Chemical Tags on a Heart Hormone Gene May Predict Stroke Risk and Recovery

A chemical code written onto the DNA of a single gene may help determine who suffers an ischemic stroke and how well they recover, according to a new study published in Epigenetics Communications. The research, led by Jialing Yao and colleagues across institutions in Suzhou, China, focused on the NPPA gene, which encodes atrial natriuretic peptide, or ANP, a hormone produced mainly by heart muscle cells. ANP is well known for its role in regulating blood pressure and fluid balance, but its connection to stroke has remained murky at the molecular level. By measuring two distinct epigenetic marks at the same locations in the NPPA promoter, the team uncovered associations that conventional methods would have blurred together, offering one of the clearest pictures yet of how this hormone gene might participate in stroke biology.

The study’s technical innovation lies in separating two DNA modifications that standard tools cannot distinguish. The most common epigenetic mark, 5-methylcytosine, or 5mC, typically silences gene expression when it appears in promoter regions. Under the action of TET enzymes, 5mC can be oxidized into 5-hydroxymethylcytosine, or 5hmC, a demethylation intermediate that often has the opposite effect and may activate genes. Traditional bisulfite sequencing, the workhorse of DNA methylation research, converts unmethylated cytosines to uracil but leaves both 5mC and 5hmC untouched, so the resulting measurement is actually a composite of the two marks. To untangle them, the researchers ran two parallel assays on the same DNA samples: targeted bisulfite sequencing to capture the combined signal, and APOBEC-coupled epigenetic sequencing, which uses an enzyme to label hydroxymethylated cytosines with glucose so they survive deamination while methylated and unmodified cytosines are converted. Subtracting the hydroxymethylation level from the composite level yielded what the team calls true methylation.

The study population was drawn from two established Chinese cohorts. Cases came from the China Antihypertensive Trial in Ischemic Stroke, a multicenter randomized trial of blood pressure management after stroke, from which 615 patients with imaging-confirmed ischemic stroke were selected. Controls, totaling 610 individuals, were frequency-matched by age and sex from the Prevention of Metabolic Syndrome and Multi-Metabolic Disorders Study, a community-based prospective cohort. The participants averaged 62 years of age, with roughly equal numbers of men and women in each group. As expected, stroke patients carried a heavier burden of metabolic risk factors, including hypertension, diabetes, obesity, hyperglycemia, and dyslipidemia, and the two groups also differed in education level. These characteristics were carefully accounted for in the statistical models, which adjusted for age, sex, education, smoking, drinking, body mass index, blood lipids, hypertension, and diabetes.

The target region spanned nine CpG sites in the NPPA promoter, located between 540 and 276 base pairs upstream of the transcription start site on chromosome 1. Genomic DNA extracted from peripheral blood mononuclear cells was processed through both sequencing pipelines, with stringent quality control: samples showing bisulfite conversion rates below 98 percent or average sequencing coverage under 20-fold were excluded. The nine CpG sites were then analyzed individually and collectively. For the gene-level analysis, the researchers employed the weighted truncated product method, which combines P values across multiple sites while weighting each by its regression coefficient, a strategy designed to detect joint effects that individual site analyses might miss when each contributes only a small share of the overall signal.

The methylation results were striking. At eight of the nine CpG sites, true methylation levels were significantly lower in stroke patients than in controls, and this hypomethylation was negatively associated with the risk of having an ischemic stroke. These associations survived correction for multiple testing using the false discovery rate approach, with all q values below 0.05. When the nine sites were considered together, the gene-based association was significant, and the mean methylation level across the region showed a clear dose-response relationship: each unit increase in mean true methylation corresponded to an odds ratio of 0.94 for stroke, with a 95 percent confidence interval of 0.91 to 0.97. Subgroup analyses revealed that the association was more pronounced among older participants, women, non-smokers, non-drinkers, and those with lower educational attainment, findings the authors suggest could guide the clinical application of epigenetic markers.

Hydroxymethylation told a different and more nuanced story. Although median hydroxymethylation levels at the assayed sites tended to be higher in stroke patients, only one CpG site, designated CpG2 and located at chromosome 1 position 11,908,348, showed a statistically significant positive association with stroke after covariate adjustment and multiple testing correction, with an odds ratio of 1.39 and a 95 percent confidence interval of 1.15 to 1.68. The gene-based test for hydroxymethylation was also significant, indicating a joint effect across the region, but the mean hydroxymethylation level alone did not distinguish cases from controls. Notably, CpG2 was the only site where both methylation and hydroxymethylation were independently associated with stroke, making it a particularly intriguing candidate for future precision interventions if the relevant technologies mature.

