Helicobacter pylori evolution during progression from chronic atrophic gastritis to gastric cancer and its impact on gastric stem cells

Giannakis et al. 10.1073/pnas.0800668105.

Supporting Information

Files in this Data Supplement:

SI Table 1
SI Table 2
SI Table 3
SI Text
SI Figure 4
SI Figure 5
SI Table 4
SI Table 5
SI Figure 6
SI Figure 7
SI Figure 8
SI Table 6
SI Table 7
SI Table 8
SI Table 9
SI Table 10
SI Table 11
SI Table 12
SI Table 13
SI Table 14
SI Table 15
SI Table 16

SI Figure 4

Fig. 4. Colony forming units in the stomachs, and IgM, IgG1, and IgG2b levels in the sera of gnotobiotic tox176 mice 4 weeks and 6 months after inoculation of the Kx1 or Kx2 strains. Each symbol at each time point represents an individual mouse (filled symbols, tox176 mice infected with Kx1; open symbols, tox176 mice infected with Kx2). *, P < 0.05.

SI Figure 5

Fig. 5. Characterization of the mGEP cell line. (A) Transmission EM of a cultured mGEP cell demonstrating features of preparietal progenitors: the arrow points to a primitive canaliculus. (B and C) mGEPs express biomarkers of gastric stem cells. (B) after fixation in paraformaldehyde, mGEPs were stained with antibodies specific for Notch3 (red). T-antigen (green), and phalloidin (blue; an actin marker) are shown. (C) mGEPs express cytoplasmic Dcamkl1 (red). T-antigen appears green, and phalloidin appears blue. (Scale bars: A, 1 mm; B and C, 20 mm.)

SI Figure 6

Fig. 6. Invasion of mGEP cells by H. pylori strain HPAG1. Arrows point to extracellular bacteria. Arrowheads indicate intracellular bacteria. See SI Text for details of the invasion assay. (Scale bars: 20 mm.)

SI Figure 7

Fig. 7. Assays of adherence of Kx1 and Kx2 to cultured mGEPs and IL-6 induction. (A) Mean values ± SD from two replicate experiments are plotted. (B) Measurements were made after infection for 24 h. Mean values ± SD from three replicate experiments are plotted.

SI Figure 8

Fig. 8. Immunohistochemistry-based invasion assay for Kx1 and Kx2 after a 24-h infection of mGEPs. Staining for Kx1 (A) and Kx2 (B) in the absence of saponin marks extracellular bacteria (purple). In the presence of saponin, a cell-permabilizing agent, no invasion is seen with the Kx1 isolate (C) but both extracellular bacteria (arrows; purple staining) and intracellular bacteria (arrowheads; red staining) are seen with the Kx2 isolate (D). See SI Methods for details of the assay. (Scale bars: 20 mm.)

SI Text

Methods

Kalixanda Study. Details concerning the design of the Kalixanda study, and the histologic criteria used to score the patient's gastric biopsies can be found in Aro et al. (1) and Storskrubb et al. (2). The Kalixanda study was approved by the Ethics Committee of Umeå University in May, 1998.

Culture of Strains and Initial Characterization of Genomic DNA. After initial isolation, all strains from the de-identified 75-year old (age at first EGD) male patient who progressed from ChAG to gastric adenocarcinoma were grown under microaerophilic conditions (5% O₂ ,10% CO₂ and 85% N₂) for 48-72h at 37°C on Brain Heart Infusion (BHI) agar, supplemented with 10% calf blood, vancomycin (6 mg/ml), trimethoprim (5 mg/ml), and amphotericin B (8 mg/ml). For liquid culture, bacteria were grown under microaerophilic conditions in Brucella Broth (BB) supplemented with 5% FCS (Sigma) and 1% IsoVitaleX (Becton-Dickinson; adjusted to pH 7.0).

DNA was isolated from bacterial strains using standard protocols (Qiagen). RAPD assays were performed with primers 1283 (5'-GCGATCCCCA-3') and 1290 (5'-GTGGATGCGA-3') as described in ref. 3.

