All patients were operated on at the Department of Surgery, Lund University Hospital. Formalin-fixed paraffin-embedded pancreatic tissue samples from 84 patients were obtained from the Department of Pathology at the same hospital. Prior to deembedding, one pathologist (U.S) reviewed all samples and approximately 100 mg of each tissue type was taken from the paraffin blocks with the tip of a scalpel. By carefully comparing the blocks with the slides, it was ascertained that a pure tissue type, i.e. tumor or normal etc, was obtained.

The tissue samples were from consecutive patients with primary exocrine cancer, predominantly of ductal type (PC) (n = 40, 20 females, mean age 59 years, range 44-77 years), neuroendocrine cancer (NE) (n = 14, 6 females, mean age 58 years, range 15-84 years), and multiple endocrine neoplasia type 1 (MEN) (n = 8, 1 female, mean age 52 years, range 42-69 years). In addition to a tumor specimen available from all 62 PC-, NE- and MEN patients, a sample of adjacent normal tissue was obtained from 41 patients (66%). Atrophic pancreatic tissue was available from 13 of the PC patients. Thus, 116 pancreatic tissue samples (1.9 samples per patient in average) were analyzed. Samples from patients with chronic pancreatitis (CP) of alcoholic, idiopathic, and epithelioid cell granulomatosis etiology (n = 5, 1 female, mean age 52 years, range 42-79 years), were also examined.

Pancreatic tissue specimens from patients with benign (other than pancreatitis) pancreatic diseases (n = 10, 8 females, mean age 55 years, range 24-71 years) such as mucinous cystadenoma (n = 5), serous cystadenoma (n = 3), pancreatic cysts (n = 2), as well as histologically normal pancreatic tissue samples from patients with cancer of ductus choledochus (n = 4, 1 female, mean age 65 years, range 55-75 years), colon (male, mean age 60 years), duodenum (male, mean age 64 years), and retroperitoneal fibrosis (male, mean age 57 years), were included as controls (collectively denoted C). Gastric tissue samples of the antrum and/or fundus, as well as duodenal samples, were obtained from 23 PC patients and four NE- and MEN patients. An average of 2.5 stomach and/or duodenum specimens was tested per patient. Specimens of the gallbladder (n = 18) and ductus choledochus (n = 8) were also obtained from PC-, NE- and MEN patients. This study was approved by the Research Ethics Committee at Lund University (LU 726-02).

Bacterial strains were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and the Culture Collection of the University of Gothenburg (CCUG, Gothenburg, Sweden). Genomic DNA of Bacteroides fragilis (ATCC 25285), H pylori (CCUG 17874), H. bilis (CCUG 38995), H. pullorum (CCUG 33840), H. sp. flexispira (CCUG 23435), H. hepaticus (CCUG 33637), Escherichia coli (ATCC 25922), Campylobacter jejuni (ATCC 33560), Clostridium difficile (ATCC 9689), Proteus mirabilis (ATCC 43071), Klebsiella pneumoniae (ATCC 10031), Enterococcus faecalis (ATCC 29212) and Enterobacter cloacae (CCUG 6323) was extracted from 10⁷ bacterial cells by the QIAamp^® DNA Mini Kit (Qiagen) as described below.

Paraffin-embedded tissue samples were heated at 60 °C for 10 min to melt excess paraffin, aseptically transferred to new micro-centrifuge tubes, washed in xylene for 2 × 5 min, rehydrated through graded ethanol (990 mL/L and 950 mL/L for 2 × 5 min and 700 mL/L for 5 min), and finally washed for 5 min in double-distilled water. Subsequently, the samples were homogenized in 170 mmol/L phosphate-buffered saline pH 7.2 by using a plastic microcentrifuge tube-adapted pestle. Pancreatic tissue samples were homogenized at 10 g/L and other gastrointestinal specimens at 40-50 g/L. DNA was extracted from 100 μL of each homogenate by the QIAamp^® DNA Mini Kit Tissue protocol (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. All DNA samples were stored at -20 °C.

