SPECIMENS

Thirty human gastric adenocarcinomas of disparate pathological stages (table 1) were collected after surgical resections performed at the Mercy Hospital, Cork, following a protocol approved by the University Teaching Hospitals ethics committee. None of the patients had received chemo-, radio- or immuno-therapy before resection. Tumours were moderately (n = 11), poorly (n = 15), or moderately to poorly (n = 4) differentiated (table 2). Of the poorly differentiated tumours, five were Signet ring cell carcinomas.

IMMUNOHISTOCHEMICAL DETECTION OF FasL PROTEIN

Paraffin wax embedded surgically resected tumour sections were deparaffinised in xylene and rehydrated before analysis. The slides were washed twice for five minutes in a wash buffer containing 50 mM Tris/HCl, pH 7.6, 50 mM NaCl, and 0.001% saponin. Endogenous peroxidase was quenched with 3.0% hydrogen peroxide in methanol for five minutes. The slides were then washed as before except that the wash buffer for this and all subsequent steps included 1% normal goat serum. They were then blocked for one hour in wash buffer containing 5% normal goat serum. This was followed by a further wash and incubation overnight at 4°C with affinity purified rabbit polyclonal anti-human FasL specific IgG (Santa Cruz Biotechnology, Santa Cruz, California, USA) at 0.1 μg/ml in wash buffer. Antibody binding was localised using a biotinylated secondary antibody, avidin conjugated horseradish peroxidase, and diaminobenzidine substrate, contained within the Vectastain ABC detection kit (Vector Laboratories, Burlingame, California, USA). The immunising peptide to which the antibody was raised (FasL: N-terminal amino acids 260–279; Santa Cruz Biotechnology) was included at 1 μg/ml during primary antibody incubation as a direct internal competitive control for antibody specificity. In a separate control for specificity of the Santa Cruz FasL specific polyclonal antibody, we previously found that specific antisense oligonucleotide mediated inhibition of FasL expression14 converted SW620 cells from positive to negative by immunohistochemistry with this antibody. The FasL specificity of the Santa Cruz antibody has also been previously verified in immunohistochemistry by others, staining FasL mRNA positive but not FasL mRNA negative cell lines,19 and staining in concordance with other FasL specific monoclonal17 and polyclonal19 antibodies. FasL detection in the gastric tumours was confirmed by immunohistochemistry with a FasL specific monoclonal antibody (Pharmingen, San Diego, California, USA; clone G247–4) as described above. The monoclonal antibody was used at a concentration of 5 μg/ml, and an isotype matched control antibody was also used. Slides were counterstained with haematoxylin.

GENERATION OF A FasL SPECIFIC RNA PROBE (RIBOPROBE)

A digoxigenin labelled RNA hybridisation probe (344 bp) was generated corresponding to codons 96–210 of the human FasL cDNA sequence. The riboprobe was synthesised by in vitro transcription using digoxygenin-11-UTP and T7 RNA polymerase (Boehringer-Mannheim GmbH, Mannheim, Germany). Template for the in vitro transcription reaction was generated by polymerase chain reaction (PCR) amplification of a fragment (codons 96–210) of FasL cDNA using a proofreading thermostable polymerase (UlTma DNA polymerase; Perkin-Elmer, Norwalk, Connecticut, USA) and an antisense primer to which a T7 promoter sequence was added.

RNA was isolated from FasL expressing cells by homogenisation in guanidine thiocyanate (Sigma Chemical Co, St Louis, Missouri, USA) followed by phenol extraction and ethanol precipitation. cDNA was synthesised using the avian myeloblastosis virus reverse transcriptase (Promega, Madison, Wisconsin, USA) and random hexanucleotide primers (Boehringer-Mannheim).

PCR was performed on the cDNA using the following sense and antisense primers respectively: FasL: dGGATTGGGCCTGGGGATGTTTCA and d[pT7]-TTGTGGCTCAGG GGCAGGTTGTTG. PCR primers were designed using the DNASTAR Lasergene Primerselect program (DNASTAR Inc, Madison, Wisconsin, USA). Primer pairs were chosen to span introns in the FasL genomic sequence, thus ensuring mRNA specific amplification. Primers that showed no significant homology to any other genes in the EMBL DNA sequence database were selected.

Thermal cycling (40 cycles) was as follows: denaturation at 96°C for 15 seconds; annealing at 55°C for 30 seconds; extension at 72°C for three minutes. Primers were used at a final concentration of 0.1 μM each, dNTPs at 50 μM, and MgCl₂ at 1.5 mM. A total of 1.0 U of UlTma DNA polymerase was used per 50 μl reaction. PCR product specificity was confirmed by restriction mapping.

This PCR amplified FasL cDNA fragment was used as template to synthesise a riboprobe by in vitro transcription using digoxygenin-11-UTP and T7 RNA polymerase (Boehringer-Mannheim) according to the manufacturer’s instructions. The nucleotide sequence of the FasL probe showed no significant homology to any other sequence in the EMBL DNA sequence database. An unlabelled riboprobe was also synthesised for use in competitive control hybridisations.

