Digestive Diseases and Sciences

, Volume 58, Issue 1, pp 140–149

H. pylori Infection Is Associated with DNA Damage of Lgr5-Positive Epithelial Stem Cells in the Stomach of Patients with Gastric Cancer

Authors

  • Takeshi Uehara
    • Department of Pathology and Laboratory MedicineUniversity of Pennsylvania School of Medicine
    • Department of Laboratory MedicineShinshu University School of Medicine
  • Deqin Ma
    • Department of Pathology and Laboratory MedicineUniversity of Pennsylvania School of Medicine
  • Yuan Yao
    • Department of Pathology and Laboratory MedicineUniversity of Pennsylvania School of Medicine
  • John P. Lynch
    • Division of Gastroenterology, Department of MedicineUniversity of Pennsylvania School of Medicine
  • Knashawn Morales
    • Department of BiostatisticsUniversity of Pennsylvania School of Medicine
  • Amy Ziober
    • Department of Pathology and Laboratory MedicineUniversity of Pennsylvania School of Medicine
  • Michael Feldman
    • Department of Pathology and Laboratory MedicineUniversity of Pennsylvania School of Medicine
  • Hiroyoshi Ota
    • Department of Laboratory MedicineShinshu University School of Medicine
    • Department of Pathology and Laboratory MedicineUniversity of Pennsylvania School of Medicine
Original Article

DOI: 10.1007/s10620-012-2360-8

Cite this article as:
Uehara, T., Ma, D., Yao, Y. et al. Dig Dis Sci (2013) 58: 140. doi:10.1007/s10620-012-2360-8

Abstract

Background

H. pylori (Hp) infection is a major risk factor in gastric carcinogenesis leading to epithelial mutagenesis, and may affect gastric epithelial stem cells.

Aims

To characterize the expression of Lgr5, a marker of epithelial stem cells in human gastric mucosa, to determine whether Hp infection affects Lgr5-positive epithelial cells (LPECs) and whether LPECs are susceptible to DNA damage associated with Hp infection.

Methods

Lgr5 expression was characterized in non-neoplastic gastric mucosa from 52 patients (34 with and 18 without gastric cancer (GC); 21 Hp-positive (Hp+) and 31 Hp-negative (Hp−)) by immunohistochemical and immunofluorescence staining. To determine the extent of DNA damage in LPECs, nuclear 8-hydroxydeoxyguanosine (8OHdG), a marker of DNA damage associated with oxidative stress, was measured by quantitative spectral image analysis.

Results

LPECs were primarily present in gastric antrum. Higher numbers of LPECs were seen in Hp+ than in Hp− non-neoplastic mucosa of GC patients, P = .006, but not in patients without GC. 8OHdG levels in LPECs were significantly higher than in Lgr5-negative epithelial cells in Hp+ GC patients (P = .012) but not in Hp− cases (P = .414), whereas no difference was seen between Hp+ and Hp− mucosa of patients without GC.

Conclusions

The Lgr5-positive epithelial stem cell pool is expanded in Hp-associated gastritis in the antrum of patients with GC. In GC patients with active Hp infection, LPECs may be more susceptible to DNA damage than Lgr5-negative epithelial cells, suggesting that Hp infection may contribute to GC risk by affecting epithelial stem cells in the human stomach.

Keywords

Helicobacter pyloriGastricDNA damage8OHdGLgr5Stem cells

Introduction

Helicobacter pylori infection of the stomach leads to chronic gastritis, intestinal metaplasia, and increased risk of gastric cancer (GC) [15]. Gastric cancer is the fourth most common cancer in the world [6], and 75 % of GC cases are attributable to H. pylori infection [5]. We have previously demonstrated that H. pylori infection leads to reduced DNA mismatch repair and enhanced mutagenesis in gastric epithelial cells [79]. However, how these induced mutagenic and regulatory alterations are transmitted to cell progeny and potentially result in cancer development remains poorly understood. It is likely that critical genomic and epigenetic changes caused by H. pylori infection affect gastric epithelial stem cells and epithelial progenitor cells, affecting their primary role of re-populating the gastric glandular units and programming them to a neoplastic pathway, consistent with the cancer stem cell hypothesis [10].

Few specific markers of gastric stem cells are available, limiting our ability to address their role in H. pylori-associated carcinogenesis. The leucine-rich repeat containing G protein coupled receptor Lgr5, also known as GPR49, is an orphan G protein-coupled receptor and a Wnt target gene that has been shown to be expressed in adult stem cells in the stomach and intestinal epithelium in a mouse model and, recently, in the human stomach [11, 12]. Using lineage tracing in mice, Barker et al. [11] showed that Lgr5 was expressed at the base of the gastric glands and that Lgr5-positive cells had the capacity of self-renewal and could build long-lived gastric antral units in vivo, thus meeting the criteria for stem cells (self-renewal and multipotency defined as the ability to give rise to daughter cells that differentiate into multiple lineages) [10]. In mice, Lgr5 expression was observed in a limited set of adult tissues including the intestine and antral epithelium of the adult stomach [13, 14]. Therefore, Lgr5 may be used as a marker of gastric epithelial stem cells. Our hypothesis is that Lgr5-positive gastric epithelial stem cells may be important in the response to H. pylori infection and associated inflammation.

