Skip to main content
NCBI home page
British Journal of Pharmacology logoLink to British Journal of Pharmacology
. 2018 Jun 21;176(12):2015–2027. doi: 10.1111/bph.14364

Human monocyte transcriptional profiling identifies IL‐18 receptor accessory protein and lactoferrin as novel immune targets in hypertension

Matthew R Alexander 1,†, Allison E Norlander 2,†, Fernando Elijovich 3, Ravi V Atreya 4, Amadou Gaye 5, Juan S Gnecco 6, Cheryl L Laffer 3, Cristi L Galindo 1, Meena S Madhur 1,2,3,✉
PMCID: PMC6534784  PMID: 29774543

Abstract

Background and Purpose

Monocytes play a critical role in hypertension. The purpose of our study was to use an unbiased approach to determine whether hypertensive individuals on conventional therapy exhibit an altered monocyte gene expression profile and to perform validation studies of selected genes to identify novel therapeutic targets for hypertension.

Experimental Approach

Next generation RNA sequencing identified differentially expressed genes in a small discovery cohort of normotensive and hypertensive individuals. Several of these genes were further investigated for association with hypertension in multiple validation cohorts using qRT‐PCR, regression analysis, phenome‐wide association study and case–control analysis of a missense polymorphism.

Key Results

We identified 60 genes that were significantly differentially expressed in hypertensive monocytes, many of which are related to IL‐1β. Uni‐ and multivariate regression analyses of the expression of these genes with mean arterial pressure (MAP) revealed four genes that significantly correlated with MAP in normotensive and/or hypertensive individuals. Of these, lactoferrin (LTF), peptidoglycan recognition protein 1 and IL‐18 receptor accessory protein (IL18RAP) remained significantly elevated in peripheral monocytes of hypertensive individuals in a separate validation cohort. Interestingly, IL18RAP expression associated with MAP in a cohort of African Americans. Furthermore, homozygosity for a missense single nucleotide polymorphism in LTF that decreases antimicrobial function and increases protein levels (rs1126478) was over‐represented in patients with hypertension relative to controls (odds ratio 1.16).

Conclusions and Implications

These data demonstrate that monocytes exhibit enhanced pro‐inflammatory gene expression in hypertensive individuals and identify IL18RAP and LTF as potential novel mediators of human hypertension.

Linked Articles

This article is part of a themed section on Immune Targets in Hypertension. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v176.12/issuetoc

Abbreviations

ARG1

arginase‐1

BMI

body mass index

CAD

coronary artery disease

DAMP

damage associated molecular pattern

DM2

diabetes mellitus type 2

eGFR

estimated GFR

GZMH

granzyme H

HER

electronic health record

IL18RAP

IL‐18 receptor accessory protein

IPA

Ingenuity Pathway Analysis

LTF

lactotransferrin

MAP

mean arterial pressure

MH‐GRID

Minority Health Genomics and Translational Research Bio‐Repository Database

OR

odds ratio

PCA

principal components analysis

PGLYRP1

peptidoglycan recognition protein 1

PheWAS

phenome‐wide association study

RPKM

reads per kilobase per million mapped reads

SNP

single nucleotide polymorphism

VSIG4

V‐set and immunoglobulin domain containing 4

Introduction

Hypertension is a leading risk factor for death and disability worldwide and a primary contributor to myocardial infarction, stroke, heart failure and chronic kidney disease. Emerging evidence from our group and others indicates that hypertension is an inflammatory disease in which antigen presenting cells, particularly monocyte‐derived dendritic cells, present immunogenic peptides to T lymphocytes which then leads to T cell activation, infiltration of target organs and production of pro‐inflammatory cytokines that cause sodium and water retention, vascular dysfunction and renal injury (Wenzel et al., 2011; Kirabo et al., 2014; McMaster et al., 2015; Saleh et al., 2015; Norlander et al., 2016).

Circulating monocytes are pluripotent cells capable of differentiating into dendritic cells, macrophages and microglial cells based on environmental cues. Animal models demonstrate a critical role for monocytes/macrophages in the development of hypertension (Wenzel et al., 2011; Justin Rucker & Crowley, 2017). In humans, monocytes from patients with hypertension secrete more pro‐inflammatory cytokines including IL‐1β and TNF‐α (Dorffel et al., 1999; Wirtz et al., 2004), demonstrate enhanced migration (Zhao et al., 2012), exhibit increased adherence to endothelial cells (Dorffel et al., 2001) and produce more ROS (Dorffel et al., 2001). However, these earlier studies used monocytes from patients who were either untreated or had temporarily stopped treatment. Interestingly, Kirabo et al. (2014) recently demonstrated that monocytes from treated hypertensive patients exhibit higher levels of isoketals, which are oxidatively modified lipids that can form adducts with proteins post‐translationally, potentially leading to the formation of neo‐antigens. We hypothesized that monocytes from conventionally treated hypertensive individuals also exhibit an altered transcriptomic or gene expression profile and that some of these genes may be novel therapeutic targets for hypertension.

To address this hypothesis, we isolated peripheral blood monocytes from normotensive and treated hypertensive individuals and performed unbiased next generation RNA sequencing of the entire transcriptome. We identified 60 genes that were significantly differentially expressed in hypertensive monocytes, many of which are involved in IL‐1 and IL‐18 signalling. Four of these genes correlated with mean arterial pressure (MAP) in multivariate analysis, and three of these four remained significantly differentially expressed in monocytes from an independent validation cohort of six normotensive and nine hypertensive subjects. Of these, IL‐18 receptor accessory protein (IL18RAP) was associated with MAP in a separate cohort of 76 hypertensive African Americans, and homozygosity for the minor allele of a missense single nucleotide polymorphism (SNP) in lactotransferrin (LTF) was increased in frequency in hypertensive individuals compared to normotensive controls in a European cohort of >15 000 individuals, suggesting that IL18RAP and LTF may be novel therapeutic targets for human hypertension.

Methods

Patients

The protocol was approved by the Vanderbilt Institutional Review Board and conforms to recognized standards of the US Federal Policy for the Protection of Human Subjects. In an initial exploratory study, whole blood (40 mL) was obtained from five normotensive and seven hypertensive subjects recruited at Vanderbilt. One normotensive subject was sampled on two separate days to obtain a technical replicate that could be used to test the internal consistency of the RNAseq assay. An additional six normotensive and nine hypertensive subjects were recruited for the validation cohort. Hypertension was defined as a systolic blood pressure greater than 140 mmHg, a diastolic blood pressure greater than 90 mmHg or treatment with antihypertensive agents, regardless of blood pressure. Exclusion criteria were confirmed or suspected secondary causes of hypertension, diabetes mellitus type 1 or 2, concomitant illness requiring corticosteroids or immunosuppressants, recent (within 3 months) vaccination against any infectious agent, active ongoing malignancy, severe psychiatric disorders or the presence of HIV/AIDS. Informed consent was obtained prior to the blood draw.