LOCALISATION OF FasL mRNA EXPRESSION BY IN SITU HYBRIDISATION

In situ hybridisation was performed on paraffin wax embedded human gastric tumour sections (4 μm thick), mounted on aminopropylethoxysilane treated slides. Prehybridisation treatments involved two washes of five minutes each in (a) phosphate buffered saline (PBS), (b) PBS/0.1 M glycine, (c) PBS/0.3% Triton X-100, and (d) PBS again. Sections were digested for 30 minutes at 37°C with proteinase K (10 μg/ml in 100 mM Tris/HCl/50 mM EDTA, pH 8.0), fixed for five minutes at 4°C in 4% paraformaldehyde/PBS and then acetylated twice for five minutes in fresh 0.25% acetic anhydride/0.1 M triethanolamine, pH 8.0. Sections were incubated at 37°C for 10 minutes in a prehybridisation buffer consisting of 50% deionised formamide in 4 × SSC (where 1 × SSC is 0.15 M NaCl/0.015 M sodium citrate). Hybridisation was performed at 42°C overnight in hybridisation buffer (50% formamide, 10% dextran sulphate, 1 × Denhardt’s reagent, 4 × SSC, 10 mM dithiothreitol, 500 μg/ml yeast tRNA, and 1 mg/ml heat denatured herring sperm DNA) containing 1 ng/μl digoxigenin labelled riboprobe. After hybridisation, tissues were washed with increasing stringency to 0.1 × SSC at 37°C. Hybridised probe was detected immunologically using alkaline phosphatase conjugated sheep anti-digoxigenin antibody (Boehringer-Mannheim) and visualised with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) (purple/black precipitating product). Control slides involved direct internal competitive inhibition of hybridisation.16 ,21This was performed exactly as described above except that a tenfold excess of unlabelled FasL specific riboprobe was added to the digoxigenin labelled FasL specific riboprobe before hybridisation. This resulted in direct competitive displacement of positive hybridisation leading to a marked reduction in signal intensity, confirming the FasL specificity of hybridisation.

LEUCOCYTE COMMON ANTIGEN (CD45) STAINING

To identify TILs, CD45 staining was performed on consecutive paraffin wax embedded sections. After deparaffinisation and rehydration, sections were pretreated by microwave irradiation in 0.01 M citrate buffer for five minutes at 370 W. The sections were rapidly cooled by immersion in 0.1 M PBS. The slides were then incubated with a mouse anti-human CD45 monoclonal IgG (clone 2B11+PD7/26; Dako Corp, Carpinteria, California, USA) at a dilution of 1:70 for one hour. All incubations were carried out at room temperature. Next the slides were washed for five minutes in 0.1 M Tris buffered saline which was used for all washes. After incubation with a secondary rabbit anti-mouse IgG (Dako Corp) at a dilution of 1:25 for 30 minutes, washing was repeated. Alkaline phosphatase conjugated anti-alkaline phosphatase (APAAP) complex (DO651; Dako Corp), at a dilution of 1:50, was incubated on the slides for 30 minutes and they were washed again. To enhance staining, the secondary antibody incubation and APAAP complex incubation were repeated once as above, but the duration of the incubation steps was reduced to 10 minutes each. Alkaline phosphatase substrate solution was then incubated on the slides for 10 minutes. The phosphatase substrate used was Fast Red (Sigma Chemical Co, St Louis, MO, USA). Slides were counterstained with haematoxylin and viewed by light microscopy.

CELL DEATH DETECTION IN SITU BY TERMINAL DEOXYNUCLEOTIDYL TRANSFERASE (TdT) MEDIATED dUTP NICK END LABELLING (TUNEL)

Cell death was detected in situ in resected tissues by enzymic labelling of DNA strand breaks using a TUNEL assay (Boehringer-Mannheim) according to the manufacturer’s instructions. Paraffin wax embedded surgically resected tumour sections were deparaffinised in xylene and rehydrated before analysis. After treatment with proteinase K (20 μg/ml in 10 mM Tris/HCl, pH 7.6) for 30 minutes, sections were washed in PBS and endogenous peroxidase activity was blocked in 0.3% hydrogen peroxide in methanol for 30 minutes. Next, the slides were washed. This and all subsequent washes were in PBS. The sections were treated with permeabilisation solution (0.1% Triton X-100 in 0.1% sodium citrate) for 15 minutes. After a wash, the labelling reaction was performed using a solution containing TdT, its buffer and fluorescein conjugated dUTP. During this step, slides were coverslipped and incubated at 37°C for 60 minutes in a humidity chamber. TdT was omitted from negative control slides. To localise cells containing labelled DNA strand breaks, sections were washed and incubated with a sheep anti-fluorescein antibody Fab fragment conjugated with horseradish peroxidase at 37°C in a humidity chamber for 30 minutes. After a wash, colour was developed by incubating sections with a diaminobenzidine substrate solution for 15 minutes. After a final wash, sections were viewed by light microscopy. Only those cells with positive TUNEL staining and of apoptotic morphology were considered apoptotic.