DNA-extracts were amplified in a GeneAmp 2700 Thermocycler (Applied Biosystems, Foster City, CA, US) by semi-nested PCR-assay for Helicobacter species as previously described[18] with primers constructed by Goto et al[19]. The forward primer 1F (5´-CTATGACGGGTATCCGGC-3´) and reverse primer 1R (5´-CTCACGACACGAGCTGAC-3´) were used in the first step. In the second step, primer 1F and reverse primer 2R (5´-TCGCCTTCGCAATGAGTATT-3´) were used. Precautions were taken to minimize the risk of PCR cross contamination as described recently[20]. Detection of PCR-products was done in agarose gels as described previously[18].

To rule out the non-specific translocation of a common gut microbe to the pancreas among the studied patients, a pancreatic tissue sample (n = 84) from each patient was amplified by nested PCR with primers for the glutamine synthase gene of Bacteroides species. Primers BFR-1 (5´-ACTCTTTGTATC CCGACGATT-3´) and BFR-2 (5´-GAGGTTGATGCCTGTATCGGT-3´), described by Kane et al[21], were used. For the second step, internal primers (forward primer BFR-3: 5´-GACAAAAACATCACCCGGGT-3´ and reverse primer BFR-4: 5´-GCCCAGCTTGTGACACTCTATT-3´), based on the sequence of the BFR1-BFR2 PCR-product, were constructed using the Vector NTI^® Suite version 8.0 (Informax, Frederick, MD, US). The specificity of the Bacteroides species PCR-assay was evaluated using strains from the ATCC and CCUG. PCR-mixtures were prepared as described previously[18]. Amplification conditions for the first step were at 94 °C for 4 min; 30 cycles at 94 °C for 30 s, at 60 °C for 30 s, at 72 °C for 45 s and finally at 72 °C for 5 min. Conditions for the second step were at 94 °C for 10 min; 35 cycles at 94 °C for 30 s, at 60 °C for 30 s, at 72 °C for 30 s and finally at 72 °C for 5 min. Genomic DNA (0.1 ng) of B fragilis was used as a positive control.

Denaturing gradient gel electrophoresis (DGGE) analysis of the V6-7 region of Helicobacter species 16S rDNA was performed as described previously[18]. Diluted (10×) first step PCR-products were amplified in the second step using forward primer GC-1F (5´-GCGGCCGCCCGTCCCGCCGCCCCCGCCCCGCCGCGGCCGCCTATGACGGGTATCCGGC-3´) and 2R as described above. Electrophoresis was carried out using a DCode™ electrophoresis unit (BioRad, Hercules, CA, US). A mixture of amplicons of type strain DNA of H pylori (CCUG 17874), H. bilis (CCUG 38995), H. pullorum (CCUG 33840), H. sp. flexispira (CCUG 23435), and H. hepaticus (CCUG 33637) was used as gel migration markers.

Nucleic acid products of the Helicobacter genus-specific PCR-assay were purified from agarose gels using the Montage DNA Gel Extraction Kit (Millipore, Bedford, MA, US), or from DGGE-gels as described previously[18]. DNA-sequence reactions were performed using the ABI PRISM™ dRhodamine Terminator Cycle Sequencing Ready Reaction Kit version 3.0 (Applied Biosystems) with modifications. One microliter of a BigDye™ mix and 1.5 μL of sequencing buffer (10 μL of 10× PCR-buffer II, 6 μL 25 mmol/L MgCl₂, 4 μL double-distilled water) were prepared in a total volume of 10 μL with primers (1F or 2R) and template according to the manufacturer’s instructions. Products of the sequence reaction were aligned and the closest homologous DNA was identified by BLASTn-analysis as described elsewhere[17].

As a rule, two pancreatic specimens were obtained from each patient in the PC-, NE- and MEN-groups. If at least one specimen was positive, the patient was considered Helicobacter-positive. Hence, 75% (tumor and/or surrounding tissue) of the PC-, 57% of the NE-, and 38% of the MEN patients were positive for the genus Helicobacter (Table 1). Three (one alcoholic, one idiopathic and one with epithelioid cell granulomatosis) of five patients with chronic pancreatitis were Helicobacter-positive. All benign tissue samples from cystadenoma- and pancreas cyst patients, as well as pancreatic tissue from the remaining C-patients, were negative for the genus Helicobacter (Table 1).