Monocyte isolation and RNA extraction

Whole blood was mixed 1:1 with cold PBS + EDTA. The mixture was layered on top of Ficoll (GE Life Sciences, Pittsburgh, PA, USA) and centrifuged for 30 min. The ‘buffy coat’ layer was isolated and washed three times with cold PBS + EDTA. The isolated PBMCs were stained using Monocyte Isolation Kit II, human (Miltenyi, Auburn, CA, USA), and the monocytes were negatively selected using an AutoMACs cell separator. Monocytes were collected after centrifugation and RNA was extracted using an RNeasy kit (Qiagen, Germantown, MD, USA), which included a DNaseI treatment step. The purity and quantity of RNA samples were measured using a Denovix DS‐11 Spectrophotometer (Denovix Inc., Wilmington, DE, USA) or a NanoDrop ND‐1000 Spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA).

RNA sequencing

The cDNA library construction and RNA sequencing was performed by VANTAGE (Vanderbilt Technologies for Advanced Genomics). The cDNA library preparation was accomplished using the Illumina TruSeq stranded mRNA Sample Preparation Kit and Rev. D of the protocol. Samples were sequenced on the Illumina HiSeq 2500 using v3 SBS chemistry. Libraries were sequenced on a Single Read 50 bp run at 30 million passing filter reads/sample.

RNA sequencing analysis

A total of 13 samples were obtained from 12 patients. Based on post‐alignment quality assurance and control (QA/QC), all samples were of high fidelity, i.e. Phred quality score = 37.00 ± 0.31, indicating a high probability that the bases were accurately identified. The Phred score = −10*log10 (Probability of Error). Samples were separated into two groups: hypertensive (n = 7) and normotensive (n = 5). Two samples were collected for one of the five normotensive subjects (technical replicates), obtained on separate days, allowing us to perform two separate sets of analyses to increase the robustness of the data and reduce false positives due to RNASeq technical variation. For all 13 samples, the base call accuracy was estimated to be >99.9%. Libraries were sequenced on a Single Read 50 bp run at 30 million passing filter reads/sample. Total reads were 35 021 078 +/− 2.7 M, and 86% +/− 0.84% of reads were aligned to the transcriptome. Statistical analysis for RNAseq was performed using Partek Flow, with the following parameters: Alignment: STAR and TopHat2/Bowtie2; Quantification: Cufflinks and P/E (Partek's optimization of the expectation–maximization algorithm) (Li et al., 2010); Statistics: Cuffdiff, GSA, and ANOVA; Annotations: RefSeq and Aceview; P value adjustment for multiple hypothesis testing: Benjamini & Hochberg. Analysis 1 included all hypertensive and normotensive subjects (true biological replicates), with the normotensive replicated subject represented by the first sample. Analysis 2 was identical to Analysis 1, except the second sample from that subject was used. While we were interested in obtaining a global view of transcriptional differences between hypertensive and normotensive subjects, we especially wanted to identify transcripts with the highest likelihood of being true and reproducible findings. For this purpose, we employed 12 separate methodologies (six each for Analysis 1 and 2), using various combinations of alignment, quantification, statistical and annotation strategies.

Principal component analysis (PCA) and Ingenuity Pathway Analysis (IPA) including upstream regulator analysis

PCA (with correlation matrix) and hierarchical clustering (with Euclidian distance and Average linkage) were performed using Partek Genomics Suite 6.6. Functional analysis was performed using IPA, Winter Release 2017 (http://www.ingenuity.com). Upstream regulator analysis, also performed within IPA, calculates two statistical values for each regulator molecule within Ingenuity's knowledge base: (i) the overlap P value, which measures whether there is a statistically significant overlap between the gene list of interest and genes that are regulated by the upstream factor and (ii) the Z score, which infers the likely activation states of upstream regulators. Upstream regulators with positive and negative Z scores of 2 or greater are considered significantly activated or inhibited respectively.

Multivariate regression analyses

Stepwise multivariate regression analyses were carried out to identify genes, the expression of which [log10 of normalized reads per kilobase per million mapped reads (RPKM)] could be a determinant of blood pressure. Univariate correlation analysis was used for hypothesis generation, that is identification of candidate regressor genes. If the expression of a gene correlated with MAP within one group of subjects (either controls or hypertensive patients) or within both, the gene was considered a candidate regressor and included in the multivariate analysis. In contrast, if the expression of a gene only correlated with MAP in all subjects analysed together, it was not included as a hypothetical regressor. This is because these apparent univariate correlations may simply be driven by the two different levels of blood pressure and gene expression in the two groups of subjects. A P value <0.05 was considered significant.

Quantitative RT‐PCR

Monocyte RNA was converted to cDNA using a High Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, Foster City, CA, USA). The cDNA was then probed for the following genes [all probes purchased from Taqman (ThermoFisher Scientific, Waltham, MA, USA)]: arginase‐1 (Arg1), the chemokine CCL4, MMP8, V‐set and immunoglobulin domain containing 4 (VSIG4), peptidoglycan recognition protein 1 (PGLYRP1) and IL18RAP. Results were normalized to glyceraldehyde‐3‐phosphate dehydrogenase and subsequently normalized to one of the normotensive patients. The data are plotted as relative quantification values.

Gene expression correlation in African Americans

RNA sequencing was performed on whole blood mRNA of self‐reported African Americans from the Morehouse School of Medicine subset of the Minority Health Genomics and Translational Research Bio‐Repository Database (MH‐GRID) as described previously (Gaye et al., 2017). For this analysis, total RNA was extracted from whole blood using MagMAX for Stabilized Blood Tubes RNA Isolation Kit (Life Technologies, Carlsbad, CA, USA). TruSeq (Illumina, San Diego, CA, USA) was then used to convert total RNA into cDNA sequencing libraries. The cDNA libraries underwent quantitative RT‐PCR using the KAPA Library Quant Kit (KAPA Biosystems; Wilmington, MA, USA). Expression levels defined as read counts were determined using a Bowtie2‐based pipeline for alignment with read counts determined using RSEM. Expression was normalized using the weighted trimmed mean of M‐values method. Correlation between the normalized expression of LTF, PGLYRP1 and IL18RAP with MAP or estimated GFR (eGFR) was assessed by linear regression in 76 hypertensive patients. Four models were fitted: (i) without covariates, (ii) adjusting for age, (iii) adjusting for gender, (iv) adjusting for body mass index (BMI) and (v) adjusting for age, gender and BMI.

Phenome‐wide association study (PheWAS)

This analysis was performed on a cohort from BioVU, the Vanderbilt DNA databank that links DNA extracted from discarded blood samples to de‐identified electronic health records (EHRs). BioVU operated on an opt‐out basis until January 2015 and on an opt‐in basis since then (Roden et al., 2008). The phenotypic data in BioVU are de‐identified, and the study was approved by the Vanderbilt Institutional Review Board as ‘non‐human subjects’ research. PheWAS results were generated on a cohort of 29 713 self‐identified Caucasian/non‐Hispanic patients of European ancestry with available HumanCoreExome BeadChip (Illumina; San Diego, CA, USA) genotyping. The mean age of the individuals was 55.4 years, and 53% were female. Ancestry was determined using STRUCTURE (Porras‐Hurtado et al., 2013). Case/control status was determined by aggregating International Classification of Diseases 9 (ICD‐9) billing codes using a phecode map (version 1.2) (Wei et al., 2017). We required phenotypes to have at least 40 cases to be included in the PheWAS. A total of 1416 phenotypes met this requirement and were tested for association with the LTF SNP (rs1126478). An association test was performed for each phenotype/genotype pair using an additive model. Logistic regression was performed with age and sex as covariates. The analysis was run using Plink version 1.90 (https://www.cog‐genomics.org/plink2).

