Gregory E. Miller’s research while affiliated with Northwestern University and other places

Problem We sought to investigate whether maternal inflammatory cytokines during pregnancy are associated with histologic inflammatory or vascular lesions in the placenta and/or correlated with gene expression patterns in the placenta. Method of Study We leveraged data from a large randomized controlled trial (RCT) at a single site. Maternal serum was collected in the second and third trimesters, and a composite inflammatory score was created using five measured biomarkers (CRP, IL‐6, IL‐1ra, IL‐10, and TNF‐α). Placentas were collected at delivery for histological analysis and four major patterns of placental injury were characterized. Fresh small chorionic villous biopsies were collected for placental genome‐wide mRNA profiling. Transcripts showing >2‐fold differential expression over the 4‐SD range of circulating inflammatory biomarkers were reported, adjusting for potential confounders. Results The primary analysis included 601 participants. A one standard deviation increase in the third‐trimester inflammatory composite was associated with increased odds of chronic inflammation in the placenta (OR: 1.23, 95% CI 1.01, 1.51;). This was driven primarily by elevations in IL‐10 (OR: 1.37; 99% CI: 1.06, 1.77). Higher maternal IL‐10 in circulation was associated with bioinformatic indications of reduced pro‐inflammatory gene regulation pathways in the placenta (AP1 decreased 25%, p = 0.003; NF‐kB decreased 53%, p = 0.003) and indications of increased STAT family signaling pathways which mediate signaling through the IL‐10 receptor (increased 73%, p = 0.002). Conclusions Our results indicate that elevated maternal circulating IL‐10 during pregnancy is associated with chronic inflammatory lesions in the placenta at delivery. Additionally, higher levels of circulating IL‐10 are associated with upregulated STAT signaling pathways in placental tissues.

Dried blood spot sampling offers a scalable strategy to close diagnostic gaps and improve global surveillance for cardiovascular‐kidney‐metabolic syndrome. However, assay performance and the extent of validity vary widely between biomarkers used in cardiovascular‐kidney‐metabolic health assessment under different settings and have not been well described. To fill this gap, we conducted a systematic search of the literature and a narrative synthesis through April 2024 and included reports with laboratory or field validation measuring biomarkers that can be used in cardiovascular‐kidney‐metabolic health assessment. We categorized assays into categories based on laboratory validation: excellent performance (r>0.95 with gold standard methods and coefficients of variation <5%), very good performance (r>0.90 and coefficients of variation <10%), reasonable performance (r>0.80 and coefficients of variation <15%), and poor performance (r<0.80 or coefficients of variation >15%). The extent of validation was determined by the total number of field validation studies with strong agreement. Hemoglobin A1c has strong laboratory and field validation and should be considered for expansion into clinical testing in low‐resource settings. Traditional lipid biomarkers showed poor performance in field validation studies, but apoB (apolipoprotein B), creatinine, cystatin C, and NT‐proBNP (N‐terminal prohormone of brain natriuretic peptide) showed promising initial laboratory validation results and deserve greater attention in field validation studies. High‐sensitivity C‐reactive protein has strong laboratory and field validation but has limited clinical utility. Dried blood spot assays have been developed for biomarkers that offer mechanistic insights including inflammatory and vascular injury markers, fatty acids, malondialdehyde, asymmetric dimethylarginine, trimethylamine N‐oxide, carnitines, and omics.

