Limits...
System-Wide Associations between DNA-Methylation, Gene Expression, and Humoral Immune Response to Influenza Vaccination.

Zimmermann MT, Oberg AL, Grill DE, Ovsyannikova IG, Haralambieva IH, Kennedy RB, Poland GA - PLoS ONE (2016)

Bottom Line: The genes and molecular functions implicated by each analysis were compared, highlighting different aspects of the biologic mechanisms of immune response affected by differential methylation.Both cis-acting (within the gene or promoter) and trans-acting (enhancers and transcription factor binding sites) sites show significant associations with measures of humoral immunity.Specifically, we identified a group of CpGs that, when coordinately hypo-methylated, are associated with lower humoral immune response, and methylated with higher response.

View Article: PubMed Central - PubMed

Affiliation: Department of Health Science Research, Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota, United States of America.

ABSTRACT
Failure to achieve a protected state after influenza vaccination is poorly understood but occurs commonly among aged populations experiencing greater immunosenescence. In order to better understand immune response in the elderly, we studied epigenetic and transcriptomic profiles and humoral immune response outcomes in 50-74 year old healthy participants. Associations between DNA methylation and gene expression reveal a system-wide regulation of immune-relevant functions, likely playing a role in regulating a participant's propensity to respond to vaccination. Our findings show that sites of methylation regulation associated with humoral response to vaccination impact known cellular differentiation signaling and antigen presentation pathways. We performed our analysis using per-site and regionally average methylation levels, in addition to continuous or dichotomized outcome measures. The genes and molecular functions implicated by each analysis were compared, highlighting different aspects of the biologic mechanisms of immune response affected by differential methylation. Both cis-acting (within the gene or promoter) and trans-acting (enhancers and transcription factor binding sites) sites show significant associations with measures of humoral immunity. Specifically, we identified a group of CpGs that, when coordinately hypo-methylated, are associated with lower humoral immune response, and methylated with higher response. Additionally, CpGs that individually predict humoral immune responses are enriched for polycomb-group and FOXP2 transcription factor binding sites. The most robust associations implicate differential methylation affecting gene expression levels of genes with known roles in immunity (e.g. HLA-B and HLA-DQB2) and immunosenescence. We believe our data and analysis strategy highlight new and interesting epigenetic trends affecting humoral response to vaccination against influenza; one of the most common and impactful viral pathogens.

Show MeSH

Related in: MedlinePlus

Interrelationships between genes associating with each humoral immune outcome.A) A Venn diagram summarizes the number of genes associated with each outcome across time points. A “core” set of 54 genes are identified using at least two outcomes. We divide genes only associated with one outcome into two groups; those that share network links with genes in the core (inner number), and those that do not (outer number). B) The fraction of genes with direct links to each other (gene-gene connectivity), within and between each outcome-specific set, is compared to observations from random gene sets of the same size. The six comparisons between HAI, B-cell ELISPOT, and gene expression are shown in the same relative position as the Venn diagram with a colored vertical bar indicating the gene-gene connectivity observed in our study and the distribution of connectivity from randomly generated genesets in gray. C) A visualization of the network is shown using the subset of links with greatest confidence and laid out similarly to the other panels. The extent of gene-gene connectivity is apparent from the number of genes (represented by colored circles) with known direct interactions (gray lines crossing between groups).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4816338&req=5

pone.0152034.g004: Interrelationships between genes associating with each humoral immune outcome.A) A Venn diagram summarizes the number of genes associated with each outcome across time points. A “core” set of 54 genes are identified using at least two outcomes. We divide genes only associated with one outcome into two groups; those that share network links with genes in the core (inner number), and those that do not (outer number). B) The fraction of genes with direct links to each other (gene-gene connectivity), within and between each outcome-specific set, is compared to observations from random gene sets of the same size. The six comparisons between HAI, B-cell ELISPOT, and gene expression are shown in the same relative position as the Venn diagram with a colored vertical bar indicating the gene-gene connectivity observed in our study and the distribution of connectivity from randomly generated genesets in gray. C) A visualization of the network is shown using the subset of links with greatest confidence and laid out similarly to the other panels. The extent of gene-gene connectivity is apparent from the number of genes (represented by colored circles) with known direct interactions (gray lines crossing between groups).

Mentions: We have identified associations between DNA methylation, gene expression, and humoral immune outcomes. Next, we compared and contrasted these analyses (Fig 4). In order to integrate across cis-acting associations using both individual CpG and promoter-averaged analyses, all results are summarized to the gene level. First, we identified cis-acting methylation with associations in at least two analyses and list them in Table 3. Two genes show association in all three analyses: ADARB2, an inhibitor of adenosine deaminase activity (RNA editing), and SPEG, a kinase with known function in myocyte development. Second, subsets of CpGs that predict late response also show concordance with low baseline activation (S6 Fig). Participants with low methylation within a particular cluster of CpGs show lower baseline log2 B-cell ELISPOT values (2.7±1.8) than participants with high methylation of levels of the same CpGs (3.4±1.4). Third, known interactions between these genes from network biology resources are gathered and the level of interaction between them tested for statistical significance. Many of the genes identified in one analysis have direct interactions (protein-protein interaction or one link apart in known pathways) with genes identified in the other analyses. For example, there are 640 genes identified in the analyses focused on B-cell ELISPOT or gene expression. Similar ratios are seen starting from B-cell ELISPOT associated genes (40%) and expression- level associated genes (44%). Beyond this highly connected core, we tested the level of interconnectivity (fraction of genes sharing links) between the genes showing association with only one outcome. Comparing our observed levels of gene-gene interconnectivity to distributions of randomly selected genes of the same sizes, the observed levels are high (p < 0.005) for all three pairwise comparisons (Fig 4).