Case–control genetic study

Cases with hypertension and controls without hypertension were identified in BioVU, a de‐identified EHR combined with genetic information as described above (Roden et al., 2008). Hypertensive cases and control non‐hypertensive individuals were identified using a rigorous, validated random forest model which utilizes age, billing codes, medications, blood pressure readings, pulse, outpatient visits, counts of types of clinical notes and counts of hypertension‐related concepts extracted from clinical notes to identify hypertensive and control individuals in the medical record with a high degree of accuracy (Teixeira et al., 2017). Cases and controls were restricted to Caucasian adults with age at last encounter less than 80 and greater than 18 years. Records were excluded with missing data for BMI or age at last encounter, or for BMI less than 16 kg·m−2 or greater than 100 kg·m−2 as these were likely erroneously recorded in the EHR. For each case and control, date of birth, age at last encounter, sex, race, date of death and median BMI were extracted. Type 2 diabetes was defined using a published algorithm involving medications, laboratory values and diagnoses (Kho et al., 2012). Coronary artery disease (CAD) was defined by counting patients who had at least two ICD codes of 410.* (acute myocardial infarction), 411.* (other acute and subacute forms of ischaemic heart disease), 412.* (old myocardial infarction), 413.* (angina pectoris), 414.* (other forms of chronic ischaemic heart disease), V45.82 (percutaneous transluminal coronary angioplasty status) or at least one CPT code of 33534–33536, 33510–33523 (coronary artery bypass graft), 92980–92982, 92984, 92995, 92996 (intracoronary stent and angioplasty). A missense SNP in LTF (rs1126478) resulting in an A −> G substitution, with the G allele being the minor allele in European cohorts (Videm et al., 2012), was examined. Genotypic (additive) models were tested by logistic regression with weights of 0 for AA (homozygous major allele), 1 for AG (heterozygous) and 2 for GG (homozygous minor allele) genotypes using XLSTAT (New York, NY, USA). Dominant and recessive genetic models were tested by contingency analysis with Fisher's exact test using GraphPad Prism (La Jolla, CA, USA). Logistic regression was performed under a recessive genetic model in the presence of covariates of age, sex, BMI, and presence of CAD and diabetes mellitus type 2 (DM2) using XLSTAT. These studies were approved by the Vanderbilt University Institutional Review Board.

Data and statistical analysis

The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015). The nature of the human monocyte profiling, PheWAS and genetic case–control studies did not permit randomization or blinding. Data are presented as mean ± SEM unless otherwise stated. RT‐PCR data between groups were analysed using Student's t‐tests or Mann–Whitney tests as appropriate. Categorical variables between groups were analysed using chi‐square tests. Fisher's exact test was used for contingency testing for the LTF SNP case–control analysis. The above tests were performed with GraphPad Prism (GraphPad Software; La Jolla, CA, USA). Univariate and multivariate logistic and linear regression were performed using XLSTAT (New York, NY, USA). A P value <0.05 was considered significant. A Bonferroni corrected P value is also shown for the PheWAS study.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017a,b,c).

Results

RNA sequencing and analysis

The demographics of the five normotensive and seven hypertensive subjects in the original cohort are shown in Supporting Information Table S1. Hypertensive patients had higher blood pressures despite treatment and slightly higher (albeit not statistically significant) BMI. The strict combinatorial analytical approach described in the Methods section yielded a total of 60 transcripts in the monocytes isolated from these subjects that were identified as significantly differentially expressed after multiple hypothesis correction (P value <0.05; fold‐difference > 1.5) between normotensive and hypertensive groups (Supporting Information Table S2). These 60 transcripts along with their fold‐difference and P values are shown in Supporting Information Table S3. A positive fold‐difference indicates the gene is up‐regulated in hypertensive individuals, and a negative fold‐difference indicates the gene is down‐regulated in hypertensive individuals. Hierarchical clustering (Figure 1A) and PCA (Figure 1B) of the 60 most robust genes demonstrated that the combined expression profile differentiated hypertensive from normotensive subjects.

Figure 1.

Figure 1

Open in a new tab

(A) Hierarchical clustering of normalized RPKM for those genes that were identified as significantly differentially expressed using RNASeq. Intensity of red or green is used to show fold increase or decrease respectively. Columns represent individual samples, and each row represents one gene. (B) PCA of the 60 significant genes. Three components were sufficient to describe 80.6% of the variability among samples (48.8, 20.2 and 11.6% on the X, Y and Z axes respectively). (C) Functional network generated from upstream regulator analysis using IPA software. The three central proteins were predicted to be activated using IPA (P values = 1.5 × 10‐11 to 1.5 × 10‐9, Z scores = 2.9 to 3.9). Peripheral genes are overlaid with RNASeq expression values. Red or green colour represents up‐ or down‐regulation respectively. Intensity of the colour represents the degree of change. Orange lines with arrows indicate that directionality of the downstream gene is consistent with predicted activation of the upstream regulator. Yellow lines indicate inconsistent results between gene expression and IPA's knowledge base for that particular interaction. Grey lines indicate that directionality cannot be resolved. Gene abbreviations are defined in Supporting Information Table S3.

Using known relationships between gene expression changes and upstream factors involved in their regulation, upstream regulatory analysis using IPA inferred activation of IL‐1β, NF‐κB and the adaptor protein MyD88, based on the differential gene expression in hypertensive monocytes. Interestingly, these three molecules are part of the IL‐1β signalling pathway. Indeed, 21 of the 60 differentially expressed genes identified by RNASeq are related to IL‐1β (Figure 1C). We also employed IPA to identify non‐biased and significantly enriched functions in these 60 genes. The most statistically enriched disorder was ‘Inflammatory Response’, with a P value range of 5.4 × 10−4 to 4.6 × 10−15 (Supporting Information Table S4 ), which essentially indicates a very low probability that this functional category would be represented by a significant portion of 60 randomly selected genes. Based on causal prediction analysis, which uses individual relationships curated from the literature to infer functional directionality from transcriptional directionality (i.e. up‐ or down‐regulation of each functionally related gene), the functions associated with inflammatory responses were activated (Z score = 2.595 to 2.943). Enriched molecular and cellular functions include immune cell migration and signalling, phagocytosis, proliferation of vascular smooth muscle cells and T cell development (Supporting Information Table S4).

