to analyze 1000 cytoplasm fields per biopsy in all patient sections. Image acquisition of 79 specimens was subsequently stratified into oral benign lesions, oral premalignant lesions, and malignant lesions. This process was performed on an electron photomicrograph, and the images were then processed and analyzed by the software. For each slide, the digitized image segmentation was controlled interactively by the red-green-blue color filter existing in the software program. The automatic cytoplasm count was determined and expressed as percentage.

2.4

HPV detection and typing

Genomic DNA was obtained from 2 sections of paraffin blocks (10 μ m each) according to the protocol proposed by Bettini et al , modified, and amplified by polymerase chain reaction (PCR) using the GP5+ and GP6+ primers , which amplify a 150–base pair DNA fragment and which were used for generic HPV amplification. Amplification was performed together with a set of primers for a β -globin (PCO3 and PCO4) gene (PCO3+ CTT CTG ACA CAA CTG TGT TCA CTA GC and PCO4+ TCA CCA CAA CTT CAT CCA CGT TCA CC) as an internal control of amplification. The generic HPV-positive samples were amplified with specific primers for HPV16, HPV18, HPV31, HPV33, HPV6, and HPV11 as described by Walboomers et al . The amplification procedure for HPV detection and typing was carried out using a final volume of 25 μ L. The reaction mixture contained 20 ng of genomic DNA, 0.20 mmol/L deoxynucleoside triphosphate (Pharmacia, Uppsala, Sweden), 0.6 μ mol/L of each primer (IDT, IA), 1.25 units Taq polymerase (Invitrogen, Brazil), 3 mmol/L MgCl ₂ , and 1x buffer (Invitrogen). The cycling conditions for almost all generic and HPV types consisted of an initial denaturation step at 94°C for 5 minutes, followed by 35 cycles at 94°C for 1 minute, at 55°C for 1 minute, and at 72°C for 1 minute, with a final extension cycle at 72°C for 10 minute. An annealing temperature of 60°C was used for HPV11 amplification. The PCR products were separated on 10% nondenaturing polyacrylamide gels followed by silver staining, as described by Sanguinetti et al .

2.5

HPV controls

The HPV-positive control and β -globin control were cervical samples collected by cytologic cytobrushing. As a negative control, all PCR reagents were added to an Eppendorf tube containing no DNA sample. For HPV6 and HPV 11 typing, the PCR controls were HPV-positive DNA collected from cervical samples. For HPV16 and HPV18 controls, DNA was extracted from SiHa and HeLa cell lines, respectively, whereas HPV31 and HPV33 were from patient samples. All controls (HPV16, 18, 31, and 33) were kindly provided by Dr Luisa Lina Villa, Ludwig Institute for Cancer Research, São Paulo, Brazil.

2.6

Statistical analysis

All statistical analyses were carried out using the GraphPad Instat software (San Diego, CA). Correlations between the HLA-G staining and the various groups of oral lesions were calculated by the χ ² test for independence. The Fisher exact test was applied to compare the levels of HLA-G molecules and HPV-infected oral lesions. One-way analysis of variance with Tukey posttest for multiple comparisons was applied to the quantitative analysis of HLA-G and oral lesions. P values were 2-sided, and the level of significance was set at > .05.

2.1

Specimens

Fifty-one formalin-fixed and paraffin-embedded oral biopsies were selected from the Department of Pathology, School of Medicine of Ribeirão Preto, University of São Paulo, archives between 1992 and 2005; and their use was approved by the local Institutional Ethical Committee on human experimentation (protocol 10108/2004). Tumor grades were evaluated histologically by experienced pathologists in tissues stained with hematoxylin and eosin. All samples were graded according to the oral lesion criteria recently proposed by van der Waal and Axéll (2002).

Patients were stratified into 3 groups according to the type of lesion. The first group consisted of 18 of 51, (35.2%) biopsies of oral benign lesions. The second group consisted of 16 of 51, (31.3%) biopsies of oral premalignant lesions and 17 of 51 (33.3%) biopsies of oral squamous cell carcinoma (OSCC), malignant lesions. Five-micrometer sections were cut, placed on organosilane-pretreated slides, and submitted to immunohistochemical assays (HLA-G). In addition, a 10- μ m section was cut for DNA extraction and HPV typing.

2.2

Immunohistochemistry (HLA-G)

A total of 51 formalin-fixed and paraffin-embedded biopsies of oral lesions were analyzed. Immunohistochemistry using the streptavidin-biotin system (EP-USA/500, Signet) was performed for the detection of HLA-G antigens. The tissue specimens were dewaxed in xylene, rehydrated in graded alcohol, and rinsed in water. For antigen retrieval, the sections were immersed in 10 mmol/L sodium citrate buffer (pH 6.0). Endogenous peroxidase blocking was performed using immersions in a hydrogen peroxide bath in absolute methanol (15 minutes each change), and nonspecific binding was performed with 3% low-fat dried milk diluted 1:100 in phosphate-buffered saline (PBS). Slides were incubated with the primary monoclonal antibody for HLA-G (diluted 1:500; clone 87G) in a humidified chamber at 4°C overnight and then incubated with the streptavidin-peroxidase complex at 37°C for 30 minutes. The sections were then incubated in a solution containing 5 mg of diaminobenzidine (GIBCO, Gaithersburg, MD) dissolved in 5 mL of PBS and 100 μ L of fresh peroxidase solution (450 μ L of PBS and 50 μ L of hydrogen peroxide) for 10 minutes, lightly counterstained with Carrazzi hematoxylin without acid for 60 seconds, exhaustively rewashed with tap water, air dried, and mounted with Permount mounting medium (MERCK, Darmstadt, Germany).

