Takeuchi C, et al. J Gastroenterol 2022;57:144-155.
2022년 1월, J Gastroenterol에 실린 내용입니다. AIG와 관련하여 epigenetics (후생유전학) 관련 첫 report 입니다. AIG는 chronic inflammation 중 하나이며, H. pylori보다는 risk가 낮지만 gastric cancer 발생과 연관성이 있다고 알려져 있습니다. 비 정상적인 DNA 메틸화는 여러 형태의 cancer에서 중요한 원인으로 알려져 있으며, H. pylori 감염은 위 점막에서 비 정상적인 DNA 메틸화를 유도하는 것으로 알려져 있습니다. 대표적으로 NF-kB활성으로 인한 TET(TET gene methylcytosine dioxygenases) expression의 저하와 NO(nitric oxide)에 노출되면서 DNMT(DNA methyltransferases) activation의 증가가 비 정상적인 DNA 메틸화로 알려져 있습니다 (아래 Figure, Takeshima H, et al. Int J Invest 2020.). 아직 AIG에서는 이러한 DNA의 비 정상적인 메틸화가 일어나는지 여부는 알려져 있지 않기 때문에 H. pylori 감염이 없는 AIG 환자의 위 점막에서 DNA 메틸화가 유도되는지, 정상인과 H. pylori associated gastritis (HPG) 환자와 비교하였고, 위 점막의 염증 반응과 관련성을 알아보는 연구 입니다.
** 여기서, DNA 메틸화는 무엇이고 CpG는 무엇인지.. 에 대해 간략하게 정의하고 가겠습니다.
CpG island: CG site라는 것은 말 그대로 C(cytosine)와 G(Guanine)가 일렬로 나란히 있는 서열을 말하며, CpG는 단순한 서열의 나열을 의미하는 것이 아니라 C와 G가 phosphate 하나로 연결된 부위를 의미합니다. 프로모터의 경우 대부분의 C는 메틸화가 되지 않습니다. 이는 발현을 위해서는 메틸화가 억제되어야 하기 때문입니다.
메틸화: CpG의 C(cytosine) 염기에는 메틸기(-CH3)가 결합하여 있는데, 이를 ‘DNA 메틸화(Methylation)’라고 부릅니다. 이러한 DNA 메틸화 현상은 유전체의 각종 반복 서열 등에서 흔히 관찰되며, 이는 유전체의 안정성 유지 등에 중요한 역할을 하는 것으로 보고 있습니다.
Gene silence: 항상 발현 유전자들의 활성을 반영구적으로 정지시켜야 할 필요가 있을 때, 그들 프로모터의 CpG에 메틸화가 일어나며, CpG 부위 거의 모두가 메틸화되어 버리면 유전자의 활성은 완전히 멈추게 되며, 이러한 경우를 유전자 침묵(gene silence)이라고 합니다.
CpG island의 메틸화와 암 발생: 생식계열 세포 즉 정자나 난자 유전체의 CpG island는 대체로 메틸화가 되어 있지 않습니다. 하지만 수정 후 배아 발달과정을 진행하는 체세포에서는 CpG island를 가진 유전자 일부가 메틸화가 됩니다. 이는 정상 배아 발달이나 세포분화에 필수적입니다. 암 발생의 경우에서도 이를 억제하는 유전자(tumor suppressor genes)의 CpG 부위에 메틸화가 일어나면 유전자는 silence 상태가 되어 무작위적으로 발생하는 암을 막지 못하게 되어 문제가 될 수 있습니다.
Study design and tissue sample collection (Multicenter observational study, University Hospital Medical Network (UMIN) 000039528)
- AIG (APCA 양성) witout H. pylori (anti-H. pylori antibody 측정)
- H. pylori associated gastritis (HPG)
- Healthy control
로 구분하였으며, gastrectomy를 받았거나 H. pylori eradication을 받은 환자는 제외하였습니다. AIG 환자의 serum gastrin levels은 모두 > 1000 pg/ml 이였으며, endoscopy 에서 AIG과 HPG group 모두 opent type Kimura-Takemoto classification (AIG는 모두 O-III) 을 보이고 있었고, 조직검사는 greater curvature in the middle corpus 에서 시행하였습니다.