Regression analysis of gene determinants of blood pressure

To narrow the list of 60 differentially expressed genes, we performed univariate and multivariate regression analyses in controls, hypertensive individuals and/or all subjects together to identify transcripts that significantly correlated with MAP. The expression of 23 genes (log10 normalized RPKM) correlated significantly with MAP in univariate analyses in normotensive controls, hypertensive individuals or in these groups and all subjects together. However, by multivariate analyses, only LTF and PGLYRP1 remained significantly correlated with MAP in normotensive controls, and granzyme H (GZMH) remained significantly correlated with MAP in hypertensive subjects. Interestingly, the expression of IL18RAP correlated significantly with MAP of controls (r = 0.93), hypertensive patients (r = 0.78) and all subjects analysed together (r = 0.89). In summary, the expression of four genes, LTF, PGLYRP1, GZMH and IL18RAP was significantly correlated with MAP in normotensive and/or hypertensive subjects by multivariate analysis. The univariate correlations of these genes with MAP are shown in Figure 2. The corresponding multivariate models for MAP based on expression of these genes are shown in Supporting Information Table S5.

Figure 2.

Figure 2

Open in a new tab

Univariate regressions of MAP and the expression (log(10)RPKMs) of those genes that significantly correlated with MAP in multivariate regression analyses: (A) LTF, (B) PGLYRP1, (C) GZMH and (D) IL18RAP. Data are from control normotensive subjects (C, n = 5) and from hypertensive subjects (HTN, n = 7). Pearson correlation coefficients ( r ), when significant (P < 0.05), are given for controls, hypertensive subjects and all subjects analysed together (All, n = 12).

Quantitative RT‐PCR in a validation cohort

To determine the robustness of our observed gene expression changes in peripheral monocytes, we recruited a separate validation cohort of six normotensive and nine hypertensive subjects. As shown in Supporting Information Table S6, age, blood pressures and BMI were slightly but not significantly higher in hypertensive individuals compared to normotensives in this cohort. Frequency and type of antihypertensive treatment also varied slightly from the original cohort. As in the earlier cohort, the majority of the patients were Caucasian. We isolated peripheral blood monocytes and examined the expression of the four genes that were found to significantly correlate with MAP (LTF, GZMH, IL18RAP and PGLYRP1) by quantitative RT‐PCR. All but GZMH remained significantly elevated in peripheral monocytes of hypertensive individuals relative to controls (with GZMH showing a trend but not reaching statistical significance) (Figure 3). As further validation for the original observed gene expression changes, we investigated the expression of four additional genes with known or predicted effects on inflammation ARG1, CCL4, MMP8 and VSIG4. Expression of ARG1, VSIG4 and MMP8 was significantly altered in the predicted direction in hypertensive individuals in the validation cohort (Supporting Information Figure S1), demonstrating the robustness and reproducibility of the gene expression changes observed in the original cohort.

Figure 3.

Figure 3

Open in a new tab

Relative RT‐PCR quantification (normalized to GAPDH) of (A) LTF, (B) PGLYRP1, (C) GZMH and (D) IL18RAP in a validation cohort of normotensive (n = 6) and hypertensive (n = 9) patients. Data are plotted as mean ± SEM. * P < 0.05, significantly different as indicated; Student's t‐test or Mann–Whitney test.

Correlation between gene expression and MAP in African Americans

We then elected to test the three differentially expressed genes in the original and validation cohorts that associated with MAP (LTF, PGLYRP1 and IL18RAP) in a larger independent cohort from the MH‐GRID (Gaye et al., 2017). This cohort importantly differs from our original and validation cohorts by being composed exclusively of hypertensive African Americans. Also, gene expression was determined by RNA sequencing of whole blood rather than peripheral monocytes. Demographics of this cohort are shown in Supporting Information Table S7. We performed Pearson correlation analysis of mRNA transcript levels of LTF, PGLYRP1 and IL18RAP versus MAP. Interestingly, IL18RAP (but not LTF or PGLYRP1) significantly correlated with MAP in these individuals (Figure 4A), and this correlation remained significant after adjustment for age, gender and BMI (Supporting Information Table S8). To determine whether IL18RAP expression may be associated with end‐organ damage from hypertension, the relationship between IL18RAP expression and eGFR was examined in these hypertensive individuals and found to exhibit a significant negative correlation (Figure 4B). This correlation remained significant after adjustment for age, gender and BMI (Supporting Information Table S8).

Figure 4.

Figure 4

Open in a new tab

Correlation between MAP and IL18RAP expression (A) and between eGFR and IL18RAP expression (B) in hypertensive individuals (n = 76). Pearson correlation coefficients ( r ) are shown when significant (P < 0.05).

Phenome‐wide association of an LTF SNP

To further evaluate the role of the differentially expressed genes in our original and validation cohorts that correlate with MAP (LTF, IL18RAP and PGLYRP1), we sought to identify missense SNPs that alter protein function in these genes. Earlier reports had demonstrated a missense SNP (A −> G) in LTF (rs1126478) that changes a lysine to arginine at position 47. The G allele (arginine) is the minor allele in European cohorts and increases LTF protein levels while decreasing its antimicrobial protein function (Fine et al., 2013; Videm et al., 2012). We were unable to identify missense SNPs altering protein function for PGLYRP1 or IL18RAP in the literature.

Using a European cohort within BioVU, a de‐identified electronic medical record coupled with genetic information at Vanderbilt University Medical Center, we performed a PheWAS on the missense LTF SNP rs1126478 (Denny et al., 2010). In our PheWAS analysis, ‘Hypertension’ and ‘Essential Hypertension’ are separate but clinically similar phecodes. Results revealed that both Hypertension [odds ratio (OR) = 1.069, P = 0.0018] and Essential Hypertension (OR = 1.067, P = 0.0027) phecodes were among the top five phenotypes associated with the minor G allele in an additive model with a P value <0.05 (Figure 5). Of note, there were no phenotypes significantly associated with the LTF SNP when using a conservative Bonferroni corrected P value of 3.5 × 10−5.

Figure 5.

Figure 5

Open in a new tab

PheWAS Manhattan plot for the LTF SNP rs1126478 showing association between this SNP and 1416 different phenotypes. Disease groups on the x‐axis are shown in different colours, and the y‐axis reflects the P value for each phenotype. Purple and orange horizontal lines represent P value of 0.05 and Bonferroni corrected P value of 3.5 × 10−5 respectively.

Hypertension case–control study of the LTF SNP

Given the significant up‐regulation of LTF in monocytes of hypertensive individuals and the over‐representation of hypertension phenotypes with a missense SNP in LTF that increases protein levels, we performed a complementary case–control analysis of this SNP in hypertensive and normotensive control patients using a European cohort within BioVU. In contrast to the PheWAS study, where disease phenotypes were defined by ICD‐9 codes, this case–control approach permits focused determination of the relationship of the LTF SNP to hypertension using a much more rigorous algorithm for accurate identification of hypertensive individuals from the medical record that incorporates ICD codes, vitals, medications and natural language processing (Teixeira et al., 2017). Using this algorithm, we identified 10 549 hypertensive individuals and 5038 controls. The hypertensive patients were older with male sex predominance, higher BMI and increased prevalence of CAD and DM2, consistent with known risk factors and comorbidities for hypertension (Supporting Information Table S9) (Gelber et al., 2007; Gu et al., 2007). We found, using a recessive genetic model, that hypertensive individuals had a significantly increased frequency of homozygosity for the minor allele (GG) of the missense LTF SNP rs1126478 compared to controls (OR 1.16, P = 0.005) (Table 1). There was a trend for significance using an additive genetic model (OR 1.05; P value 0.08) and no significant association of rs1126478 with hypertension using a dominant (OR 1.02; P value 0.53) genetic model.