To date there are limited data regarding the characterization of Lgr5-positive cells in gastric disease, namely the features of Lgr5-positive cells in human gastric mucosa and the genomic changes of this stem cell population in response to enhanced oxidative stress in the milieu of H. pylori gastritis.

8-Hydroxydeoxyguanosine (8OHdG) is believed to be one of the main DNA modifications induced by reactive oxygen species (ROS) in inflammatory conditions such as H. pylori gastritis [15, 16]. Interestingly, the levels of 8OHdG in gastric mucosa have been shown to decrease substantially after eradication of H. pylori infection [15, 16]. 8OHdG may also be responsible for DNA base mutations, because the adducts can accumulate and may be only partially repaired through enzyme pathways that may, in turn, cause further DNA damage [17].

In this study, we characterized the expression of Lgr5 in human gastric mucosa and determined 8OHdG levels in LPECs to determine:
  1. 1.

    whether the Lgr5-positive stem cell population is susceptible to oxidative damage in the setting of chronic H. pylori infection; and

     
  2. 2.

    whether responses are different for patients with and without gastric cancer.

     

Materials and Methods

Cases

Gastric tissue samples from fifty-two patients were included in the study. Thirty-four patients had gastric cancer and 18 did not (Supplement Table 1). Cases were selected consecutively from archived formalin fixed-paraffin embedded (FFPE) gastrectomy specimens from 2003 to 2008, at the Hospital of the University of Pennsylvania, Philadelphia, PA, USA. Gastric tissue samples from nine patients with H. pylori gastritis without gastric cancer were from the Department of Laboratory Medicine, Shinshu University School of Medicine, Matsumoto, Japan. Representative tissue blocks containing gastric antral mucosa and/or oxyntic mucosa were selected when available from each individual gastrectomy. In each case, gastric mucosa with none or small areas of intestinal metaplasia were selected for the study (Supplement Table 2).

Immunohistochemistry and Immunofluorescence of Stomach Tissues

FFPE tissue sections cut at 5 μm thickness were used. Antigen retrieval was performed with Target Retrieval Solution (Dako North America, Carpinteria, CA, USA) by boiling the slides for 20 min followed by cooling at room temperature for 20 min. The sections were then incubated with Protein Block Serum Free solution (Dako North America) at room temperature for 1 h. The individual primary antibodies for detection of Lgr5, H. pylori, 8OHdG, CD34, CD45, cytokeratin AE1/AE3, CD44, vimentin, and GFP (Supplement Table 3) were applied and incubated with the tissue sections for 1 h at room temperature in a humidified chamber. The slides were then washed in phosphate-buffered saline and incubated with secondary antibody using the Dako LSAB2 System-HRP Kit according to the manufacturer’s instructions (Dako North America).

Immunohistochemistry for H. pylori was performed with a rabbit polyclonal antibody (Supplement Table 3) using Bond Polymer Refine Detection reagents as recommended by the manufacturer (Vision Biosystems, Buffalo Grove, IL, USA).

For immunofluorescence, after the sections were incubated with the primary antibody, the slides were washed in PBS and incubated with anti-rabbit antibody labeled with Alexa Fluor 488 or anti-mouse antibody labeled with Alexa Fluor 568 (Invitrogen, Carlsbad, California, USA) for 1 h at room temperature in the dark. For negative controls, the primary antibodies were omitted. The slides were then washed in PBS and mounted using Vectashield (Vector Laboratories, Burlingame, USA). The immunofluorescence images and quantitative multispectral data were obtained using an argon laser (420–720 nm) connected to an inverted fluorescence microscope (Nuance Trio; CRI, San Francisco, CA, USA). Images were captured using a 40× objective lens. Digital image overlays were processed using CRI Maestro systems (CRI).

Histology and Scoring of Immunohistochemical and Immunofluorescence Stains

The number of Lgr5-positive cells in each case was scored on the basis of the highest number of Lgr5-positive cells per high-power field (400×) in gastric epithelium, examined over the entire tissue section. The intensity of Lgr5 stain was reported as three grades as follows: no stain (0); intermediate stain (1); and strong stain (2). Grades 1 and 2 were regarded as positive Lgr5 expression. H. pylori immunohistochemistry was scored as positive or negative. The inflammatory grade, including neutrophils and mononuclear inflammatory cells in the gastric mucosa, and the extent of intestinal metaplasia were scored using the updated Sydney system (Houston classification) of gastritis, with scores ranging from 0 to 3 [4].