System-Wide Associations between DNA-Methylation, Gene Expression, and Humoral Immune Response to Influenza Vaccination.

Zimmermann MT, Oberg AL, Grill DE, Ovsyannikova IG, Haralambieva IH, Kennedy RB, Poland GA - PLoS ONE (2016)

Interrelationships between genes associating with each humoral immune outcome.A) A Venn diagram summarizes the number of genes associated with each outcome across time points. A “core” set of 54 genes are identified using at least two outcomes. We divide genes only associated with one outcome into two groups; those that share network links with genes in the core (inner number), and those that do not (outer number). B) The fraction of genes with direct links to each other (gene-gene connectivity), within and between each outcome-specific set, is compared to observations from random gene sets of the same size. The six comparisons between HAI, B-cell ELISPOT, and gene expression are shown in the same relative position as the Venn diagram with a colored vertical bar indicating the gene-gene connectivity observed in our study and the distribution of connectivity from randomly generated genesets in gray. C) A visualization of the network is shown using the subset of links with greatest confidence and laid out similarly to the other panels. The extent of gene-gene connectivity is apparent from the number of genes (represented by colored circles) with known direct interactions (gray lines crossing between groups).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4816338&req=5

pone.0152034.g004: Interrelationships between genes associating with each humoral immune outcome.A) A Venn diagram summarizes the number of genes associated with each outcome across time points. A “core” set of 54 genes are identified using at least two outcomes. We divide genes only associated with one outcome into two groups; those that share network links with genes in the core (inner number), and those that do not (outer number). B) The fraction of genes with direct links to each other (gene-gene connectivity), within and between each outcome-specific set, is compared to observations from random gene sets of the same size. The six comparisons between HAI, B-cell ELISPOT, and gene expression are shown in the same relative position as the Venn diagram with a colored vertical bar indicating the gene-gene connectivity observed in our study and the distribution of connectivity from randomly generated genesets in gray. C) A visualization of the network is shown using the subset of links with greatest confidence and laid out similarly to the other panels. The extent of gene-gene connectivity is apparent from the number of genes (represented by colored circles) with known direct interactions (gray lines crossing between groups).
Mentions: We have identified associations between DNA methylation, gene expression, and humoral immune outcomes. Next, we compared and contrasted these analyses (Fig 4). In order to integrate across cis-acting associations using both individual CpG and promoter-averaged analyses, all results are summarized to the gene level. First, we identified cis-acting methylation with associations in at least two analyses and list them in Table 3. Two genes show association in all three analyses: ADARB2, an inhibitor of adenosine deaminase activity (RNA editing), and SPEG, a kinase with known function in myocyte development. Second, subsets of CpGs that predict late response also show concordance with low baseline activation (S6 Fig). Participants with low methylation within a particular cluster of CpGs show lower baseline log2 B-cell ELISPOT values (2.7±1.8) than participants with high methylation of levels of the same CpGs (3.4±1.4). Third, known interactions between these genes from network biology resources are gathered and the level of interaction between them tested for statistical significance. Many of the genes identified in one analysis have direct interactions (protein-protein interaction or one link apart in known pathways) with genes identified in the other analyses. For example, there are 640 genes identified in the analyses focused on B-cell ELISPOT or gene expression. Similar ratios are seen starting from B-cell ELISPOT associated genes (40%) and expression- level associated genes (44%). Beyond this highly connected core, we tested the level of interconnectivity (fraction of genes sharing links) between the genes showing association with only one outcome. Comparing our observed levels of gene-gene interconnectivity to distributions of randomly selected genes of the same sizes, the observed levels are high (p < 0.005) for all three pairwise comparisons (Fig 4).

Bottom Line: The genes and molecular functions implicated by each analysis were compared, highlighting different aspects of the biologic mechanisms of immune response affected by differential methylation.Both cis-acting (within the gene or promoter) and trans-acting (enhancers and transcription factor binding sites) sites show significant associations with measures of humoral immunity.Specifically, we identified a group of CpGs that, when coordinately hypo-methylated, are associated with lower humoral immune response, and methylated with higher response.

View Article: PubMed Central - PubMed

Affiliation: Department of Health Science Research, Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota, United States of America.

ABSTRACT
Failure to achieve a protected state after influenza vaccination is poorly understood but occurs commonly among aged populations experiencing greater immunosenescence. In order to better understand immune response in the elderly, we studied epigenetic and transcriptomic profiles and humoral immune response outcomes in 50-74 year old healthy participants. Associations between DNA methylation and gene expression reveal a system-wide regulation of immune-relevant functions, likely playing a role in regulating a participant's propensity to respond to vaccination. Our findings show that sites of methylation regulation associated with humoral response to vaccination impact known cellular differentiation signaling and antigen presentation pathways. We performed our analysis using per-site and regionally average methylation levels, in addition to continuous or dichotomized outcome measures. The genes and molecular functions implicated by each analysis were compared, highlighting different aspects of the biologic mechanisms of immune response affected by differential methylation. Both cis-acting (within the gene or promoter) and trans-acting (enhancers and transcription factor binding sites) sites show significant associations with measures of humoral immunity. Specifically, we identified a group of CpGs that, when coordinately hypo-methylated, are associated with lower humoral immune response, and methylated with higher response. Additionally, CpGs that individually predict humoral immune responses are enriched for polycomb-group and FOXP2 transcription factor binding sites. The most robust associations implicate differential methylation affecting gene expression levels of genes with known roles in immunity (e.g. HLA-B and HLA-DQB2) and immunosenescence. We believe our data and analysis strategy highlight new and interesting epigenetic trends affecting humoral response to vaccination against influenza; one of the most common and impactful viral pathogens.

Show MeSH
Related in: MedlinePlus