Table 1.

Case–control analysis using recessive genetic model for rs1126478

Normotensive controls (n = 5038) Hypertensive patients (n = 10 549)
AA+AG 4591 (91.1%) 9462 (89.7%)
GG 447 (8.9%) 1087 (10.3%)
Open in a new tab

A, major allele; G, minor allele. Odds ratio 1.16; 95% confidence interval 0.75–0.95; Fisher's exact P = 0.005.

To control for potential confounders that differed between cases and controls (Supporting Table S9), logistic regression was performed with the recessive genetic model for rs1126478 incorporating age, sex, BMI and presence of CAD and/or DM2. Results revealed that the homozygous minor allele of rs1126478 remained significantly associated with hypertension after inclusion of these covariates with a similar OR (1.18; P = 0.017) (Supporting Information Table S10).

Discussion

To elucidate changes in monocyte gene expression and identify novel immune targets in hypertension, we employed RNA sequencing of human peripheral blood monocytes and identified 60 differentially expressed transcripts between normotensive versus hypertensive subjects in a small discovery cohort. Multivariate regression analyses revealed that expression of four genes (LTF, PGLYRP1, GZMH and IL18RAP) correlated with MAP in control and/or hypertensive subjects. Of these, increased expression in hypertensive individuals was confirmed for IL18RAP, LTF and PGLYRP1 in an independent validation cohort, despite the fact that the two cohorts differed in their anti‐hypertensive medication profile. Of note, three out of four additional genes examined (ARG1, MMP8 and VSIG4) were also changed in the predicted direction in our validation cohort. In a larger cohort of 76 hypertensive African Americans, IL18RAP expression in whole blood significantly correlated with MAP and inversely correlated with eGFR. In addition, we found that homozygosity for the minor allele of a missense SNP in LTF that increases protein levels while decreasing antimicrobial function significantly increased the odds of hypertension in a European cohort of >15 000 individuals, even after controlling for covariates. Taken together, these results demonstrate that monocytes of hypertensive patients exhibit an enhanced pro‐inflammatory gene expression profile and highlight IL18RAP and LTF as potential novel mediators and therapeutic targets of human hypertension.

IPA revealed that 21 of the 60 genes are related to the IL‐1β pathway (Figure 1C). These results are consistent with earlier work demonstrating increased IL‐1β production (Dorffel et al., 1999; Wirtz et al., 2004) in monocytes from hypertensive patients. The CANTOS trial recently demonstrated that inhibiting IL‐1β with canakinumab reduces cardiovascular events in post‐myocardial infarction patients with elevated C‐reactive protein (CRP) (Ridker et al., 2017). While results on blood pressure or incidence of hypertension were not reported in this trial, it is interesting to speculate that perhaps some of the benefit in cardiovascular risk reduction may be through a reduction in hypertension or hypertension‐associated inflammation, particularly given that CRP elevations are associated with hypertension as well as CAD (Davey Smith et al., 2005). Earlier studies in animal models have indeed demonstrated an important role of IL‐1 signalling in promoting hypertension (Ling et al., 2017; Zhang et al., 2016). Of note, it is possible that some of our observed transcriptional changes are due to phosphorylation and activation of STAT3 as STAT3 has been shown to enhance IL‐1β production from LPS stimulated macrophages (Samavati et al., 2009) and from monocytes/macrophages from patients with CAD (Shirai et al., 2016). Furthermore, mice with reduced STAT3 activity display blunted angiotensin II induced hypertension (Zouein et al., 2013).

Our results suggest an important role for IL18RAP (also known as IL18Rβ) in hypertension, as expression was increased in hypertensive monocytes from two separate cohorts and expression correlated with MAP in our primarily Caucasian discovery cohort as well as a larger African American cohort. IL18RAP binds to the IL‐18 receptor α subunit and enhances IL‐18 signalling through the heterodimeric receptor complex (Krishnan et al., 2014). Both IL‐1β and IL‐18 are products of the inflammasome, a multimeric protein complex important in innate immunity, and have been shown to be elevated in hypertensive patients (Krishnan et al., 2014; Malik and Kanneganti, 2017). Drummond and colleagues have provided evidence that the inflammasome and its products play a critical role in hypertension (Krishnan et al., 2016; Krishnan et al., 2014). Of note, the inflammasome is activated by damage associated molecular pattern receptors (DAMPs) and pathogen associated molecular pattern receptors (PAMPs) that bind to toll‐like receptors and signal through NF‐κB. The exact DAMP or PAMP that activates the inflammasome in hypertension is unknown. While one would speculate that a DAMP might contribute to immune activation in hypertension, our results suggest that the PAMP, PGLYRP1, might also play an important role. PGLYRP1 binds to triggering receptor expressed on myeloid cells (TREM)‐1 either as a complex with bacterial peptidoglycan or in the absence of bacterial peptidoglycan when it is multimerized (Read et al., 2015). Signalling through TREM‐1 results in activation of NF‐κB and NFAT‐mediated transcription. Interestingly, PGLYRP1 has been implicated in ROS formation in neutrophils (Dziarski et al., 2003). As hypertensive monocytes express increased isoketals, a by‐product of oxidative injury (Kirabo et al., 2014), it is attractive to hypothesize that PGLYRP1 may be upstream of ROS production in monocytes. Notably, Rohatgi et al. (2009) showed that increased PGLYRP1 levels are associated with all major cardiovascular risk factors (including hypertension) and with inflammatory markers such as CRP and IL‐18. Further studies are needed to investigate whether PGLYRP1 might increase isoketal formation in monocytes and what effect IL‐18 signalling through IL8RAP may have on monocyte activation and function, including adherence to endothelial cells. Interestingly, IL‐18 levels are increased in obesity and metabolic syndrome disorders, and polymorphisms in IL18RAP have been shown to influence susceptibility to obesity (Martinez‐Barquero et al., 2017).