Figure 1. Representative cases and flowchart of the study.
a. Representative cases with autoimmune gastritis (AIG) and H. pylori associated gastritis (HPG). Endoscopic view of gastric body and antrum, and H&E staining of gastric body (x 100).
b. Flow chart of the study UMIN000039528. From March 2020 to November 2021, a total of 48 people were enrolled. Eleven people were excluded due to insufficient clinical data for diagnosis, and 12 people with AIG without H. pylori infection, 14 with HPG, and 10 healthy volunteers were included. 11 AIG samples, 9 HPG samples, and 4 normal samples were selected for genome-wide DNA methylation analysis. 8 of each of the AIG samples, HPG samples, and normal samples were used for quantitative RT-PCR.
Table S1. Characteristics of people enrolled in sutdy UMIN000039528.
Data acquisition from DNA methylation array
Methylation levels were prsented by a β value: 0 (unmethylated), 1 (fully methylated)
Estimation and corredction of leukocyte fractions
Estimation and correction of leukocyte fractions (3 leukocyte groups)
: T cells (CD4+, CD8+ T cells, CD56+ NK cells)
: B cells (CD19+)
: Monocytes/granulocytes (CD14+ monocytes, neutrophils, eosinophil)
3 groups에서 methylated CpG sites를 selection 하였고, fractions을 계산하였습니다.
Informatics of DNA methylation analysis
총 439,031 probes에서 Transcription start sites (TSSs)와 CpG islands (CGIs)에 대한 locations에 따라 probes를 collection 하였습니다.
Promotor CpG island에서 genomic block의 CpG sites를 분석하였으며, CDKN2A (p16)를 확인하기 위해 TSS의 genomic block을 바로 downstream 시켜서 분석하였습니다 (TSS200에 위치하는 genomic block이 없었기 떄문).
Gene ontology analysis (GSEA v4.1.0 & DAVID)
Quntitative real-time PCR (expresstion levels of inflammation-related genes)
결론을 살펴 보겠습니다.
AIG displayed unique methylation profiles against HPG
* DNA methylation (CpG sites, (P < 0.05), infiltrating leukocytes 포함)
AIG: Hypermethylated (256,440), Hypomethylated (188,276) / HPG: Hypermethylated (283,295), Hypomethylated (202,303)
* DNA methylation (CpG sites, (P < 0.05), infiltrating leukocytes 제거)
AIG: Hypermethylated (122,444), Hypomethylated (87,241) / HPG: Hypermethylated (135,671), Hypomethylated (102,022)
AIG와 HPG의 samples 중에서 가장 높은 SD를 가진 20,000 CpG sites (gastric cancer 포함)은 Nromal 과 비교하여 비정상적인 DNA methylation을 보였습니다 (Fig. 2a. Left). Leukocyte contamination으로 인한 영향을 제거한 후 DNA methylation은 AIG와 HPG에서 서로 차이를 보였습니다 (Fig. 2a. Right). AIG와 HPG사의 관계를 비교해 보았을 때, gastric cancer samples에서 증가된 DNA methylation 부위에 서로간 차이가 있었습니다 (Fig. S1a).
Leukocyte fractions을 살펴보면, T cell infiltration은 AIG와 HPG samples이 Normal 보다 높았고, monocyte/granulocyte와 B-cell infiltration은 AIG 보다 HPG에서 높게 나타나서 (mean: 4.9% and 16.0% vs. 1.8% and 9.1%, P=0.01 and <0.01; Fig. 2b), HPG samples에서 monocyte/granulocyte와 B-cell infiltration의 정도가 stronger methylation induction과 관련이 있다고 볼 수 있습니다.
Figure2. Results of genome-wide DNA methylation analysis among the AIG (n = 11), HPG (n = 9), and normal samples (n = 4), including the samples with gastric cancer.