Genome Sequencing and Assembly. Strains Kx1 and Kx2 were grown separately on BHI agar plates as described above. Cells were scraped off the plates and genomic DNA was prepared. DNA from each strain was used for a single run of the 454 Life Sciences GS20 pyrosequencer and contigs were assembled by using the Newbler Assembler (4).

Gene Identification by BLAST. All annotated protein-coding genes in the finished genomes of H. pylori strains HPAG1, J99, and 26695 were combined into a single protein BLAST database. The assembled nucleotide sequences from Kx1 and Kx2 were put into separate BLAST databases. Each protein sequence from the three previously sequenced genomes was used as a query for a TBLASTN search against the Kx1 and Kx2 nucleotide databases. A gene was considered present if it or an ortholog of it [as defined in Chen et al. (5)] had a TBLASTN hit with an E value of less than 1 ´ 10^-4. To verify these presence calls, we masked the Kx1 and Kx2 sequences with the TBLASTN hits. The remaining sequence was used in a BLASTX search against all of the annotated proteins in the 26695, J99 and HPAG1 genomes. After applying an E value cutoff of 1 ´ 10^-4, we found that no genes that we called absent had been hit by any of the remaining Kx1 or Kx2 sequences. We chose this E value cutoff (1 ´ 10^-4) because it should identify both orthologous and homologous genes in the Kx strains: it is expected to produce a high false positive rate: i.e., genes from sequenced H. pylori strains that may only have homologs and not orthologs in Kx1 and/or Kx2 would be called as present in Kx1 and Kx2. However, this approach allowed us to be very confident that our identification of genes found in previously sequenced H. pylori strains but not in Kx1 or Kx2 would be conservative, and would suffer a low false positive rate.

Gene Differences Between Kx1 and Kx2. All genes (or gene fragments) that were identified in Kx1 and Kx2 in the above TBLASTN and BLASTX searches against the previously sequenced H. pylori genomes were collected. Each gene (or gene fragment) identified in Kx1 was then used as a query in a BLASTN search against the Kx2 gene/gene fragment database. If any of the following criteria were met, the gene was considered to be present in Kx2: (i) at least 90% identity over at least 100 nt; (ii) at least 95% identity over at least 50 nt; (iii) 100% identity over at least 30 nt; (iv) 100% identity over at least 15 nt if the alignment was truncated by the end of a contig; and (v) 100% identity over at least 90% of the gene fragment length. Sequences that did not pass any of these criteria were collected. Similarly, genes from Kx2 were used in a search against the Kx1 database to find sequences in Kx2 but not in Kx1. These were regarded as candidate gene differences. We manually culled this list using the TBLASTN and BLASTX data above: if multiple gene fragments from the same genome (Kx1 or Kx2) had the same hit in the BLAST database consisting of proteins from all three sequenced genomes (26695, J99. and HPAG1), they were only counted once. Genes and gene fragments that passed these tests were designated "gene differences" between Kx1 and Kx2, and annotated based on their best BLAST hit to the sequenced H. pylori genomes.

Analysis of Positive Selection. A test for positive selection was done on individual genes as described in Chen et al. (5). Consensus gene sequences for Kx1 and Kx2 were inferred from raw 454 sequencing data. Briefly, raw reads were mapped to a reference finished ChAG-associated H. pylori genome (HPAG1), using BLASTN with a 1 ´ 10^-4 E value cutoff. For each position within the HPAG1 genome, all reads covering that position were collected, and the most frequent base at that position in the 454 dataset (as aligned by BLASTN) was called the consensus base for Kx1/Kx2. Inferred consensus sequences for Kx1/Kx2 were considered to be orthologous to the reference HPAG1 gene and its corresponding orthologs in the J99 and 26695 genomes. All subsequent recombination, phylogenetic tree inference, and maximum likelihood analyses were performed by using programs and control files detailed in Chen et al. (5). Genes found to be under positive selection were further verified manually using the same procedures, but with Newbler assembled 454 data instead of raw 454 data. In this verification step, consensus gene sequences for Kx1 and Kx2 were based on BLASTN of the appropriate HPAG1 gene to the assembled Kx1/Kx2 sequences.