To determine the 8OHdG maximum fluorescence intensity in Lgr5-positive versus negative epithelial cells, at least five cells including both Lgr5-positive and negative epithelial cell were counted in each image. Similarly, in the lamina propria, at least five non-epithelial cells including both Lgr5-positive and negative cells were counted in each image. Therefore, data from at least 10 Lgr5-positive and 10 Lgr5-negative epithelial and lamina propria cells were obtained from each case. The 8OHdG maximum fluorescence intensities of Lgr5-positive cells were compared with those of Lgr5-negative cells by use of image-analysis software.

To determine the co-localization of CD45, CD34, and cytokeratin in the Lgr5-positive cells in the gastric antral epithelium and lamina propria, at least 10 Lgr5-positive cells from each of the glandular epithelium and lamina propria were counted. To determine whether there is co-localization of vimentin (a mesenchymal cell marker) and CD44 (one of a number of gastric stem cell markers [12]) in Lgr5-positive cells, immunofluorescence was performed as described above.

Immunofluorescence Stains for Lgr5 in Mouse Stomach and Gastric Cell Lines

Heterozygous Lgr5-EGFP-IRES-CreERT2 mice in a C57BL/6 genetic background (Lgr5 mice) harbor an Lgr5-EGFP-IRES-CreERT2 “knock-in” allele that abolishes Lgr5 gene function and expresses green fluorescent protein (EGFP) and CreERT2 fusion protein from the Lgr5 promoter/enhancer elements. In Lgr5 mice EGFP fluorescence is observed in crypt base cells in mouse small intestine and colon and in epithelial stem cells of the stomach [11, 13, 18].

Heterozygous Lgr5 mice were obtained from Jackson Laboratories (Bar Harbor, ME, USA). For immunofluorescence studies, 5 µm thick sections from FFPE mouse stomach were used. Sections were treated with Protein Block Serum Free solution (Dako North America) at room temperature for 1 h. The slides were then incubated with a goat anti-GFP antibody for 1 h (Supplement Table 3). The slides were covered with Vectashield HardSet mounting medium with DAPI (Vector Laboratories). Dako NP Universal Negative Control Mouse IgG (NP015) was used as negative control.

The gastric cell lines AGS (CRL-1739; ATCC, Manassas, VA, USA), and SNU638 were grown in RPMI-1640 medium containing 10 % fetal bovine serum, penicillin, and streptomycin (Gibco, Gaithersburg, MD, USA) [8, 19]. The cell lines were plated in Lab-Tek II slide chambers (Thermo Fisher Scientific, Rochester, NY, USA), with 1 × 105 cells per well and grown for 24 h to reach 90 % confluence. The culture medium was removed, and the cells were washed with PBS and then fixed with 2 % paraformaldehyde in PBS. The slides were then stained with anti-Lgr5 antibody, followed by incubation with Alexa Fluor 488 conjugated goat anti-rabbit secondary antibody (Invitrogen) and covered with Vectashield HardSet mounting medium with DAPI.

Statistical Analysis

Statistical analysis used the two-tailed T test or Mann–Whitney rank sum test, multiple group comparison analysis with ANOVA, and pairwise correlations by Pearson product moment correlation. Analysis was performed with Sigmaplot 11.2 software (Systat Software, Point Richmond, CA, USA). P values <.05 were considered significant.

Results

Expression of Lgr5 in Gastric Mucosa

Lgr5 expression was examined in non-neoplastic gastric mucosa from 52 patients. Thirty-four patients had gastric adenocarcinoma (mean age 68.6 years) and eighteen did not (mean age 49.6 years) (Supplement Table 1). Overall Lgr5-positive cells were predominantly identified as rare epithelial cells in gastric mucous glands, with fewer positive cells present in the mucous necks of the glands; no Lgr5-positive cells were found on the surface or gastric pit epithelium (Fig. 1). Lgr5 appeared as a diffuse cytoplasmic stain and Lgr5-positive epithelial cells often had somewhat triangular shapes, with, occasionally, laterally elongated ends, and were located in the outer aspect of the glands, closer to the lamina propria rather than toward the lumen of the glands (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2360-8/MediaObjects/10620_2012_2360_Fig1_HTML.jpg
Fig. 1

Lgr5 expression in gastric mucosa detected by immunohistochemistry. aH. pylori-positive gastric antral mucosa with multiple Lgr5-positive cells in the gastric glands (arrow). The gastric pit and surface epithelium (SP), gland mucous neck region (MNR), and deep mucous gland region (DMG) are indicated. b Lgr5-positive epithelial cells in gastric antrum from another case (arrow). c Lgr5-positive cells showing triangular shape and basal orientation (arrow) in the glandular profiles. d Lgr5-positive cells (arrows) in a focus of intestinal metaplasia. e Rare Lgr-5 positive cells in the lamina propria (arrows). a and b magnification ×200; ce magnification ×400

In addition to Lgr5-positive epithelial cells we also found rare Lgr5-positive cells in the lamina propria, characterized by plasmacytoid morphology (Fig. 1).