Beyond stroke risk, the researchers asked whether these epigenetic marks could predict recovery. They followed 598 of the 615 stroke patients, after excluding 17 lost to follow-up, and assessed functional outcome at three months using the modified Rankin Scale, an ordinal measure ranging from no symptoms to death. In ordered logistic regression models that additionally adjusted for stroke subtype, time from onset to hospitalization, stroke severity on the National Institutes of Health Stroke Scale, and treatment assignment, true methylation levels at all eight significant CpG sites were associated with a lower risk of scoring higher on the scale, meaning better neurological recovery. A complementary binary analysis defining poor outcome as a modified Rankin Scale score of 3 or higher produced consistent results. Hydroxymethylation, by contrast, showed no significant association with functional outcome at any site, and the gene-based test for hydroxymethylation and outcome was not significant.

The findings fit into a broader scientific context. Circulating ANP levels have previously been linked to ischemic stroke and to 90-day mortality and functional outcomes after stroke onset, and genetic variants in NPPA have been associated with both ANP levels and stroke susceptibility in studies of Chinese and Korean populations. Because promoter methylation generally suppresses gene expression, the observed hypomethylation in stroke patients is consistent with reports that stroke patients carry higher serum ANP levels, though the authors caution that a direct causal link between NPPA promoter methylation and ANP expression has not yet been verified; in their own earlier work, methylation at these same sites did not significantly predict serum proANP levels. Animal studies add another layer: mice experiments have shown that 5hmC can protect the brain from ischemic injury, and knockdown of the TET enzymes that generate 5hmC worsened brain degeneration in models of focal ischemia.

The study’s implications reach toward prevention and treatment. Because epigenetic modifications are chemically reversible, unlike fixed genetic variants, the CpG sites identified here could serve not only as biomarkers for risk stratification and prognosis but also as potential therapeutic targets. This is particularly relevant given the troubled history of ANP-targeted drugs such as sacubitril/valsartan and recombinant carperitide, whose clinical use has been limited by hypotension, angioedema, and fatalities in hospital settings. A deeper understanding of how the NPPA gene is regulated at the epigenetic level might open alternative routes to modulating the ANP pathway more safely.

The authors are careful to note the limitations of their work. As an observational study, it cannot determine whether altered methylation and hydroxymethylation in the NPPA promoter are causes, consequences, or mere companions of ischemic stroke. Residual confounding cannot be fully excluded despite extensive adjustment, and the findings, derived from Chinese adults, may not generalize to other ethnic or age groups. The absence of circulating ANP measurements in the current dataset leaves the mechanistic bridge between promoter marks and hormone levels unconfirmed. Nevertheless, by simultaneously quantifying true methylation and hydroxymethylation at identical CpG sites in a large case-control sample with prospective follow-up, the study demonstrates that the two marks carry distinct information about stroke risk and recovery, and it makes a compelling case that separating them should become standard practice in epigenetic epidemiology of complex disease.

Subject of Research: Associations of NPPA promoter DNA methylation and hydroxymethylation with ischemic stroke risk and functional outcome

Article Title: Associations of NPPA promoter true methylation and hydroxymethylation with ischemic stroke and its functional outcome

Article References: Yao, J., Song, L., Jiang, J., Zhang, J., Chen, L., Fan, W., Lu, Y., Zhang, X., Jing, J., Jin, Y., Zhang, M., Hao, Y., & Peng, H. (2024). Associations of NPPA promoter true methylation and hydroxymethylation with ischemic stroke and its functional outcome. Epigenetics Communications, 4(1), Article 6. https://doi.org/10.1186/s43682-024-00029-5

Image Credits: AI Generated

DOI: 10.1186/s43682-024-00029-5

Keywords: ischemic stroke, NPPA, atrial natriuretic peptide, DNA methylation, DNA hydroxymethylation, epigenetics, bisulfite sequencing, APOBEC-coupled epigenetic sequencing, modified Rankin Scale, functional outcome, biomarkers, case-control study