Colonization of Germ-Free Atp4b-tox176 Mice. Germ-free FVB/N tox176 mice and their wild-type littermates were maintained in plastic gnotobiotic isolators (6) under a strict 12-h light cycle. Animals were given an autoclaved chow diet (B&K Universal, East Yorkshire, U.K.) ad libitum. All manipulations of mice were performed by using protocols approved by the Washington University Animal Studies Committee. The Kx1 and Kx2 isolates were grown under microaerophilic conditions in Brucella broth containing 5% FBS (HyClone) and 1% IsoVitaleX (Becton Dickinson) to log phase overnight, pelleted at 7,000 ´g for 5 min at 25°C, washed once with PBS, and resuspended in PBS. Each infected mouse was gavaged once with 2 ´10⁸ CFU of Kx1 or Kx2 in 100 ml of PBS.

ELISA of Sera. Bacterial antigens were extracted by growing Kx1 and Kx2 to log phase in BB supplemented with FBS and IsoVitaleX. Cells were harvested by centrifugation (1,000 ´ g for 5 min at 25°C), followed by a wash in PBS (4°C), resuspension in PBS, vortexing (1 min at 25°C), and incubation of the resulting suspension at room temperature for 30 min. This step was followed by centrifugation at 20,000 ´g at 4°C for 20 min, collection of the supernatant, and quantification of protein levels.

Each well of the 96-well plastic ELISA plates (Nunc, MaxiSorp) was coated with 10 mg/ml of Kx1 or Kx2 antigen in sodium bicarbonate coating buffer (15 mM sodium carbonate, 35 mM sodium bicarbonate, 3 mM sodium azide, pH 8.5). Serial dilutions of each serum sample were then incubated in blocking buffer (PBS plus 1% BSA and 0.05% Tween-20) in the ELISA plates for 2h at room temperature. After three washes with PBS/0.05% Tween-20, goat horseradish peroxidase (HRP)-conjugated anti-mouse IgG1, anti-mouse IgG2b or anti-mouse IgM (Southern Biotech) were added for 1 h at 25°C at 1:1,000 dilutions (in blocking buffer). Development was done with 2,2'-azino-bis (3-ethylbenzthiazoline-6-sulfonic acid) (ABTS, 1 mM; Roche) in citrate buffer (100 mM citric acid, 50 mM sodium phosphate, pH 4.2) containing 0.03% H₂O₂ for 30 min, and the plates then read in an ELISA plate reader (ThermoMax, Molecular Devices) at OD 405 nm. Readings that fell within the linear range of signal intensity produced by a reference control serum sample obtained from a mouse colonized with H. pylori strain Hp1 (from a Peruvian patient with gastritis; ref. 7) were expressed as a percentage of that control (n = 2 assays per serum sample).

Cell Culture. mGEPs were grown in 75 cm² cell culture flasks (Corning) at 37°C in RPMI medium 1640 (Sigma) supplemented with 10% FBS (HyClone) under an atmosphere of 5% CO₂/95% air to 70% confluency. Nonprogenitor Gastric Epithelial Cells (npGECs) were grown in RPMI medium 1640 (Sigma) supplemented with 5% FBS, 1 mg/ml insulin (Sigma), 5 units/ml mouse Inf-g (Peprotech) at 33°C to 70% confluency, and then shifted to 37°C in the absence of Inf-g and insulin for 3d before infection.

Immunohistochemical Studies of Uninfected mGEP Cells. mGEP cells were plated at a density of 10⁵cells per well in 24-well plates (Corning). Twenty-four hours later, medium was removed, and cells were fixed in 4% paraformaldehyde/PBS. After incubation in blocking buffer (1% BSA, 0.3% Triton-X, PBS), the following primary antibodies were added (final dilution of 1:200 in blocking buffer) for 1 h at 25°C: rabbit anti-mouse double cortin and calcium/calmodulin-dependent protein kinase-like-1 (Dcamkl1; a gift from Christopher Walsh, Harvard Medical School, Boston, MA), rabbit anti-mouse Notch3 (Santa Cruz), and goat anti-SV40 TAg (8). Bound antigen-antibody complexes were detected with Alexa-Fluor 488-labeled donkey anti-goat Ig (Molecular Probes) and Alexa-Fluor 594-labeled donkey anti-rabbit Ig (Molecular Probes). Alexa-Phalloidin-647 (Molecular Probes) was used to label actin filaments.