Overall, 33 antral and 23 oxyntic mucosa samples were examined (Table 1). The mean number of Lgr5-positive epithelial cells was 11.4 and 3.7 in antral and oxyntic mucosa, respectively (P = .001), Table 1.
Table 1

Lgr5-positive epithelial cell numbers in the different groups

Groups

Agea

Mean (Std Dev)

Number of samplesb

Lgr5+ epithelial cells Mean (SD)

Test

P value

Mucosa type

All Antrumc

58.6 (17.9)

33

11.4 (7.0)

.001

All Oxynticc

65.2 (13.8)

23

3.7 (6.4)

H. pylori status

All positive

57.4 (19.6)

21

12.0 (6.4)

.001

All negative

65.1 (14.1)

35

6.0 (7.7)

GC+

H. pylori positive

66.4 (11.1)

12

11.7 (8.1)

.006

H. pylori negative

69.7 (12.9)

22

4.4 (4.9)

GC−

H. pylori positive

45.3 (22.5)

9

12.3 (3.7)

.314

H. pylori negative

53.8 (10.3)

13

8.7 (10.6)

GC+ (antrum mucosa)

H. pylori positive

69.0 (10.2)

10

13.5 (7.5)

.036

H. pylori negative

64.4 (14.6)

8

6.1 (5.7)

GC+ (oxyntic mucosa)

H. pylori positive

53.5 (3.5)a

2

2.5 (0.7)

.87

H. pylori negative

72.8 (11.3)

14

3.4 (4.3)

Patients with GC (GC+); patients without GC (GC−)

aThere were no statistically significant differences between the age groups (P > .05), except for the mean ages of H. pylori-positive versus H. pylori-negative patients in the sub-group GC+ (Oxyntic Mucosa), P < .05

bWhen available, separate samples of antral and oxyntic mucosa from each patient were studied

cInclude H. pylori-positive and negative mucosa from patients with and without gastric cancer

Focal intestinal metaplasia was present in 17 gastric antrum samples with a mean score of 0.85, out of possible scores ranging from 0 to 3 (Supplement Table 2). In 11 cases, intestinal metaplasia was present in the tissue sections used for Lgr5 immunohistochemistry. In the foci of intestinal metaplasia, few positive Lgr5-positive cells were identified in the base of intestinalized glands (Fig. 1), with a mean number of Lgr5-positive cells of 1.5, ranging from 1 to 4 Lgr5-positive cells per 400× field.

Expression of Lgr5 in Gastric Cancer Cell Lines and in the Mouse Stomach

Immunofluorescence stains were used to determine the expression of Lgr5 in cultured gastric cancer cell lines. We identified a population of Lgr5-positive cells (5.0 %) in the SNU638 cell line, Supplement Fig. 1. There were only rare Lgr5-positive cells in the AGS cell line (0.42 %). Therefore, the SNU638 cell line may be useful as a model to study the response of Lgr5-positive cells to H. pylori and other agents.

Previous studies in mice revealed expression of Lgr5 in a limited set of adult tissues including the antrum of the adult stomach [20]. Because anti-Lgr5 antibodies that recognize mouse Lgr5 are not currently available, we used an anti-EGFP antibody to identify putative Lgr5-positive cells in the stomach of Lgr5-EGFP-IRES-creERT2 mice. In the mouse stomach, we identified EGFP-positive cells in the base of glands in the gastric antrum, consistent with reported studies and with our findings in the human stomach described above [13] (Supplement Fig. 2).

Effect of H. pylori Infection and Inflammation on Lgr5 Expression in Gastric Mucosa

The H. pylori status of gastric mucosa samples was examined by detection of H. pylori on H&E stained sections and by immunohistochemistry. The grades of inflammation in the mucosa were scored by evaluating the grade of chronic inflammation (lymphocytes and plasma cells or other mononuclear inflammatory cells) and activity grade (neutrophils). Overall, when H. pylori-positive (n = 21) were compared with H. pylori-negative mucosa samples (n = 35), the number of Lgr5-positive epithelial cells was higher in H. pylori-positive cases (P = .001), Table 1. Further analyses showed that gastric cancer patients with H. pylori-positive gastritis in the background mucosa had a higher number of Lgr5-positive epithelial cells than H. pylori-negative patients (P = .006), whereas in patients without gastric cancer the number of Lgr5-positive epithelial cells was higher in H. pylori-positive than H. pylori-negative mucosa but did not reach statistical significance (P = .314), Table 1. Additionally, the number of Lgr5-positive epithelial cells in the mucosa of H. pylori-positive gastric cancer patients as compared with H. pylori-negative gastric cancer patients was significantly higher in the gastric antrum (P = .036) but not in the oxyntic mucosa (P = .87), Table 1.

Interestingly, no significant difference between the numbers of Lgr5-positive epithelial cells in the mucosa of gastric cancer-positive versus gastric cancer-negative patients was found when H. pylori status was not factored in (P > .05). These data suggest interaction of H. pylori with host susceptibility factors that affect the response of Lgr5-positive epithelial cells in the group of gastric cancer patients.