(a) Unsupervised hierarchical cluster analysis of methylation levels of the three groups of samples. β values of the 20,000 CpG sites with the highest standard deviation (SD) in overall CpG sites were used. Before correction of leucocyte fractions (left), the AIG and HPG samples were not clearly separated. After correction (right), the AIG samples presented unique methylation profiles against the HPG samples.
(b) Leucocyte fractions in the samples. Monocyte/granulocyte and B-cell infiltration was prominent in the HPG samples. Both AIG and HPG samples showed increased infiltration of T cells, compared with the normal samples. *P < 0.05, †P < 0.01
Figure S1. Results of genome-wide DNA methylation analysis among the AIG (n=11), HPG (n=9), and normal samples (n=4), including the samples with gastric cancer.
(a) Unsupervised hierarchical cluster analysis of methylation levels of AIG and HPG samples after correction of leucocyte fractions. β values of the 20,000 CpG sites with the highest standard deviation (SD) in overall CpG sites were used. The sample(s) with gastric cancer were clustered separately with unique methylation patterns compared with the samples without gastric cancer, and a cluster of CpG sites (shown by a yellow rectangle) had increased DNA methylation.
(b) Leucocyte fractions in the samples. The fractions of the three leucocyte groups (T cells, B cells, and monocytes/granulocytes) were calculated based upon their specific methylation patterns.
AIG showed weaker metylation induction than HPG
Figure 3. The weaker induction of aberrant DNA methylation by AIG than by HPG.
(a) Scatter plot analysis using the methylation levels in overall CpG sites and in promoter CpG islands. Both the AIG (n = 11) and HPG (n = 9) samples showed aberrant DNA methylation, compared with the normal samples (n = 4) (left). Comparison between the AIG and HPG samples showed that AlG-induced methylation, but less than HPG (right).
(b) Volcano plot analysis using the methylation levels of genomic blocks in promoter CpG islands. The number of genomic blocks with large Ap values (> 0.05) and small P values (-1og10(P values) > 1.301) greatly increased both in the AIG and HPG samples compared with the normal samples. Moreover, the number of genomic blocks more methylated was remarkably higher in the HPG samples than in the AIG samples. 
Figure S2. The weaker induction of aberrant DNA methylation by AIG than by HPG.
(a) Scatter plot analysis before correction of leucocyte fractions. The methylation levels in overall CpG sites and in promoter CpG islands were analyzed between the AIG (n=11), HPG (n=9), and normal samples (n=4). Both the AIG and HPG samples showed aberrant DNA methylation, compared with the normal samples (left). Comparison between the AIG and HPG samples showed that HPG further induced methylation upon AIG (right).
(b) Volcano plot analysis before correction of leucocyte fractions. The number of genomic blocks with large Δβ values (≥ 0.05) and small P values (–log10(P values) > 1.301) greatly increased both in the AIG and HPG samples compared with the normal samples. Moreover, the number of genomic blocks more methylated was remarkably higher in the HPG samples than in the AIG samples.
Driver genes were hypermethylation in AIG, but more so in HPG
4가지 Tumor-suppressor genes의 methylation levels을 비교하였습니다.
* CDKN2A (p16): methylated in gastric cancer (higher AIG < HPG)
* CDH1 (E-cadherin): inactivation is associated with diffuse-type gastric cancer (higher AIG < HPG)
* IGFBP7 & MIR34B: p53 pathway (higher AIG < HPG)
* DKK3 & SFRP1: negative regulator genes (higher AIG < HPG)
* MLH1: mismatch repair genes (little methylation)
--> AIG samples 에서 (HPG에서는 보다 더) cancer driver genes의 methylatio이 높다는 것이고, AIG는 상대적으로 HPG보다 낮은 carcinogenic potential을 갖는 다는 것입니다.