Bacterial Adherence Assay. mGEP cells were seeded in 24-well polystyrene plates (Corning, 1 ´ 10⁵ cells per well), grown for 24h to 70% confluence, and infected with Kx1 and Kx2 (1 ´10⁵ to 6 ´ 10⁶ bacteria per well) for 1h at 25°C, with continuous shaking under room atmosphere. Cells were subsequently washed with PBS to remove nonadherent bacteria (three times; 25°C). Cells and adherent bacteria were then fixed with 4% paraformaldehyde for 15 min, washed twice with PBS, and incubated for 1h at 25°C with blocking buffer (PBS containing 1% BSA and 0.3% Triton-X). Rabbit anti-H. pylori (DAKO, 1:1,000) was added overnight at 4°C. Cells were washed (three times in PBS, 5min/cycle), and Alexa-Fluor-labeled 594 donkey anti-rabbit Ig was added together with Alexa-Fluor 488-phalloidin (Molecular Probes; 1:500) for 1 h at 25°C. Nuclei were stained with 4',6'-diamidino-2-phenylindole (DAPI; 2 mg/ml). Attached bacteria were visualized by fluorescence microscopy (Zeiss Axiovert microscope). In two independent experiments, and for each isolate, we scored and averaged out the number of bacteria attached to >8,000 mGEP cells.

Bacterial Invasion Assay. mGEP cells were seeded in 24-well plates (Corning; 1 ´ 10⁵ cells per well), and cultured for 20h at 37°C. Approximately 4.5 ´ 10⁸ CFU of Kx1 and Kx2 were then added at 37°C. Three hours later, cells were washed with PBS (four times; 25°C) to remove nonattached bacteria, and then incubated with fresh medium for 20 h at 37°C. Cells were washed with PBS (three times; 25°C), fixed in 4% paraformaldehyde for 15 min, and attached extracellular bacteria marked with rabbit polyclonal antibodies to H. pylori surface proteins (DAKO, 1:1,000 dilution; overnight incubation at 4°C). Bound antibodies were visualized with biotin-tagged anti-rabbit Ig (1:500; 1h at 25°C) followed by streptavidin-Alexa-Fluor 350 (1:500; 1 h at 25°C). After washing three times in PBS (5 min per cycle at 25°C) to remove unbound antibodies, mGEP cells in some wells were permeabilized [1% saponin (Sigma), 3% BSA in PBS; 15 min at 37°C]. Antibodies to H. pylori were added (DAKO; 1h at 25°C, 1:1,000 dilution) to label both extra- and intracellular bacteria, and bacteria were visualized with Cy3-tagged donkey anti-rabbit Ig. Intracellular bacteria were defined as those that were only stained with Cy3 and not with Alexa-Fluor 350. Alexa-Fluor 488-phalloidin (Molecular Probes; 1:500) was used to label actin filaments.

Navigated Laser Capture Microdissection (n-LCM). Gnotobiotic mice (6-month time point) were treated with BrdU 90 min before killing to label cells in S-phase. Stomachs were excised, cut in half along their cephalocaudal axis, and the resulting half-stomachs either plated for CFU determination after homogenization in PBS or embedded in OCT compound (Sakura). Cryosections (7 mm-thick) were cut and numbered sequentially. For every group of three slides, the middle one was taken for fixation in methanol for 5 min at -20°C, rinsed with de-ionized water (5 min), washed with PBS (three cycles of 5 min each at 25°C), placed in blocking buffer (1% BSA/0.3% Triton X-100 in PBS for 30 min) and incubated with rabbit anti-H. pylori sera (final dilution in blocking buffer, 1:100; DAKO), rat anti-mouse E-cadherin sera (1:500, Zymed), and goat anti-BrdU (1:1,000) overnight at 4°C. Bound primary antibodies were detected by using Alexa-Fluor-488, -594 and cyanine (Cy)5-conjugated secondary Igs. Nuclei were stained with DAPI.