Importantly, our results also suggest a novel role for lactoferrin (also known as LTF or lactotransferrin) in the pathogenesis of hypertension. LTF is an iron‐binding glycoprotein present at high levels in exocrine secretions such as saliva and milk as well as in neutrophil granules (Kruzel et al., 2017). LTF exerts antimicrobial activity through multiple mechanisms including binding LPS, inhibition of bacterial adhesion to host cells and surfaces, and iron sequestration (Rosa et al., 2017). LTF can also modulate the migration, maturation and functions of immune cells, thus influencing both innate and adaptive immunity (Legrand & Mazurier, 2010). Interestingly, recombinant human LTF induces maturation of human monocyte‐derived dendritic cells, leading to up‐regulation of HLA class II, CD83, CD80 and CD86 costimulatory molecules, and CXCR4 and CCR7 chemokine receptors (Spadaro et al., 2008). We showed that monocyte LTF production is increased in hypertensive individuals and correlates with MAP. Moreover, we provide functional evidence for a role of LTF in human hypertension by demonstrating that homozygosity for the minor allele of a missense SNP (A −> G) in LTF (rs1126478) that changes a lysine to arginine at position 47 is more prevalent in hypertensive individuals of European descent compared to normotensive controls. The G allele of this SNP is associated with increased LTF protein levels but interestingly, with decreased antimicrobial activity against dental caries‐provoking microbes such as S. mutans (Fine et al., 2013; Videm et al., 2012). Of note, there is a well‐established association between periodontal disease and cardiovascular disease, including hypertension (Leong et al., 2014; Rosa et al., 2017). While the cause for this association is likely to be multifactorial, including increased production of pro‐inflammatory cytokines, it is interesting to speculate that changes in lactoferrin levels or activity may be partly responsible for the connection between poor oral health and hypertension. It should be noted that the minor G allele frequency in our BioVU cohort was 0.31, similar to that seen in other European cohorts (Videm et al., 2012). However, in African cohorts, the G allele is the dominant allele with a frequency of 0.9–1.0 [National Center for Biotechnology Information. Reference snp (refsnp) cluster report]. This suggests that varying selection pressures, perhaps related to altered infectious organisms or a protective effect of enhanced iron‐sequestration capacity in diseases such as sickle cell anaemia, may have driven different allele frequencies in the different populations. The predominance of the G allele, which corresponds to higher LTF levels, may also partly explain the increased incidence of hypertension in African populations. Further evidence for a link between LTF and hypertension comes from a study showing that chorionic villous expression of LTF is higher in patients with pre‐eclampsia compared to controls (Farina et al., 2009). Interestingly, LTF stimulation of intestinal epithelial cells or macrophages has been shown to induce IL‐18 production (Iigo et al., 2004; Wang et al., 2005), raising the intriguing possibility that LTF and IL18RAP may act through a common pathway in the pathogenesis of hypertension. Of note, in a Spanish population, higher circulating LTF levels and the presence of the G allele for rs1126478 was associated with lower fasting triglyceride concentration and higher HDL cholesterol concentrations, suggesting that LTF may have a beneficial effect on lipid profiles (Moreno‐Navarrete et al., 2008). Future studies are needed to understand precisely how LTF may contribute to the pathogenesis of hypertension and other metabolic disorders.

There are several limitations of our studies that are worth noting. First, nearly all the hypertensive patients in our original, validation and MH‐GRID cohorts were on antihypertensive medications (Supporting Information Tables S1, S6 and S7). While the wide variety of antihypertensive medication classes and combinations used in these patients reduces the risk that our observed gene expression differences were influenced by an individual medication class, it remains possible that the presence of antihypertensive therapy could have influenced some of our findings. Second, we isolated total peripheral monocytes for analysis of gene expression in the original and validation cohorts in order to determine overall differences in gene expression in monocytes in hypertension. Given the earlier studies demonstrating differences in inflammatory cytokine production in different monocyte subsets (Chimen et al., 2017) and dietary factors such as high salt altering proportions of peripheral monocyte subsets (Zhou et al., 2013), future studies extending our results to distinct monocyte subsets in hypertension will be of interest. Finally, while our studies are significant in that they are conducted in humans and utilize several distinct approaches to determine the importance of the observed gene expression changes in the pathogenesis of human hypertension, they are limited by the difficulty in establishing definitive causality. However, our results do establish an important foundation for further studies to more definitively test causality and mechanisms in animal models of hypertension with genetic and/or pharmacological inhibition of IL18RAP and/or LTF.

In conclusion, our unbiased investigation of monocyte gene expression demonstrates a pro‐inflammatory expression profile in hypertensive individuals, with IL‐1β as a prominent driver of the observed gene expression changes. Furthermore, additional studies in larger African and European cohorts provide compelling and provocative evidence that IL18RAP and LTF may play an important role in hypertension and potentially serve as novel immune targets for this disease.

Author Contributions

M.R.M., A.E.N., F.E., R.V.A., A.G., J.S.G., C.L.L., C.L.G. and M.S.M. made substantial contributions to conception and design, acquisition of data and/or analysis/interpretation of data. M.R.M., A.E.N., F.E., C.L.G. and M.S.M. were involved in drafting the manuscript and revising it critically for important intellectual content.

Conflict of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organisations engaged with supporting research.

Supporting information

Figure S1 Relative RT‐PCR quantification (normalized to GAPDH) of (A) arginase 1 (Arg1), (B) chemokine (C‐C motif) ligand 4 (CCL4), (C) matrix metalloproteinase 8 (MMP8), and (D) V‐set and immunoglobulin domain containing 4 (VSIG4) in a validation cohort of control normotensive (n = 6) and hypertensive (n = 9) patients. Data are plotted as mean±SEM. *P < 0.05 by one‐tailed Mann‐Whitney test.

Table S1 Demographics table for original cohort.

Table S2 Overview of Statistical Results.

Table S3 List of the 60 Significant Genes and their fold‐change.

Table S4 Functional Analysis of Significant Genes.

Table S5 Multivariate Regression Models for MAP using expression levels of LTF, PGLYRP1, GZMH, and/or IL18RAP.

Table S6 Demographics Table for Validation Cohort.

Table S7 Demographics of hypertensive individuals from MH‐GRID.

Table S8 Linear Regression of IL18RAP expression versus MAP or eGFR with and without adjustment for age, gender and BMI.

Table S9 Demographics of case and control individuals in BioVU.

Table S10 Logistic regression for LTF SNP rs1126478 with hypertension including covariates.

Click here for additional data file. (259.7KB, pdf)

Acknowledgements

We would like to thank the Vanderbilt Technologies for Advanced Genomics Core for performing the RNA sequencing and the Vanderbilt Clinical Trials Center for assisting with subject recruitment and sample collection. We would also like to thank Lisa Bastarache and Janey Wang for help with PheWAS analysis as well as Rakale C. Quarells and Gary H. Gibbons for providing data from ‘The Minority Health Genomics and Translational Research Bio‐Repository Database (MH‐GRID)’. This work was supported by a training grant from the National Institutes of Health (NIH T32 HL069765‐11A1) and an NIH NRSA Award (F31 HL127986) to A.E.N., an NIH T32 HL007411‐37 to M.R.A., an NIH K01 Award (HL121045) to C.L.G. and an NIH K08 Award (HL121671) to M.S.M. The PheWAS and case–control datasets were obtained from Vanderbilt University Medical Center's BioVU which is supported by institutional funding and by the CTSA grant ULTR000445 from NCATS/NIH. PheWAS Core resource for analysis was supported by R01 LM010685. Genome‐wide genotyping was funded by NIH grants RC2GM092618 from NIGMS/OD and U01HG004603 from NHGRI/NIGMS.

Alexander M. R., Norlander A. E., Elijovich F., Atreya R. V., Gaye A., Gnecco J. S., Laffer C. L., Galindo C. L., and Madhur M. S. (2019) Human monocyte transcriptional profiling identifies IL‐18 receptor accessory protein and lactoferrin as novel immune targets in hypertension, British Journal of Pharmacology, 176, 2015–2027, doi: 10.1111/bph.14364.