Figure 4. The DNA methylation levels of cancer driver genes among the AIG (n = l1), HPG (n = 9), and normal samples (n = 4) after correction of leucocyte fractions. (a) CDKN2A was highly methylated in the AIG samples, but more so in the HPG samples. (b) CDH1 was highly methylated in the AIG samples, but more so in the HPG samples. (c) IGFBP7 and MIR34B showed high methylation levels in the AIG samples, but more so in the HPG samples. (d) DKK3 and SFRP1 showed high methylation levels in the AIG samples, but more so in the HPG samples. (e) MLH1 showed little methylation in any types of the samples. †P < 0.01, n.s. not significant
Inflammation-related genes were more upregulated in HPG and in AIG
4가지 inflammation-related genes의 methylation levels을 비교하였습니다.
TNFα, IL1B, NOS2: DNA methylation induction에 관여하는 물질들입니다
* TNFα: secreted by all types of leucocytes (highly expressed in the AIG & HPG)
* IL1B, IL8: secreted by macrophage (induce neutrophil migration) (highly expressed in the HPG: likely reflecting the stronger monocyte/granulocyte infiltration)
* NOS2: (highly expressed in the AIG (AIG vs. HPG: no difference)
Figure 5. The expression levels of four infl ammation-related genes among the AIG (n = 8), IIPG (n = 8), and normal samples (n = 8). IL1B and IL8 showed remarkably high expression levels in the HPG samples. TNFα and NOS2 showed high expression levels in the AIG samples, but no significant difference between the AIG and HPG samples. *P < 0.05, †P < 0.01, n.s. not significant
이번 연구들을 살펴보면, AIG의 경우 위 점막에서 비 정상적인 DNA 메틸화가 일어난다는 것이고, HPG와 비교하여 독립적인 methylation profiles이 존재한다는 것입니다. 그리고, AIG-induced gastric cancer에 epigenetic metahnism이 작용한다는 것을 의미하기도 합니다. 하지만, AIG는 HPG 보다는 비 정상적인 DNA 메틸화가 덜 일어나는 것 같고, H. pylori infection이 없는 AIG에서는 동반된 AIG 보다 더 낮은 gastric cancer incidence를 보일 수 있다는 것입니다. 또한, DNA 메틸화를 유도하는 것에 있어서 T 또는 B cells 보다 macrophages와 neutrophils이 더 중요하다는 것이고, macrophage-induced inflammation이 DNA 메틸화 유도의 main mechanism 이라는 것을 알 수 있었습니다. 앞으로 AIG환자에서 HPG가 함께 동반 되었을 때 발생하는 DNA 메틸화 유도에 미치는 영향과 gastric cancer의 높은 incidence 연관성 등이 연구되어져야 할 것 같습니다.
Abstract
Background Autoimmune gastritis (AIG) is a chronic inflammatory condition in gastric mucosa and is associated with increased cancer risk, though not as high as that by Helicobacter pylori (H. pylori)-associated gastritis (HPG), Although aberrant DNA methylation is induced by HPG and the level correlates with the risk of gastric cancer, DNA methylation induction by AIG is unknown.
Methods Gastric mucosa samples from the corpus were collected from 12 people with AIG without H. pylori infection, 10 people with IIPG, and eight healthy volunteers. Genome-wide DNA methylation analysis was conducted using Infinium Methylation EPIC array. Gene expression was analyzed by quantitative RT-PCR.
Results The AIG samples had extensive aberrant DNA methylation but presented unique methylation profiles against the HPG samples after correction of leucocyte fractions. Comparison between the AIG and HPG samples showed that AIG induced methylation, but less than HPG, in overall CpG sites and also in promoter CpG islands. Promoter CpG islands of tumor-suppressor genes in the pathway of cell cycle, cell adhesion, p53, and WNT were highly methylated in the AIG samples, but more so in the HPG samples. The expression levels of ILIB and, Il3, secreted by macrophage, were significantly lower in the AIG samples than in the HPG samples, suggesting that a difference in inflammatory response affected the degree and pattems of aberrant DNA methylation.
Conclusions AIG induced aberrant DNA methylation in gastric mucosa. However, the degree of DNA methylation was less than that by HPG, which reflected carcinogenic risk.