Introduction: Poor pre-pregnancy cardiovascular health is associated with adverse pregnancy outcomes (APOs). We sought to investigate whether estimated 30-year CVD risk from the PREVENT equation similarly predicts APOs. Hypothesis: We hypothesized that higher 30-year CVD risk mid-pregnancy would be associated with greater risk of APOs, including preterm delivery and preeclampsia. Methods: Participants are from the Stress, Pregnancy, and Health (SPAH) study, conducted in Evanston, IL between 2018-2023. During the 2 nd trimester study visit, blood pressure was measured and participants provided a blood sample. Nuclear magnetic resonance was used to measure serum cholesterol and creatinine (used to estimate eGFR). Medication use, pre-gestational diabetes diagnosis, and APOs, including preterm delivery (<37 weeks gestation), early preterm delivery (<34 weeks) and preeclampsia, were abstracted from medical records. The PREVENT equation was used to estimate 30-year CVD risk based on 2 nd trimester measures. Logistic models were used to test associations between CVD risk in pregnancy and APOs, adjusted for race/ethnicity, as a marker of structural factors, socioeconomic position, based on household educational attainment and occupational prestige, and gestational age at the study visit. Results: SPAH included 605 participants, 524 of whom had complete data to estimate 30-year CVD risk from the 2 nd trimester study visit (mean gestational age at visit: 23.7 weeks, SD: 1.7). The mean age of participants was 33.3 years (SD: 5.5) and 18% identified as Black, 21% as Hispanic, 49% as White, and 12% as another group. The mean 30-year CVD risk during pregnancy was 7.6% (SD: 6.6), 11% delivered preterm, and 10% developed preeclampsia. In the adjusted model, a 1 SD increase in CVD risk was associated with 41% greater odds of preterm delivery (95% CI: 1.14, 1.74), 67% greater odds of early preterm delivery (95% CI: 1.27, 2.19), and 71% greater odds of developing preeclampsia (95% CI: 1.36, 2.14; Figure 1). Conclusions: Results suggest that a higher PREVENT 30-year estimated CVD risk in mid-pregnancy is associated with greater odds of delivering preterm and developing preeclampsia. These APOs are associated with long-term offspring CVD risk, also highlighting the intergenerational implications of PREVENT CVD risk estimates.

Introduction: Many adverse pregnancy outcomes (APOs) associated with long-term CVD risk, such as preeclampsia, are mediated by placental function. Notably, the vascular and inflammatory placental pathology underlying APOs are also found in 20-35% of healthy pregnancies. Thus, we aimed to investigate the link between 30-year CVD risk and placental pathology. Hypothesis: We hypothesized that higher 30-year CVD risk, as estimated based on the PREVENT equation calculated in mid-pregnancy, would be associated with greater likelihood of placental pathology, including chronic inflammation (CI), maternal vascular malperfusion (MVM), and fetal vascular malperfusion (FVM). Methods: Data are from the Stress, Pregnancy, and Health (SPAH) study, a prospective pregnancy cohort conducted from 2018-2023 in Evanston, IL. The PREVENT 30-year CVD risk estimate was calculated based on the 2 nd trimester study visit. At the study visit, blood pressure was assessed, and a blood sample was collected and used to measure cholesterol and estimate eGFR from creatinine. Participants self-reported smoking status. Age, pre-gestational diabetes diagnosis, and use of anti-hypertensive or lipid-lowering medications were abstracted from medical records. Following delivery, placentas were collected and reviewed by a perinatal pathologist and patterns of placental injury (CI, MVM, FVM) were identified based on Amsterdam consensus criteria. Logistic models were used to test associations between CVD risk in pregnancy and placental pathology, adjusted for race/ethnicity, socioeconomic position, and gestational age at the study visit. Results: The SPAH study included 605 participants, 505 of whom had placental pathology exams and complete data to estimate 30-year CVD risk during pregnancy. The mean gestational age at the study visit was 23.7 weeks (SD: 1.7) and the mean age of study participants was 33.3 years (SD: 5.6). The mean 30-year CVD risk was 7.6% (SD: 6.7) and 54% had CI, 29% had MVM, and 34% had FVM. In the adjusted model, a 1 SD increase in CVD risk was associated with 24% greater odds of having CI (95% CI: 1.01, 1.52) and 25% greater odds of having MVM (95% CI: 1.04, 1.50; Figure 1) at delivery. CVD risk was not associated with FVM. Conclusions: Results demonstrate that PREVENT 30-year CVD risk estimated in the second trimester is associated with CI and MVM in the placenta at delivery, which may provide insight into the mechanisms linking CVD risk and APOs.