Helicobacter DNA was detected in 48% of tumors of PC patients (11 PC patients were PCR-negative in the tumor but positive in surrounding normal or atrophic tissue, hence, a higher number of PC patients [75%] compared with that of PC tumor patients [48%] were Helicobacter-positive), 57% of neuroendocrine tumors, and 38% of MEN tumors (Table 2). Atrophic and normal pancreatic tissues of the PC patients were positive in 69% and 36% of the samples, respectively, whereas only 5% of normal pancreatic tissues surrounding NE and MEN tumors were helicobacter-positive (Table 2). Commonly, if a PC patient was positive for Helicobacter species in a tumor sample, normal tissue was often negative and vice versa. Positive atrophic tissue was common in both tumor-negative and positive samples (Table 2).

Stomach and/or duodenum samples demonstrated a Helicobacter-positive result in 33% of the patients (Table 1). Of the 40 PC patients, stomach and/or duodenum was obtained from 23. Among these 23 patients, 30% were positive in the stomach and/or duodenum whereas 74% were positive in the pancreas. Four PC patients and one NE patient were simultaneously positive both in the pancreas and in a gastric and/or duodenal specimen (Figure 1). All gallbladder as well as ductus choledochus tissue samples were negative in the Helicobacter genus-specific PCR (Table 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Agarose gel electrophoresis (A), DGGE (B), and the outcome of a public database comparison of Helicobacter 16S ribosomal DNA PCR-product sequences (C). From 7 pancreatic cancer patients (P), amplicons from the duodenum (D), fundus ventriculi (Fv), and antrum ventriculi (Av) were compared to amplicons from normal (Pn), atrophic (Pa) and pancreatic tumor (Pt) tissues. Migration markers (M) were (from top to bottom) amplified from DNA of H. bilis (1), H. pullorum (2), H pylori (3), H. sp. flexispira (4) and H. hepaticus (5).

The expected 600-bp product was amplified with genomic DNA of B. fragilis as a template in the first step of the assay. However, after nested analysis, the expected fragment of 228-bp was also produced with E. cloacae DNA. Extracted genomic DNA of the other tested reference strains was negative. A pancreatic sample from each patient (n = 84), including tumor-, benign-, and normal tissue, was analyzed. None of the samples showed positive amplification for the genus Bacteroides in the nested PCR-assay.

Forty-six Helicobacter 16S ribosomal DNA PCR-products were identified by DGGE, and 20 of those were subjected to DNA-sequence analysis. DGGE revealed four migration profiles, similar to the reference strains of H pylori, H. sp. flexispira, H. bilis and H. hepaticus (Figure 1). The two methods were in concordance except for two PCR-products migrating with H. sp. flexispira in DGGE analysis but with the highest similarity to H. cinaedi after sequence- and BLASTn analysis. Twenty-seven (of 29) identified H pylori-sequences were amplified from pancreatic tissue samples. Sequences of H. sp. flexispira (n = 7) and H. cinaedi (n = 2) were only detectable in pancreatic tumor specimens (Table 3). Seven (of 23) PC patients were helicobacter-positive in gastric and/or duodenal samples, but only one of them was identified as H pylori. Eight PCR-products related to enteric Helicobacter spp. (H. bilis, n = 6; H. hepaticus, n = 2) were identified in gastric and duodenal specimens from 7 pancreatic cancer patients (H. bilis was detected in the stomach as well as duodenum in one PC patient) (Table 3). Five patients (four with PC and one with NE) were simultaneously Helicobacter DNA-positive in the pancreas and the stomach and/or duodenum. Four patients had different species in the different tissues, whereas one duodenal sample of a MEN-patient with H pylori DNA in the pancreas demonstrated both H. pylori and H. hepaticus in the DGGE analysis (Figure 1, patient no. 7). Agarose- and DGGE gel images as well as the result of DNA-sequencing of pancreatic as well as gastric and/or duodenal samples of 7 patients are shown in Figure 1.