Images of the stained middle section were taken with a Zeiss Axiovert microscope and Zeiss AxioCam, and used to direct capture of Kx2-associated GEPs (and GEPs from an uninfected mouse) from adjacent cryosections that had been lightly stained with methyl green and eosin to outline cellular architecture while preserving RNA integrity. n-LCM was performed by using the PixCell II system (Arcturus; 7.5-mm-diameter laser spot) and CapSure HS LCM Caps (Arcturus). RNA was isolated from 30 capture areas (~30 Kx2-associated progenitor cells and ~ 30 noninfected progenitor cells) using the PicoPure RNA isolation Kit (Arcturus). Unamplified RNA (for Azin1) and RNA treated with RiboAmp HS RNA Amplification kit (Arcturus) (for Odc1), were used as templates for cDNA synthesis (random primers, Superscript II Reverse Transcriptase kit; Invitrogen) before performing qRT-PCR.

qRT-PCR. cDNA was added to a 25 ml of reaction mix containing 12.5 ml of 2X SYBR green master mix (Applied Biosystems), UDP-N-glycosidase (Invitrogen; 0.25units/reaction) and gene-specific primers (100 nM). Each assay was performed in triplicate and the data were normalized to L32 mRNA levels (DDC_T method). Primers used included: L32-F: 5'-CCTCTGGTGAAGCCCAAGATC, L32-R: 5'-TCTGGGTTTCCGCCAGTTT, Azin1-F: 5'-ATTGACGATGCGAACTACTCCG, Azin1-R: 5'-TTCCCAAGATCCCCCACAAAA, Odc1-F: 5'-GACGAGTTTGACTGCCACATC, Odc1-R: 5'-CGCAACATAGAACGCATCCTT, Cdkn1a-F: 5'-CCTGGTGATGTCCGACCTG, Cdkn1a-R: 5'- CCATGAGCGCATCGCAATC, Kangai1-F: 5'-TTCGGGGTGTGGATTCTTGC, Kangai1-R: 5'-AGGAAGCCCATCACTATGGTG. Incubations were performed by using the Stratagene Mx3000P QPCR system.

SI Results

Identifying Genes Under Positive Selection in the Kx1 and Kx2 Strains Compared to Three Previously Sequenced Strains. Using the three finished genome sequences, and methods described in SI Methods and ref. 5, we identified five genes that were under positive selection in the Kx1 and Kx2 strains: hofC, which encodes an outer membrane protein whose function remains unknown; ftsZ, a GTPase required for septum formation and cell division (9); pgm, a phosphoglycerate mutase; tetA(P) (see below), and a gene encoding a hypothetical protein (SI Table 3). Phosphoglycerate mutase (E.C. 5.4.2.1) not only catalyzes the inter-conversion of 2- and 3-phosphoglycerate in the glycolytic pathway but also physically interacts with a RNA polymerase sigma-80 factor (rpoD) (10), which has an increased rate of divergence in H. pylori compared to other bacteria (11). tetA(P) has homology to drug efflux proteins of the major facilitator superfamily (12), but its ligands are not known.

Establishment of Assays for H. pylori Attachment and Invasion of mGEPs Using HPAG1. HPAG1 is a type 1 H. pylori strain (13) that has been sequenced, annotated, and functionally characterized in our laboratory (14). It invades GEPs in tox176 mice (15). HPAG1 was inoculated into 24-well plates containing 70% confluent monolayers of mGEPs (10⁶ bacteria per well). mGEP infections were carried out for 30 min, 3h, and 24 h. Cells were washed (3X in PBS; 5 min/cycle at 25°C), fixed in 2% paraformaldehyde for 15 min, and extracellular bacteria were marked with rabbit polyclonal antibodies to H. pylori surface proteins (DAKO, 1:1,000 dilution; 1-h incubation at 25°C). Bound antibodies were visualized with biotin-tagged anti-rabbit Ig followed by streptavidin-Alexa-Fluor350. After washing (three times in PBS, 5 min per cycle) to remove unbound primary antibodies, an aliquot of mGEP cells was permeabilized (1% saponin, 3% BSA in PBS; 15 min), antibodies to H. pylori were added (1h, 1:1,000 dilution) to label both extra- and intracellular bacteria, and bacteria were visualized with Cy3-tagged antibodies to rabbit Ig. Intracellular bacteria were defined as those that only stained with Cy3 and not with AlexaFluor350.