EXPRESSION OF FasL PROTEIN BY GASTRIC ADENOCARCINOMAS

With the use of an affinity purified FasL specific rabbit polyclonal IgG (Santa Cruz Biotechnology) raised against a synthetic FasL peptide (N-terminal amino acids 260–279), FasL expression by tumour cells was immunohistochemically detected in all (30/30) surgically resected gastric adenocarcinomas (fig 1A). FasL specificity was confirmed using the FasL peptide immunogen (FasL amino acids 260–279) as an internal competitive control. Inclusion of the soluble peptide immunogen during immunohistochemistry resulted in direct competitive displacement of positive staining in consecutive control sections (fig 1B). The FasL specificity of the Santa Cruz polyclonal antibody used (C-20) had been previously verified by us14and others.17 ,19 FasL detection in the gastric tumours was confirmed using immunohistochemistry with a FasL specific monoclonal antibody (Pharmingen; FasL clone G247–4), which resulted in a pattern of staining identical with that detected with the Santa Cruz polyclonal antibody (not shown). An isotype-matched monoclonal antibody did not stain in control sections.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Expression of Fas ligand (FasL) in human gastric adenocarcinomas. Immunoperoxidase staining using a FasL specific rabbit polyclonal IgG was performed on paraffin wax embedded gastric carcinoma sections. Slides were counterstained with haematoxylin. (A) FasL positive immunohistochemical staining (brown) is shown in a representative gastric adenocarcinoma. (B) As a control for specificity of antibody detection, the FasL immunising peptide was included during primary antibody incubation. Competitive displacement of staining by the immunising peptide confirms FasL specificity. Tumour sections sequential to those shown in (A) and (B) were used to detect FasL mRNA by in situ hybridisation, using a digoxigenin labelled FasL specific riboprobe. (C) Positive purple hybridisation signal was obtained from the tumour area that stained immunohistochemically positive for FasL protein. (D) In a control hybridisation, a tenfold excess of unlabelled probe caused direct competitive displacement of the labelled probe, confirming the specificity of hybridisation. These results are representative of 30 adenocarcinomas of the stomach.

Positive staining of neoplastic tissue varied in intensity and extent, both within individual tumours and between tumours. Intensity of staining varied from weakly positive neoplastic areas to intensely staining tumour regions, where staining was stronger than that observed in local FasL positive TILs. FasL staining intensity was locally uniform within nests of tumour cells. FasL positive and negative staining neoplastic areas were frequently found to occur within the same tumour. However, all tumours examined were predominantly FasL positive (>70% of tumour area).

LOCALISATION OF FasL mRNA TO NEOPLASTIC GASTRIC EPITHELIAL CELLS

A digoxigenin labelled FasL specific riboprobe was synthesised using the T7 RNA polymerase and a PCR generated cDNA template. The nucleotide sequence of the FasL riboprobe showed no significant homology to any other sequence within the EMBL DNA sequence database. Using in situ hybridisation with this probe, FasL mRNA expression was detected in tumour cells in resected gastric adenocarcinomas. Areas of positive hybridisation occurred within neoplastic cells throughout large areas of the tumours (fig 1C). The pattern of FasL mRNA detection corresponded closely to that of FasL immunostaining performed on consecutive tumour sections. Co-localisation of FasL mRNA and protein confirmed that the transformed gastric epithelial cells expressed FasL. Hybridisation was also detected in cells of lymphoid morphology—that is, TILs. Hybridisation specificity was confirmed by adding an excess of unlabelled FasL specific riboprobe to the digoxigenin labelled FasL specific riboprobe as an internal competitive control during hybridisation. This resulted in direct competitive displacement of positive hybridisation signals in consecutive control sections (fig1D).

APOPTOSIS OF TILs

CD45 immunohistochemistry showed immunocyte infiltration in all 30 carcinomas (fig 2B). Tonsil sections included as positive controls stained strongly, and omission of the primary antibody abolished staining. Most of the CD45 positive cells were of lymphoid morphology. Recently, FasL expression by human oesophageal carcinoma was found to be associated with increased cell death of TILs.20Apoptosis was detected by TUNEL among TILs adjacent to FasL positive areas of gastric carcinomas. These TUNEL positive TILs exhibited morphological features of apoptosis, including nuclear condensation and fragmentation (fig 2C). This was a consistent finding in all the tumours examined (n = 8).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Apoptotic lymphocytes adjacent to FasL expressing carcinoma. (A) FasL expression (brown) detected in an area of gastric carcinoma by immunoperoxidase staining. (B) CD45 (leucocyte common antigen) immunophosphatase (alkaline phosphatase conjugated anti-alkaline phosphatase (APAAP)) staining (red) was performed on a sequential section. CD45 positive cells (red) of lymphoid morphology are present adjacent to areas of carcinoma. Slides were counterstained with haematoxylin. Isotype matched control sections were negative (not shown). (C) Cell death detection in situ by terminal deoxynucleotidyl transferase mediated dUTP nick end labelling (TUNEL). Tumour sections sequential to those shown in (A) and (B) were used to detect cell death by enzymic labelling of DNA strand breaks using TUNEL. Only those cells with positive TUNEL staining (brown) and exhibiting apoptotic morphology were considered apoptotic. Control sections (not shown) where the labelling enzyme was omitted were negative.