References

  1. Alexander SPH, Fabbro D, Kelly E, Marrion NV, Peters JA, Faccenda E et al (2017a). The Concise Guide to PHARMACOLOGY 2017/18: Catalytic receptors. Br J Pharmacol 174: S225–S271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alexander SPH, Fabbro D, Kelly E, Marrion NV, Peters JA, Faccenda E et al (2017b). The Concise Guide to PHARMACOLOGY 2017/18: Enzymes. Br J Pharmacol 174: S272–S359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Alexander SPH, Christopoulos A, Davenport AP, Kelly E, Marrion NV, Peters JA et al (2017c). The Concise Guide to PHARMACOLOGY 2017/18: G protein‐coupled receptors. Br J Pharmacol 174: S17–S129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chimen M, Yates CM, McGettrick HM, Ward LS, Harrison MJ, Apta B et al (2017). Monocyte subsets coregulate inflammatory responses by integrated signaling through TNF and IL‐6 at the endothelial cell interface. J Immunol 198: 2834–2843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Curtis MJ, Bond RA, Spina D, Ahluwalia A, Alexander SPA, Giembycz MA et al (2015). Experimental design and analysis and their reporting: new guidance for publication in BJP. Br J Pharmacol 172: 3461–3471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Davey Smith G, Lawlor DA, Harbord R, Timpson N, Rumley A, Lowe GD et al (2005). Association of C‐reactive protein with blood pressure and hypertension: life course confounding and mendelian randomization tests of causality. Arterioscler Thromb Vasc Biol 25: 1051–1056. [DOI] [PubMed] [Google Scholar]
  7. Denny JC, Ritchie MD, Basford MA, Pulley JM, Bastarache L, Brown‐Gentry K et al (2010). PheWAS: demonstrating the feasibility of a phenome‐wide scan to discover gene‐disease associations. Bioinformatics 26: 1205–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dorffel Y, Franz S, Pruss A, Neumann G, Rohde W, Burmester GR et al (2001). Preactivated monocytes from hypertensive patients as a factor for atherosclerosis? Atherosclerosis 157: 151–160. [DOI] [PubMed] [Google Scholar]
  9. Dorffel Y, Latsch C, Stuhlmuller B, Schreiber S, Scholze S, Burmester GR et al (1999). Preactivated peripheral blood monocytes in patients with essential hypertension. Hypertension 34: 113–117. [DOI] [PubMed] [Google Scholar]
  10. Dziarski R, Platt KA, Gelius E, Steiner H, Gupta D (2003). Defect in neutrophil killing and increased susceptibility to infection with nonpathogenic gram‐positive bacteria in peptidoglycan recognition protein‐S (PGRP‐S)‐deficient mice. Blood 102: 689–697. [DOI] [PubMed] [Google Scholar]
  11. Farina A, Morano D, Arcelli D, De Sanctis P, Sekizawa A, Purwosunu Y et al (2009). Gene expression in chorionic villous samples at 11 weeks of gestation in women who develop preeclampsia later in pregnancy: implications for screening. Prenat Diagn 29: 1038–1044. [DOI] [PubMed] [Google Scholar]
  12. Fine DH, Toruner GA, Velliyagounder K, Sampathkumar V, Godboley D, Furgang D (2013). A lactotransferrin single nucleotide polymorphism demonstrates biological activity that can reduce susceptibility to caries. Infect Immun 81: 1596–1605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gaye A, Gibbons GH, Barry C, Quarells R, Davis SK (2017). Influence of socioeconomic status on the whole blood transcriptome in African Americans. PLoS One 12: e0187290. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gelber RP, Gaziano JM, Manson JE, Buring JE, Sesso HD (2007). A prospective study of body mass index and the risk of developing hypertension in men*. Am J Hypertens 20: 370–377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gu D, Wildman RP, Wu X, Reynolds K, Huang J, Chen CS et al (2007). Incidence and predictors of hypertension over 8 years among Chinese men and women. J Hypertens 25: 517–523. [DOI] [PubMed] [Google Scholar]
  16. Harding SD, Sharman JL, Faccenda E, Southan C, Pawson AJ, Ireland S et al (2018). The IUPHAR/BPS Guide to PHARMACOLOGY in 2018: updates and expansion to encompass the new guide to IMMUNOPHARMACOLOGY. Nucl Acids Res 46: D1091–D1106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Iigo M, Shimamura M, Matsuda E, Fujita K‐i, Nomoto H, Satoh J et al (2004). Orally administered bovine lactoferrin induces caspase‐1 and interleukin‐18 in the mouse intestinal mucosa: a possible explanation for inhibition of carcinogenesis and metastasis. Cytokine 25: 36–44. [DOI] [PubMed] [Google Scholar]
  18. Justin Rucker A, Crowley SD (2017). The role of macrophages in hypertension and its complications. Pflugers Archiv 469: 419–430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kho AN, Hayes MG, Rasmussen‐Torvik L, Pacheco JA, Thompson WK, Armstrong LL et al (2012). Use of diverse electronic medical record systems to identify genetic risk for type 2 diabetes within a genome‐wide association study. J Am Med Inform Assoc 19: 212–218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kirabo A, Fontana V, de Faria AP, Loperena R, Galindo CL, Wu J et al (2014). DC isoketal‐modified proteins activate T cells and promote hypertension. J Clin Invest 124: 4642–4656. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Krishnan SM, Sobey CG, Latz E, Mansell A, Drummond GR (2014). IL‐1beta and IL‐18: inflammatory markers or mediators of hypertension? Br J Pharmacol 171: 5589–5602. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Krishnan SM, Dowling JK, Ling YH, Diep H, Chan CT, Ferens D et al (2016). Inflammasome activity is essential for one kidney/deoxycorticosterone acetate/salt‐induced hypertension in mice Br J Pharmacol 173: 752–765. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kruzel ML, Zimecki M, Actor JK (2017). Lactoferrin in a context of inflammation‐induced pathology. Front Immunol 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Legrand D, Mazurier J (2010). A critical review of the roles of host lactoferrin in immunity. Biometals 23: 365–376. [DOI] [PubMed] [Google Scholar]
  25. Leong XF, Ng CY, Badiah B, Das S (2014). Association between hypertension and periodontitis: possible mechanisms. Sci World J 2014: 768237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li B, Ruotti V, Stewart RM, Thomson JA, Dewey CN (2010). RNA‐Seq gene expression estimation with read mapping uncertainty. Bioinformatics 26: 493–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ling YH, Krishnan SM, Chan CT, Diep H, Ferens D, Chin‐Dusting J et al (2017). Anakinra reduces blood pressure and renal fibrosis in one kidney/DOCA/salt‐induced hypertension. Pharmacol Res 116: 77–86. [DOI] [PubMed] [Google Scholar]