H. pylori infection is always associated with variable degrees of chronic and active inflammation in the gastric mucosa, the latter consisting of neutrophils. The mean acute and chronic inflammation scores were significantly higher in H. pylori-positive than in H. pylori-negative mucosa samples (P = .017) (Supplement Table 4). When the inflammation scores of antrum and oxyntic mucosa were scored separately, significantly higher scores for both neutrophils and mononuclear inflammatory cells were found in H. pylori-positive antrum (P = .001 and P = .017, respectively), whereas in oxyntic mucosa only the neutrophil scores were significantly higher in H. pylori-positive mucosa samples (P = .002) (Supplement Table 4). Interestingly, the numbers of Lgr5-positive epithelial cells were positively correlated with the acute inflammation scores in gastric antral mucosa (R = 0.407, P = .018; Pearson product moment correlation), whereas other pairwise comparisons of inflammation and Lgr5 epithelial expression resulted in no significant correlation.

Quantitative Analysis of 8OHdG in Lgr5-Positive and Negative Cells in Gastric Mucosa

The finding of increased numbers of Lgr5-positive epithelial cells in H. pylori-positive patients, particularly in the gastric cancer patient group, suggests that these cells may biologically respond to H. pylori infection. Therefore, we hypothesized that they may be susceptible to genomic injury associated with oxidative stress induced by the active inflammatory response to H. pylori infection and possibly by the bacterium or its released products. We performed quantitative analysis of 8OHdG as a marker of oxidative DNA damage in the nuclei of Lgr5-positive cells compared with Lgr5-negative cells in the non-neoplastic antral mucosa, by using quantitative spectral image analysis. Tissue sections were co-stained with immunofluorescence to detect 8OHdG and Lgr5. 8OHdG levels in Lgr5-positive versus Lgr5-negative epithelial cells in 33 cases including patients with or without gastric cancer and with or without H. pylori gastritis are summarized in Table 2.
Table 2

Levels of 8OHdG in the nuclei of epithelial cells in antral mucosa

Epithelial cells (EP)

Mean 8OHdG levels, SD

P value

All casesa: N = 33

Lgr5+ cells

264.97 (148.7)

.073

Lgr5− cells

247.44 (136.8)

Hp+ (GC+ and GC−): N = 18

Lgr5+ cells

200.15 (117.5)

.087

Lgr5− cells

179.15 (97.0)

Hp− (GC+ and GC−): N = 15

Lgr5+ cells

342.76 (147.9)

.414

Lgr5− cells

329.39 (134.8)

GC+ (Hp+ and Hp−): N = 16

Lgr5+ cells

313.98 (165.6)

.049

Lgr5− cells

287.72 (157.9)

GC− (Hp+ and Hp−): N = 17

Lgr5+ cells

218.84 (117.8)

.525

Lgr5− cells

209.53 (104.5)

GC+ and HP+: N = 9

Lgr5+ cells

263.59 (126.9)

.012

Lgr5− cells

215.00 (106.5)

GC− and Hp+: N = 9

Lgr5+ cells

136.70 (64.8)

.611

Lgr5− cells

143.00 (75.9)

GC+ and Hp−: N = 7

Lgr5+ cells

378.77 (195.9)

.878

Lgr5− cells

381.22 (170.5)

GC− and Hp−: N = 8

Lgr5+ cells

311.25 (284.1)

.341

Lgr5− cells

92.17 (80.0)

aAll cases include: GC+, GC−, Hp+, and Hp− cases

Variable levels of 8OHdG were detected in Lgr5-positive and negative epithelial cells (Fig. 2). The overall 8OHdG maximum fluorescence intensity in Lgr5-positive epithelial cells (mean 264.97) was higher than in Lgr5-negative epithelial cells (mean 247.44) but did not reach statistical significance (P = .073), Table 2. In the gastric mucosa of patients with gastric cancer, significantly higher levels of 8OHdG were detected in Lgr5-positive epithelial cells than in Lgr5-negative epithelial cells (P = .049), Table 2. Further, among gastric cancer patients with H. pylori gastritis, 8OHdG levels were higher in Lgr5-positive than in Lgr5-negative epithelial cells (P = .012), Table 2. The 8OHdG maximum fluorescence intensity in Lgr5-positive epithelial cells was higher in H. pylori-positive patients, irrespective of gastric cancer status, but did not reach statistical significance (P = .087). No significant differences were found for other group comparisons (Table 2). These data further suggest an interaction of H. pylori with host susceptibility factors that affect the response of Lgr5-positive epithelial cells in gastric cancer patients.
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2360-8/MediaObjects/10620_2012_2360_Fig2_HTML.jpg
Fig. 2

Detection of Lgr5 and 8OHdG in epithelial cells of gastric glands by immunofluorescence. a In the glandular epithelial cells, 8OHdG fluorescence (red punctate nuclear stain) was detected in nuclei of Lgr5-positive (green cytoplasmic stain, long arrow), and Lgr5-negative cells (short arrow). b In the lamina propria, 8OHdG was detected in nuclei of Lgr5-positive (long arrows) and Lgr5-negative cells (short arrows). Nuclei are stained with DAPI. Magnification ×400

In the mucosa of intestinal metaplasia-positive cases the difference between 8OHdG maximum fluorescence intensity of Lgr5-positive epithelial cells (mean 289.71, Std Dev 147.24) and 8OHdG maximum fluorescence intensity of Lgr5-negative epithelial cells (mean 268.48, Std Dev 127.91), was not statistically significant (P = .144).