BACKGROUND Evidence suggests that the intrauterine environment shapes offspring cardiovascular disease risk. Although placental dysfunction may be an important pathophysiologic pathway, numerous parental and pregnancy characteristics that influence offspring blood pressure are strong confounders of the mechanistic role of the placenta in observational analyses of singletons. Therefore, we leverage twin- and sibling-based comparison designs to determine whether placental pathology is associated with offspring blood pressure at age 7 while mitigating major sources of confounding. METHODS Data are from pregnant participants and their offspring in the Collaborative Perinatal Project, a longitudinal pregnancy cohort conducted from 1959 to 1965 in the United States. After delivery, placentas were systematically examined for lesions indicative of maternal vascular malperfusion (MVM) and acute inflammation. Blood pressure was assessed at a follow-up research visit when the offspring were 7 years old. Linear fixed-effects models were used to estimate associations between within-twin or sibling discordance in placental pathology and differences in blood pressure at age 7. RESULTS Overall, 193 twin pairs were eligible for inclusion, and 23.3% had placentas discordant for MVM. In a fixed-effect analysis, a twin with high-grade MVM had a higher systolic blood pressure Z score by 0.56 SDs than their co-twin without MVM (95% CI, 0.06–1.05) or a 5.7-mm Hg difference (95% CI, 0.6–10.8). Associations were consistent in a sensitivity analysis restricted to dichorionic twins and in a secondary analysis of 759 MVM-discordant sibling pairs. Acute placental inflammation was not associated with blood pressure at age 7. CONCLUSIONS MVM in the placenta is associated with higher offspring blood pressure in mid-childhood, independent of parental and pregnancy characteristics that twins have in common. The findings support the role of the placenta and the intrauterine environment in the developmental origins of cardiovascular health.

Background: Several area-level indices exist that integrate a variety of social drivers of health. However, whether certain indices better capture neighborhood-level variability in cardiovascular-kidney-metabolic (CKM) outcomes is not well known. Aims: We sought to examine the associations of census tract-level indices of social disadvantage with CKM outcomes. Methods: Exposures for the analysis included seven indices of social disadvantage (ADI, COI, EJI, NDI, SDI, SREI, SVI) based on census tract-level data from the 2010-2019 American Community Survey (ACS) and census tract-level median household income from the 2015-2019 ACS data. Outcomes included the census tract-level prevalence of CKM outcomes (obesity, hypertension, diabetes, hyperlipidemia, chronic kidney disease, coronary heart disease [CHD], and stroke) from the 2019 Behavioral Risk Factor Surveillance System. All indices and median income were standardized, and we used linear regression to examine the associations between each census tract-level social measure and CKM outcomes, adjusted for population size and median age. Next, we calculated the △r ² in adjusted models for the association of median income with each CKM outcome when each index was added. Results: Among 65,476 US census tracts, median (IQR) population size was 4068 (2969, 5374), age was 39 years (35, 44), and household income was $58,870 (44276, 79802). Median income as well as each index was significantly associated with census tract-level prevalence of each CKM outcome assessed (p < 0.05). The r ² (Panel A) and △r ² values (Panel B) varied for each of the social measures and CKM outcomes, with higher r ² values when each index was individually added to median income compared with income alone. For example, for CHD, the r ² ranged between 0.39 to 0.67 and △r ² ranged from 0.01 to 0.10. Conclusions: Neighborhood-level measures of social disadvantage are differentially associated with CKM outcomes, with a large proportion of the variability explained by median income. Identifying the advantages and disadvantages of each index and comparison with median income can inform the prioritization of measures for specific outcomes.

DNA methylation (DNAm) is a key epigenetic mark that modulates regulatory elements and gene expression, playing a crucial role in mammalian development and physiological function. Despite extensive characterization of DNAm profiles across species, little is known about its evolutionary conservation. Here, we conducted a comparative epigenome-wide analysis of great apes to identify and characterize sequence- and methylation-conserved CpGs (MCCs). Using 202 DNAm arrays, alongside 6 matched genotype and 13 matched transcriptomic datasets, we identified 11,500 MCCs for which methylation was evolutionarily related to sequences of CpGs and methylation quantitative trait loci. MCCs were the most stable across healthy human tissues and exhibited weaker genetic associations than other CpGs. Moreover, MCCs showed minimal associations with demographic, environmental factors, and noncancer diseases, yet demonstrated stronger associations with certain cancers than other CpGs, particularly gastrointestinal cancers. Functional enrichment analysis revealed that genes associated with MCC methylation in cancer were enriched for cancer driver genes and canonical cancer pathways, highlighting a significant regulatory role for MCCs in tumorigenesis. Collectively, our findings reveal the extent of DNAm conservation in great ape evolution, its association with genetic conservation, and its relevance to human diseases. These integrative analyses offer evolutionary insights into epigenetic variation and its functional implications in human populations.