Intracellular H. pylori were most prominent after a 24-h infection (SI Fig. 6); no intracellular HPAG1 were observed by immunohistochemistry at the 30-min time point. Intracellular H. pylori were viable: when mGEP cells were grown to 70% confluency (seeded at 5 ´ 10⁴ cells per well then incubated at 37°C for 2d) and subsequently infected for 24 h with 2 ´ 10⁸ CFU of strain HPAG1, we obtained a cell-associated population of 3 ´ 10⁶ bacteria with an internalized population of 2 ´ 10⁴ [defined by gentamycin protection assays performed according to Kwok et al. (16)]. Similarly, infection of npGECs with HPAG1 under the same conditions, yielded 9 ´10⁵ cell-associated and 1.7 ´ 10⁴ internalized bacteria.

1. Aro P, et al. (2004) Valid symptom reporting at upper endoscopy in a random sample of the Swedish adult general population: The Kalixanda study. Scand J Gastroenterol 39:1280-1288.

2. Storskrubb T, et al. (2005) A negative Helicobacter pylori serology test is more reliable for exclusion of premalignant gastric conditions than a negative test for current H. pylori infection: A report on histology and H. pylori detection in the general adult population. Scand J Gastroenterol 40:302-311.

3. Akopyanz N, Bukanov NO, Westblom TU, Kresovich S, Berg DE (1992) PCR-based RFLP analysis of DNA sequence diversity in the gastric pathogen Helicobacter pylori. Nucleic Acids Res 20:5137-5142.

4. Margulies M, et al. (2005) Genome sequencing in microfabricated high-density picolitre reactors. Nature 437:376-380.

5. Chen SL, et al. (2006) Identification of genes subject to positive selection in uropathogenic strains of Escherichia coli: A comparative genomics approach. Proc Natl Acad Sci USA 103:5977-5982.

6. Hooper LV, et al. (2002) in Methods in Microbiology: Molecular Cellular Microbiology eds Sansonetti P, Zychlinksy A. (Academic, San Diego), Vol. 31, pp 559-589.

7. Guruge JL, et al. (1998) Epithelial attachment alters the outcome of Helicobacter pylori infection. Proc Natl Acad Sci USA 95:3925-3930.

8. Hu Y, Ippolito JE, Garabedian EM, Humphrey PA, Gordon JI (2002) Molecular characterization of a metastatic neuroendocrine cell cancer arising in the prostates of transgenic mice. J Biol Chem 277:44462-44474.

9. Rothfield L, Taghbalout A, Shih YL (2005) Spatial control of bacterial division-site placement. Nat Rev Microbiol 3; 959-968.

10. Rain JC, et al. (2001) The protein-protein interaction map of Helicobacter pylori. Nature 409:211-215.

11. Solnick JV, Hansen LM, Syvanen M (1997) The major sigma factor (RpoD) from Helicobacter pylori and other gram-negative bacteria shows an enhanced rate of divergence. J Bacteriol 179:6196-6200.

12. Lewinson O, Adler J, Sigal N, Bibi E (2006) Promiscuity in multidrug recognition and transport: The bacterial MFS Mdr transporters. Mol Microbiol 61:277-284.

13. Enroth H, Kraaz W, Engstrand L, Nyren O, Rohan T (2000) Helicobacter pylori strain types and risk of gastric cancer: a case-control study. Cancer Epidemiol Biomarkers Prev 9:981-985.

14. Oh JD, et al. (2006) The complete genome sequence of a chronic atrophic gastritis Helicobacter pylori strain: evolution during disease progression. Proc Natl Acad Sci USA 103:9999-10004.

15. Oh JD, Karam S, Gordon JI (2005) Intracellular Helicobacter pylori in gastric epithelial progenitors. Proc Natl Acad Sci USA 102:5186-5191.

16. Kwok T, Backert S, Schwarz H, Berger J, Meyer TF (2002) Specific entry of Helicobacter pylori into cultured gastric epithelial cells via a zipper-like mechanism. Infect Immun 70:2108-2120.