Examples of Lgr5-positive cells in the lamina propria with 8OHdG accumulation are shown in Fig. 2. The overall 8OHdG maximum fluorescence intensity in lamina propria Lgr5-positive cells (mean 268.8) was significantly higher than in Lgr5-negative lamina propria cells (mean 247.67), P = .004 (Supplement Table 5). Surprisingly, the levels of 8OHdG were higher in Lgr5-positive lamina propria cells in patients who were H. pylori-negative and gastric cancer-negative.

Co-Expression of CD45, Cytokeratin, CD34, CD44, and Vimentin in Lgr5-Positive Cells in Gastric Mucosa

Both immunohistochemistry and immunofluorescence staining of gastric mucosa with anti-Lgr5 antibody showed that, in addition to Lgr5-positive epithelial cells, there were rare scattered Lgr5-positive cells in the lamina propria, particularly in patients with H. pylori gastritis. To determine whether the Lgr5-positive cells in the gastric glands and in the lamina propria had features of epithelial, leukocyte or endothelial lineages we evaluated co-expression of CD45, cytokeratin AE1/AE3, CD34, and Lgr5 in three representative gastric antral mucosa samples of patients with H. pylori gastritis (Table 3).
Table 3

Co-expression of cytokeratin AE1/AE3 (CK), CD45, or CD34 in Lgr5-positive cells within the glandular epithelium (GE) and lamina propria (LP), detected by immunofluorescence staining

Co-expression of Lgr5, CK, CD45 (%)

Case SR50

Case SR67

Case SR144

GE: CK+ Lgr5+ cells

22/22 (100)

32/32 (100)

19/19 (100)

GE: CD45+ Lgr5+ cells

0/39 (0)

0/30 (0)

0/16 (0)

GE: CD34+ Lgr5+ cells

0/24 (0)

0/32 (0)

0/21 (0)

LP: CK+ Lgr5+ cells

0/19 (0)

0/18 (0)

0/11 (0)

LP: CD45+ Lgr5+ cells

18/18 (100)

17/22 (72.3)

10/10 (100)

LP: CD34+ Lgr5+ cells

0/16 (0)

0/18 (0)

0/20 (0)

The numbers of CK, CD45, or CD34-positive cells co-expressing Lgr5 were characterized in antral mucosa of three H. pylori-positive cases (SR50, SR67, and SR144)

In the epithelium of gastric glands, Lgr5-positive cells were always positive for cytokeratin, confirming an epithelial lineage (Table 3, Fig. 3). Further, in gastric epithelium, Lgr5-positive cells did not co-express CD45 (Table 3, Fig. 4), CD34, or vimentin (Table 3 and Supplement Figs. 3 and 4). To determine whether epithelial Lgr5+ cells co-express other stem cell markers we performed immunofluorescence staining of Lgr5 and CD44 in the same sections (Supplement Figs. 5 and 6).We found that most CD44+ epithelial cells were Lgr5−, and essentially all Lgr5+ cells seemed to co-express CD44 (Supplement Figs. 5 and 6) .
https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2360-8/MediaObjects/10620_2012_2360_Fig3_HTML.jpg
Fig. 3

Co-localization of Lgr5 and cytokeratin in epithelial cells of gastric antral glands. Panel a: Overlapped immunofluorescence images show co-localization of Lgr5 and cytokeratin in epithelial cells (yellow cells, short arrow). Panel a was created by overlapping the cytokeratin AE1/AE3 immunofluorescence stain (shown in panel b), where red stain and the long thin arrow indicate cytokeratin positive cells, with the Lgr5 immunofluorescence stain (shown in panel c), where green stain and the long thick arrow indicate Lgr5-positive epithelial cells. Panel d: Overlapped immunofluorescence images show lack of co-localization of cytokeratin in Lgr5-positive lamina propria cells (short arrows). The overlapped images in panel d were the cytokeratin immunofluorescence stain (shown in panel e), and the Lgr5 immunofluorescence stain (shown in panel f). Two Lgr5-positive lamina propria cells are indicated by the long thick arrows in panel f, and these two cells are negative for cytokeratin (panel e, long thin arrows). Nuclei are stained with DAPI. Original magnification ×400

https://static-content.springer.com/image/art%3A10.1007%2Fs10620-012-2360-8/MediaObjects/10620_2012_2360_Fig4_HTML.jpg
Fig. 4