Takeuchi C, et al. J Gastroenterol 2022;57:144-155.
2022년 1월, J Gastroenterol에 실린 내용입니다. AIG와 관련하여 epigenetics (후생유전학) 관련 첫 report 입니다. AIG는 chronic inflammation 중 하나이며, H. pylori보다는 risk가 낮지만 gastric cancer 발생과 연관성이 있다고 알려져 있습니다. 비 정상적인 DNA 메틸화는 여러 형태의 cancer에서 중요한 원인으로 알려져 있으며, H. pylori 감염은 위 점막에서 비 정상적인 DNA 메틸화를 유도하는 것으로 알려져 있습니다. 대표적으로 NF-kB활성으로 인한 TET(TET gene methylcytosine dioxygenases) expression의 저하와 NO(nitric oxide)에 노출되면서 DNMT(DNA methyltransferases) activation의 증가가 비 정상적인 DNA 메틸화로 알려져 있습니다 (아래 Figure, Takeshima H, et al. Int J Invest 2020.). 아직 AIG에서는 이러한 DNA의 비 정상적인 메틸화가 일어나는지 여부는 알려져 있지 않기 때문에 H. pylori 감염이 없는 AIG 환자의 위 점막에서 DNA 메틸화가 유도되는지, 정상인과 H. pylori associated gastritis (HPG) 환자와 비교하였고, 위 점막의 염증 반응과 관련성을 알아보는 연구 입니다.
** 여기서, DNA 메틸화는 무엇이고 CpG는 무엇인지.. 에 대해 간략하게 정의하고 가겠습니다.
CpG island: CG site라는 것은 말 그대로 C(cytosine)와 G(Guanine)가 일렬로 나란히 있는 서열을 말하며, CpG는 단순한 서열의 나열을 의미하는 것이 아니라 C와 G가 phosphate 하나로 연결된 부위를 의미합니다. 프로모터의 경우 대부분의 C는 메틸화가 되지 않습니다. 이는 발현을 위해서는 메틸화가 억제되어야 하기 때문입니다.
메틸화: CpG의 C(cytosine) 염기에는 메틸기(-CH3)가 결합하여 있는데, 이를 ‘DNA 메틸화(Methylation)’라고 부릅니다. 이러한 DNA 메틸화 현상은 유전체의 각종 반복 서열 등에서 흔히 관찰되며, 이는 유전체의 안정성 유지 등에 중요한 역할을 하는 것으로 보고 있습니다.
Gene silence: 항상 발현 유전자들의 활성을 반영구적으로 정지시켜야 할 필요가 있을 때, 그들 프로모터의 CpG에 메틸화가 일어나며, CpG 부위 거의 모두가 메틸화되어 버리면 유전자의 활성은 완전히 멈추게 되며, 이러한 경우를 유전자 침묵(gene silence)이라고 합니다.
CpG island의 메틸화와 암 발생: 생식계열 세포 즉 정자나 난자 유전체의 CpG island는 대체로 메틸화가 되어 있지 않습니다. 하지만 수정 후 배아 발달과정을 진행하는 체세포에서는 CpG island를 가진 유전자 일부가 메틸화가 됩니다. 이는 정상 배아 발달이나 세포분화에 필수적입니다. 암 발생의 경우에서도 이를 억제하는 유전자(tumor suppressor genes)의 CpG 부위에 메틸화가 일어나면 유전자는 silence 상태가 되어 무작위적으로 발생하는 암을 막지 못하게 되어 문제가 될 수 있습니다.
Study design and tissue sample collection (Multicenter observational study, University Hospital Medical Network (UMIN) 000039528)
- AIG (APCA 양성) witout H. pylori (anti-H. pylori antibody 측정)
- H. pylori associated gastritis (HPG)
- Healthy control
로 구분하였으며, gastrectomy를 받았거나 H. pylori eradication을 받은 환자는 제외하였습니다. AIG 환자의 serum gastrin levels은 모두 > 1000 pg/ml 이였으며, endoscopy 에서 AIG과 HPG group 모두 opent type Kimura-Takemoto classification (AIG는 모두 O-III) 을 보이고 있었고, 조직검사는 greater curvature in the middle corpus 에서 시행하였습니다.