All scoring and qualitative interpretations of the immunohistochemical results were carried out by an experienced pathologist; and the results were classified as negative when the cytoplasm was immunolabelling (≤25%), and positive (immunolabeling <25% until 100%). The mAb 87G (HLA-G1 and HLA-G5 antigen-specific HLA-G) was kindly provided by Dr Edgardo D Carosella from the Service de Recherche en Hémato-Immunologie, Hôpital Saint Louis, Institut Universitaire d’Hématologie, Paris, France. This specific clone has served as a probe in several studies. It is able to recognize both the membrane-bound and soluble forms of the HLA-G antigen . The positive control was a section of trophoblastic tissue from a third-trimester human placenta. The negative control was the same human tissue used as positive control, in which the primary antibody was omitted from the assay and replaced with PBS.

2.3

Image acquisition and quantification of HLA-G by image analysis

Positive cytoplasms were automatically quantified with a computer-assisted system (Image-Pro Plus; Cybernetics, Bethesda, MD) consisting of a microscope, a digital camera, and a software package. A mean of 10 random microscope fields was selected to analyze 1000 cytoplasm fields per biopsy in all patient sections. Image acquisition of 79 specimens was subsequently stratified into oral benign lesions, oral premalignant lesions, and malignant lesions. This process was performed on an electron photomicrograph, and the images were then processed and analyzed by the software. For each slide, the digitized image segmentation was controlled interactively by the red-green-blue color filter existing in the software program. The automatic cytoplasm count was determined and expressed as percentage.

2.4

HPV detection and typing

Genomic DNA was obtained from 2 sections of paraffin blocks (10 μ m each) according to the protocol proposed by Bettini et al , modified, and amplified by polymerase chain reaction (PCR) using the GP5+ and GP6+ primers , which amplify a 150–base pair DNA fragment and which were used for generic HPV amplification. Amplification was performed together with a set of primers for a β -globin (PCO3 and PCO4) gene (PCO3+ CTT CTG ACA CAA CTG TGT TCA CTA GC and PCO4+ TCA CCA CAA CTT CAT CCA CGT TCA CC) as an internal control of amplification. The generic HPV-positive samples were amplified with specific primers for HPV16, HPV18, HPV31, HPV33, HPV6, and HPV11 as described by Walboomers et al . The amplification procedure for HPV detection and typing was carried out using a final volume of 25 μ L. The reaction mixture contained 20 ng of genomic DNA, 0.20 mmol/L deoxynucleoside triphosphate (Pharmacia, Uppsala, Sweden), 0.6 μ mol/L of each primer (IDT, IA), 1.25 units Taq polymerase (Invitrogen, Brazil), 3 mmol/L MgCl ₂ , and 1x buffer (Invitrogen). The cycling conditions for almost all generic and HPV types consisted of an initial denaturation step at 94°C for 5 minutes, followed by 35 cycles at 94°C for 1 minute, at 55°C for 1 minute, and at 72°C for 1 minute, with a final extension cycle at 72°C for 10 minute. An annealing temperature of 60°C was used for HPV11 amplification. The PCR products were separated on 10% nondenaturing polyacrylamide gels followed by silver staining, as described by Sanguinetti et al .

2.5

HPV controls

The HPV-positive control and β -globin control were cervical samples collected by cytologic cytobrushing. As a negative control, all PCR reagents were added to an Eppendorf tube containing no DNA sample. For HPV6 and HPV 11 typing, the PCR controls were HPV-positive DNA collected from cervical samples. For HPV16 and HPV18 controls, DNA was extracted from SiHa and HeLa cell lines, respectively, whereas HPV31 and HPV33 were from patient samples. All controls (HPV16, 18, 31, and 33) were kindly provided by Dr Luisa Lina Villa, Ludwig Institute for Cancer Research, São Paulo, Brazil.

2.6

Statistical analysis

All statistical analyses were carried out using the GraphPad Instat software (San Diego, CA). Correlations between the HLA-G staining and the various groups of oral lesions were calculated by the χ ² test for independence. The Fisher exact test was applied to compare the levels of HLA-G molecules and HPV-infected oral lesions. One-way analysis of variance with Tukey posttest for multiple comparisons was applied to the quantitative analysis of HLA-G and oral lesions. P values were 2-sided, and the level of significance was set at > .05.