Lgr5-positive cells in lamina propria of gastric antrum co-express CD45. Panel a. Overlapped immunofluorescence images show lack of co-localization of CD45 in Lgr5-positive epithelial cells (short arrows). Panel a was created by overlapping the CD45 stain (shown in panel b) with the Lgr5 stain (shown in panel c). Lgr5-positive epithelial cells stain green and are indicated by long thick arrows in panel c. The CD45 stain is negative in Lgr5-positive epithelial cells, as indicated by the long thin arrows in panel b. Panel d was created by overlapping the CD45 stain (shown in panel e) with the Lgr5 stain (shown in panel f) and shows co-localization of CD45 in Lgr5-positive lamina propria cells (yellow cells, short arrows). Lgr5-positive lamina propria cells stain green and are indicated by long thick arrows in panel f. The CD45 stain is positive in Lgr5-positive lamina propria cells, as indicated by the long thin arrows in panel e. Nuclei are stained with DAPI. Original magnification ×400

In the lamina propria, most Lgr5-positive cells co-expressed CD45, indicating they are of leukocyte lineage (Table 3, Fig. 4). This is consistent with their plasmacytoid morphology and location amid the inflammatory infiltrate in the lamina propria. In two cases (SR50 and SR54), all lamina propria Lgr5-positive cells co-expressed CD45, whereas in case SR67 a proportion of lamina propria Lgr5-positive cells did not (27.7 %), suggesting that there may be another lineage contributing to the Lgr5 cell population in the gastric mucosa (Table 3 and Supplement Fig. 7). Lamina propria Lgr5+ cells were found to be vimentin positive, consistent with the leukocyte lineage indicated by the fact that these cells are also CD45+, however most vimentin+ lamina propria cells were Lgr5-negative (Supplement Fig. 4).

Discussion

This is the first study characterizing the distribution and oxidative DNA damage of Lgr5-positive epithelial cells, a population of gastric epithelial stem cells in human stomach, and their association with H. pylori gastritis and gastric cancer. A contribution of epithelial stem cells to gastric carcinogenesis, including H. pylori-associated carcinogenesis, has been inferred from studies of mouse models [10, 11, 21, 22] and, therefore, involvement of epithelial stem cells in the development of human gastric cancer has been proposed [23].

In this study, we observed that Lgr5-positive epithelial cells occur almost exclusively in the human gastric antrum, whereas smaller numbers were observed in the oxyntic mucosa of the gastric body or fundus. Similar to data reported recently by Simon et al. [12], we identified small numbers of Lgr5-positive cells in mucous glands of the gastric mucosa. Whereas we observed higher numbers of Lgr5+ cells in the deep mucous glands than in the neck of these glands, Simon et al. reported higher numbers in the neck regions. This difference in distribution may be explained by the fact that the two studies used different antibodies against Lgr5, which could detect different Lgr5 protein variants, or, alternatively, by differences in the assessment of the neck region versus deeper glandular region, as these are not rigidly delimited in the mucosa.

The numbers of Lgr5-positive epithelial cells were increased in the antral mucosa of patients with H. pylori gastritis and gastric cancer but not in patients without gastric cancer. This finding suggests that H. pylori infection-associated factors may disturb the normal regulation of Lgr5-positive epithelial stem cells, leading to alteration of the normal pattern of asymmetric versus symmetric division, with a possible increase in symmetric division. Lineage tracing experiments in the Lgr5-EGFP-IRES-creERT2 mouse model may be useful to address the underlying mechanisms. The presence of multiple Lgr5-positive cells in gastric glands has also been found in mouse stomach and it is still unclear how many of these are bonafide stem cells. Lgr5 may mark a somewhat broader range of progenitors with proliferative ability, including those with less long-term self-renewal potential [10].

The near absence of Lgr5-positive cells in the gastric body suggests that other stem cell markers are needed to identify the stem cell population of oxyntic glands in the gastric body and fundus. In the mouse model, Barker et al. [11] showed that Lgr5 was expressed at the base of the gastric glands and that Lgr5-positive cells had the capacity of self-renewal and could build long-lived gastric antral units in vivo. Consistent with the findings reported in the mouse model, we observed that most Lgr5-positive cells were located at the base of gastric antral glands in the human gastric mucosa, and confirmed a similar pattern of expression in the mouse stomach. We also showed that the gastric cancer cell line SNU638 expresses Lgr5, providing a cell culture model for future studies of the involvement of Lgr5 in gastric carcinogenesis. Future studies to determine whether co-cultures of AGS or SNU638 cells with H. pylori increase the number of Lgr5-positive cells, as was seen in patients infected with H. pylori, and whether different virulence types of H. pylori, for example the cagPAI, have differential effects in Lgr5+ cells, are warranted.