Figure 1. Representative cases and flowchart of the study.
a. Representative cases with autoimmune gastritis (AIG) and H. pylori associated gastritis (HPG). Endoscopic view of gastric body and antrum, and H&E staining of gastric body (x 100).
b. Flow chart of the study UMIN000039528. From March 2020 to November 2021, a total of 48 people were enrolled. Eleven people were excluded due to insufficient clinical data for diagnosis, and 12 people with AIG without H. pylori infection, 14 with HPG, and 10 healthy volunteers were included. 11 AIG samples, 9 HPG samples, and 4 normal samples were selected for genome-wide DNA methylation analysis. 8 of each of the AIG samples, HPG samples, and normal samples were used for quantitative RT-PCR.
Table S1. Characteristics of people enrolled in sutdy UMIN000039528.
Data acquisition from DNA methylation array
Methylation levels were prsented by a β value: 0 (unmethylated), 1 (fully methylated)
Estimation and corredction of leukocyte fractions
Estimation and correction of leukocyte fractions (3 leukocyte groups)
: T cells (CD4+, CD8+ T cells, CD56+ NK cells)
: B cells (CD19+)
: Monocytes/granulocytes (CD14+ monocytes, neutrophils, eosinophil)
3 groups에서 methylated CpG sites를 selection 하였고, fractions을 계산하였습니다.
Informatics of DNA methylation analysis
총 439,031 probes에서 Transcription start sites (TSSs)와 CpG islands (CGIs)에 대한 locations에 따라 probes를 collection 하였습니다.
Promotor CpG island에서 genomic block의 CpG sites를 분석하였으며, CDKN2A (p16)를 확인하기 위해 TSS의 genomic block을 바로 downstream 시켜서 분석하였습니다 (TSS200에 위치하는 genomic block이 없었기 떄문).
Gene ontology analysis (GSEA v4.1.0 & DAVID)
Quntitative real-time PCR (expresstion levels of inflammation-related genes)
결론을 살펴 보겠습니다.
AIG displayed unique methylation profiles against HPG
* DNA methylation (CpG sites, (P < 0.05), infiltrating leukocytes 포함)
AIG: Hypermethylated (256,440), Hypomethylated (188,276) / HPG: Hypermethylated (283,295), Hypomethylated (202,303)
* DNA methylation (CpG sites, (P < 0.05), infiltrating leukocytes 제거)
AIG: Hypermethylated (122,444), Hypomethylated (87,241) / HPG: Hypermethylated (135,671), Hypomethylated (102,022)
AIG와 HPG의 samples 중에서 가장 높은 SD를 가진 20,000 CpG sites (gastric cancer 포함)은 Nromal 과 비교하여 비정상적인 DNA methylation을 보였습니다 (Fig. 2a. Left). Leukocyte contamination으로 인한 영향을 제거한 후 DNA methylation은 AIG와 HPG에서 서로 차이를 보였습니다 (Fig. 2a. Right). AIG와 HPG사의 관계를 비교해 보았을 때, gastric cancer samples에서 증가된 DNA methylation 부위에 서로간 차이가 있었습니다 (Fig. S1a).
Leukocyte fractions을 살펴보면, T cell infiltration은 AIG와 HPG samples이 Normal 보다 높았고, monocyte/granulocyte와 B-cell infiltration은 AIG 보다 HPG에서 높게 나타나서 (mean: 4.9% and 16.0% vs. 1.8% and 9.1%, P=0.01 and <0.01; Fig. 2b), HPG samples에서 monocyte/granulocyte와 B-cell infiltration의 정도가 stronger methylation induction과 관련이 있다고 볼 수 있습니다.
Figure2. Results of genome-wide DNA methylation analysis among the AIG (n = 11), HPG (n = 9), and normal samples (n = 4), including the samples with gastric cancer.