Although the association between H. pylori and gastric cancer is well established [5, 24, 25], the critical molecular mechanisms and cellular targets that drive the development and progression of gastric cancers remain poorly understood. Factors known to be involved in H. pylori gastric carcinogenesis include the inflammation associated with the infection, increased oxidative DNA damage [26], altered DNA repair in the epithelium, and increased mutagenesis of the epithelial genome [7, 8, 27]. In this study we showed that oxidative DNA damage of Lgr5-positive epithelial cells in gastric mucosa, determined by the levels of nuclear 8OHdG, was increased compared with Lgr5-negative epithelial cells in patients with gastric cancer and H. pylori infection but not in gastric cancer patients without H. pylori infection. These findings are consistent with results showing that inflammation associated with H. pylori infection increased oxidative DNA damage [26], and impaired DNA repair in the epithelium with increased mutagenesis of the epithelial genome [7, 8, 27]. Further, our results suggest that oxidative damage of Lgr5-positive epithelial cells may be mediated by the interaction of H. pylori with host susceptibility factors that affect the response of Lgr5-positive epithelial cells in gastric cancer patients. Because oxidative damage can lead to pathogenic mutations that may reprogram cells to a neoplastic pathway, the oxidative DNA damage caused by H. pylori infection in Lgr5-positive epithelial stem cells of particularly susceptible individuals may be a critical mechanism of gastric carcinogenesis. This notion is consistent with reported lineage-tracing studies in mice showing that a deletion mutant of the adenomatous polyposis coli gene (APC) in Lgr5-positive gastric progenitor cells was sufficient to give rise to antral tumors, whereas APC deletion in more differentiated cells did not [10, 11].

Of interest, the presence of intestinal metaplasia, which is a metaplastic cell lineage in the early sequence that leads from H. pylori gastritis to gastric carcinoma, did not result in significant differences between oxidative DNA damage of Lgr5-positive versus negative epithelial cells.

We also observed that, in addition to Lgr5-positive epithelial cells within the gastric glands, there were rare scattered Lgr5-positive cells in the lamina propria, which had plasmacytoid morphology. These findings are similar to those previously reported in colonic mucosa [14], where rare scattered Lgr5-positive cells were observed in the lamina propria, suggesting they may be hematopoietic lineage-derived cells, because mRNA-expression studies have shown Lgr5 is expressed in bone marrow [28].

To characterize the cellular lineages of Lgr5-positive cells in the gastric mucosa we performed co-localization immunofluorescence studies. We showed that all Lgr5-positive cells in the gastric glands co-stained with cytokeratin and did not co-stain with vimentin, indicating their epithelial lineage. Co-immunofluorescence staining with Lgr5 and CD44, another gastric stem cell marker, led to findings similar to those reported by Simon et al. [12], in that most CD44+ epithelial cells were Lgr5−, and essentially all Lgr5+ cells seemed to co-express CD44, consistent with a gastric stem cell lineage. In contrast, most Lgr5-positive cells in the lamina propria co-expressed CD45 and vimentin, indicating a leukocyte lineage, with co-expression of CD45 absent from a few cells. The rare CD45− Lgr5+ cells may be an additional Lgr5-positive cell lineage, for example of hematopoietic origin. Surprisingly, the levels of 8OHdG in Lgr5-positive cells in the lamina propria were higher in patients negative for gastric cancer and H. pylori, contrasting with the response seen for Lgr5-positive cells in gastric glands. The significance of this finding is unclear; it is, however, possible that the Lgr5-positive lamina propria cells undergo faster turnover, and thus accumulate less 8OHdG adducts in patients with H. pylori gastritis and gastric cancer, for whom variable degrees of inflammation in the background gastric mucosa may be observed.

In summary, this is the first study characterizing the distribution of Lgr5-positive cells in the human stomach and implicating Lgr5-positive epithelial stem cells in the adaptive response during H. pylori gastritis. This is characterized by an increase in the number of possibly dysregulated Lgr5-positive epithelial stem cells, as they show increased sensitivity to oxidative DNA damage in response to H. pylori infection, particularly in patients with co-existing gastric cancer. Overall our results suggest involvement of Lgr5-positive epithelial stem cells in H. pylori-associated gastric carcinogenesis.

Conflict of interest

None.

Supplementary material

10620_2012_2360_MOESM1_ESM.docx (29 kb)
Supplementary material 1 (DOCX 29 kb)
10620_2012_2360_MOESM2_ESM.tif (9.8 mb)
Supplementary material 2 (TIFF 10023 kb)
10620_2012_2360_MOESM3_ESM.tif (24.5 mb)
Supplementary material 3 (TIFF 25119 kb)
10620_2012_2360_MOESM4_ESM.tif (24.8 mb)
Supplementary material 4 (TIFF 25390 kb)
10620_2012_2360_MOESM5_ESM.tif (6.7 mb)
Supplementary material 5 (TIFF 6833 kb)
10620_2012_2360_MOESM6_ESM.tif (8 mb)
Supplementary material 6 (TIFF 8158 kb)
10620_2012_2360_MOESM7_ESM.tif (7 mb)
Supplementary material 7 (TIFF 7180 kb)
10620_2012_2360_MOESM8_ESM.tif (31.6 mb)
Supplementary material 8 (TIFF 32346 kb)
10620_2012_2360_MOESM9_ESM.docx (35 kb)
Supplementary material 9 (DOCX 35 kb)

Copyright information

© Springer Science+Business Media, LLC 2012