(a) Unsupervised hierarchical cluster analysis of methylation levels of the three groups of samples. β values of the 20,000 CpG sites with the highest standard deviation (SD) in overall CpG sites were used. Before correction of leucocyte fractions (left), the AIG and HPG samples were not clearly separated. After correction (right), the AIG samples presented unique methylation profiles against the HPG samples.
(b) Leucocyte fractions in the samples. Monocyte/granulocyte and B-cell infiltration was prominent in the HPG samples. Both AIG and HPG samples showed increased infiltration of T cells, compared with the normal samples. *P < 0.05, †P < 0.01
Figure S1. Results of genome-wide DNA methylation analysis among the AIG (n=11), HPG (n=9), and normal samples (n=4), including the samples with gastric cancer.
(a) Unsupervised hierarchical cluster analysis of methylation levels of AIG and HPG samples after correction of leucocyte fractions. β values of the 20,000 CpG sites with the highest standard deviation (SD) in overall CpG sites were used. The sample(s) with gastric cancer were clustered separately with unique methylation patterns compared with the samples without gastric cancer, and a cluster of CpG sites (shown by a yellow rectangle) had increased DNA methylation.
(b) Leucocyte fractions in the samples. The fractions of the three leucocyte groups (T cells, B cells, and monocytes/granulocytes) were calculated based upon their specific methylation patterns.
AIG showed weaker metylation induction than HPG
Figure 3. The weaker induction of aberrant DNA methylation by AIG than by HPG.
(a) Scatter plot analysis using the methylation levels in overall CpG sites and in promoter CpG islands. Both the AIG (n = 11) and HPG (n = 9) samples showed aberrant DNA methylation, compared with the normal samples (n = 4) (left). Comparison between the AIG and HPG samples showed that AlG-induced methylation, but less than HPG (right).
(b) Volcano plot analysis using the methylation levels of genomic blocks in promoter CpG islands. The number of genomic blocks with large Ap values (> 0.05) and small P values (-1og10(P values) > 1.301) greatly increased both in the AIG and HPG samples compared with the normal samples. Moreover, the number of genomic blocks more methylated was remarkably higher in the HPG samples than in the AIG samples.
Figure S2. The weaker induction of aberrant DNA methylation by AIG than by HPG.
(a) Scatter plot analysis before correction of leucocyte fractions. The methylation levels in overall CpG sites and in promoter CpG islands were analyzed between the AIG (n=11), HPG (n=9), and normal samples (n=4). Both the AIG and HPG samples showed aberrant DNA methylation, compared with the normal samples (left). Comparison between the AIG and HPG samples showed that HPG further induced methylation upon AIG (right).
(b) Volcano plot analysis before correction of leucocyte fractions. The number of genomic blocks with large Δβ values (≥ 0.05) and small P values (–log10(P values) > 1.301) greatly increased both in the AIG and HPG samples compared with the normal samples. Moreover, the number of genomic blocks more methylated was remarkably higher in the HPG samples than in the AIG samples.
Driver genes were hypermethylation in AIG, but more so in HPG
4가지 Tumor-suppressor genes의 methylation levels을 비교하였습니다.
* CDKN2A (p16): methylated in gastric cancer (higher AIG < HPG)
* CDH1 (E-cadherin): inactivation is associated with diffuse-type gastric cancer (higher AIG < HPG)
* IGFBP7 & MIR34B: p53 pathway (higher AIG < HPG)
* DKK3 & SFRP1: negative regulator genes (higher AIG < HPG)
* MLH1: mismatch repair genes (little methylation)
--> AIG samples 에서 (HPG에서는 보다 더) cancer driver genes의 methylatio이 높다는 것이고, AIG는 상대적으로 HPG보다 낮은 carcinogenic potential을 갖는 다는 것입니다.
Figure 4. The DNA methylation levels of cancer driver genes among the AIG (n = l1), HPG (n = 9), and normal samples (n = 4) after correction of leucocyte fractions. (a) CDKN2A was highly methylated in the AIG samples, but more so in the HPG samples. (b) CDH1 was highly methylated in the AIG samples, but more so in the HPG samples. (c) IGFBP7 and MIR34B showed high methylation levels in the AIG samples, but more so in the HPG samples. (d) DKK3 and SFRP1 showed high methylation levels in the AIG samples, but more so in the HPG samples. (e) MLH1 showed little methylation in any types of the samples. †P < 0.01, n.s. not significant
Inflammation-related genes were more upregulated in HPG and in AIG
4가지 inflammation-related genes의 methylation levels을 비교하였습니다.
TNFα, IL1B, NOS2: DNA methylation induction에 관여하는 물질들입니다
* TNFα: secreted by all types of leucocytes (highly expressed in the AIG & HPG)
* IL1B, IL8: secreted by macrophage (induce neutrophil migration) (highly expressed in the HPG: likely reflecting the stronger monocyte/granulocyte infiltration)
* NOS2: (highly expressed in the AIG (AIG vs. HPG: no difference)
Figure 5. The expression levels of four infl ammation-related genes among the AIG (n = 8), IIPG (n = 8), and normal samples (n = 8). IL1B and IL8 showed remarkably high expression levels in the HPG samples. TNFα and NOS2 showed high expression levels in the AIG samples, but no significant difference between the AIG and HPG samples. *P < 0.05, †P < 0.01, n.s. not significant
이번 연구들을 살펴보면, AIG의 경우 위 점막에서 비 정상적인 DNA 메틸화가 일어난다는 것이고, HPG와 비교하여 독립적인 methylation profiles이 존재한다는 것입니다. 그리고, AIG-induced gastric cancer에 epigenetic metahnism이 작용한다는 것을 의미하기도 합니다. 하지만, AIG는 HPG 보다는 비 정상적인 DNA 메틸화가 덜 일어나는 것 같고, H. pylori infection이 없는 AIG에서는 동반된 AIG 보다 더 낮은 gastric cancer incidence를 보일 수 있다는 것입니다. 또한, DNA 메틸화를 유도하는 것에 있어서 T 또는 B cells 보다 macrophages와 neutrophils이 더 중요하다는 것이고, macrophage-induced inflammation이 DNA 메틸화 유도의 main mechanism 이라는 것을 알 수 있었습니다. 앞으로 AIG환자에서 HPG가 함께 동반 되었을 때 발생하는 DNA 메틸화 유도에 미치는 영향과 gastric cancer의 높은 incidence 연관성 등이 연구되어져야 할 것 같습니다.
Abstract
Background Autoimmune gastritis (AIG) is a chronic inflammatory condition in gastric mucosa and is associated with increased cancer risk, though not as high as that by Helicobacter pylori (H. pylori)-associated gastritis (HPG), Although aberrant DNA methylation is induced by HPG and the level correlates with the risk of gastric cancer, DNA methylation induction by AIG is unknown.
Methods Gastric mucosa samples from the corpus were collected from 12 people with AIG without H. pylori infection, 10 people with IIPG, and eight healthy volunteers. Genome-wide DNA methylation analysis was conducted using Infinium Methylation EPIC array. Gene expression was analyzed by quantitative RT-PCR.
Results The AIG samples had extensive aberrant DNA methylation but presented unique methylation profiles against the HPG samples after correction of leucocyte fractions. Comparison between the AIG and HPG samples showed that AIG induced methylation, but less than HPG, in overall CpG sites and also in promoter CpG islands. Promoter CpG islands of tumor-suppressor genes in the pathway of cell cycle, cell adhesion, p53, and WNT were highly methylated in the AIG samples, but more so in the HPG samples. The expression levels of ILIB and, Il3, secreted by macrophage, were significantly lower in the AIG samples than in the HPG samples, suggesting that a difference in inflammatory response affected the degree and pattems of aberrant DNA methylation.
Conclusions AIG induced aberrant DNA methylation in gastric mucosa. However, the degree of DNA methylation was less than that by HPG, which reflected carcinogenic risk.