  28. Malik A, Kanneganti TD (2017). Inflammasome activation and assembly at a glance. J Cell Sci 130: 3955–3963. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Martinez‐Barquero V, Marco G, Martinez‐Hervas S, Adam‐Felici V, Perez‐Soriano C, Gonzalez‐Albert V et al (2017). Are IL18RAP gene polymorphisms associated with body mass regulation? A cross‐sectional study. BMJ Open 7: e017875. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. McMaster WG, Kirabo A, Madhur MS, Harrison DG (2015). Inflammation, immunity, and hypertensive end‐organ damage. Circ Res 116: 1022–1033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Moreno‐Navarrete JM, Ortega FJ, Bassols J, Castro A, Ricart W, Fernandez‐Real JM (2008). Association of circulating lactoferrin concentration and 2 nonsynonymous LTF gene polymorphisms with dyslipidemia in men depends on glucose‐tolerance status. Clin Chem 54: 301–309. [DOI] [PubMed] [Google Scholar]
  32. National Center for Biotechnology Information . Reference snp (refsnp) cluster report: rs1126478. https://wwwncbinlmnihgov/projects/SNP/snp_refcgi?rs=1126478 (accessed 7/2/2018).
  33. Norlander AE, Saleh MA, Kamat NV, Ko B, Gnecco J, Zhu L et al (2016). Interleukin‐17A regulates renal sodium transporters and renal injury in angiotensin ii‐induced hypertension. Hypertension 68: 167–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Porras‐Hurtado L, Ruiz Y, Santos C, Phillips C, Carracedo Á, Lareu MV (2013). An overview of STRUCTURE: applications, parameter settings, and supporting software. Front Genet 4: 98. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Read CB, Kuijper JL, Hjorth SA, Heipel MD, Tang X, Fleetwood AJ et al (2015). Cutting edge: identification of neutrophil PGLYRP1 as a ligand for TREM‐1. J Immunol 194: 1417–1421. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ridker PM, Everett BM, Thuren T, MacFadyen JG, Chang WH, Ballantyne C et al (2017). Antiinflammatory therapy with canakinumab for atherosclerotic disease. N Engl J Med 377: 1119–1131. [DOI] [PubMed] [Google Scholar]
  37. Roden DM, Pulley JM, Basford MA, Bernard GR, Clayton EW, Balser JR et al (2008). Development of a large‐scale de‐identified DNA biobank to enable personalized medicine. Clin Pharmacol Ther 84: 362–369. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rohatgi A, Ayers CR, Khera A, McGuire DK, Das SR, Matulevicius S et al (2009). The association between peptidoglycan recognition protein‐1 and coronary and peripheral atherosclerosis: observations from the Dallas Heart Study. Atherosclerosis 203: 569–575. [DOI] [PubMed] [Google Scholar]
  39. Rosa L, Cutone A, Lepanto MS, Paesano R, Valenti P (2017). Lactoferrin: a natural glycoprotein involved in iron and inflammatory homeostasis. Int J Mol Sci 18: 1985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Saleh MA, McMaster WG, Wu J, Norlander AE, Funt SA, Thabet SR et al (2015). Lymphocyte adaptor protein LNK deficiency exacerbates hypertension and end‐organ inflammation. J Clin Invest 125: 1189–1202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Samavati L, Rastogi R, Du W, Huttemann M, Fite A, Franchi L (2009). STAT3 tyrosine phosphorylation is critical for interleukin 1 beta and interleukin‐6 production in response to lipopolysaccharide and live bacteria. Mol Immunol 46: 1867–1877. [DOI] [PubMed] [Google Scholar]
  42. Shirai T, Nazarewicz RR, Wallis BB, Yanes RE, Watanabe R, Hilhorst M et al (2016). The glycolytic enzyme PKM2 bridges metabolic and inflammatory dysfunction in coronary artery disease. J Exp Med 213: 337–354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Spadaro M, Caorsi C, Ceruti P, Varadhachary A, Forni G, Pericle F et al (2008). Lactoferrin, a major defense protein of innate immunity, is a novel maturation factor for human dendritic cells. FASEB J 22: 2747–2757. [DOI] [PubMed] [Google Scholar]
  44. Teixeira PL, Wei WQ, Cronin RM, Mo H, VanHouten JP, Carroll RJ et al (2017). Evaluating electronic health record data sources and algorithmic approaches to identify hypertensive individuals. J Am Med Inform Assoc 24: 162–171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Videm V, Dahl H, Walberg LE, Wiseth R (2012). Functional polymorphisms in the LTF gene and risk of coronary artery stenosis. Hum Immunol 73: 554–559. [DOI] [PubMed] [Google Scholar]
  46. Wang WP, Iigo M, Sato J, Sekine K, Adachi I, Tsuda H (2005). Activation of intestinal mucosal immunity in tumor‐bearing mice by lactoferrin. Jpn J Cancer Res 91: 1022–1027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Wei WQ, Bastarache LA, Carroll RJ, Marlo JE, Osterman TJ, Gamazon ER et al (2017). Evaluating phecodes, clinical classification software, and ICD‐9‐CM codes for phenome‐wide association studies in the electronic health record. PLoS One 12: e0175508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wenzel P, Knorr M, Kossmann S, Stratmann J, Hausding M, Schuhmacher S et al (2011). Lysozyme M‐positive monocytes mediate angiotensin II‐induced arterial hypertension and vascular dysfunction. Circulation 124: 1370–1381. [DOI] [PubMed] [Google Scholar]
  49. Wirtz PH, von Kanel R, Frey K, Ehlert U, Fischer JE (2004). Glucocorticoid sensitivity of circulating monocytes in essential hypertension. Am J Hypertens 17: 489–494. [DOI] [PubMed] [Google Scholar]
  50. Zhang J, Rudemiller NP, Patel MB, Karlovich NS, Wu M, McDonough AA et al (2016). Interleukin‐1 receptor activation potentiates salt reabsorption in angiotensin II‐induced hypertension via the NKCC2 Co‐transporter in the Nephron. Cell Metab 23: 360–368. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhao Z, Ni Y, Chen J, Zhong J, Yu H, Xu X et al (2012). Increased migration of monocytes in essential hypertension is associated with increased transient receptor potential channel canonical type 3 channels. PLoS One 7: e32628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhou X, Zhang L, Ji WJ, Yuan F, Guo ZZ, Pang B et al (2013). Variation in dietary salt intake induces coordinated dynamics of monocyte subsets and monocyte‐platelet aggregates in humans: implications in end organ inflammation. PLoS One 8: e60332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zouein FA, Zgheib C, Hamza S, Fuseler JW, Hall JE, Soljancic A et al (2013). Role of STAT3 in angiotensin II‐induced hypertension and cardiac remodeling revealed by mice lacking STAT3 serine 727 phosphorylation. Hypertens Res 36: 496–503. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1 Relative RT‐PCR quantification (normalized to GAPDH) of (A) arginase 1 (Arg1), (B) chemokine (C‐C motif) ligand 4 (CCL4), (C) matrix metalloproteinase 8 (MMP8), and (D) V‐set and immunoglobulin domain containing 4 (VSIG4) in a validation cohort of control normotensive (n = 6) and hypertensive (n = 9) patients. Data are plotted as mean±SEM. *P < 0.05 by one‐tailed Mann‐Whitney test.

Table S1 Demographics table for original cohort.

Table S2 Overview of Statistical Results.

Table S3 List of the 60 Significant Genes and their fold‐change.

Table S4 Functional Analysis of Significant Genes.

Table S5 Multivariate Regression Models for MAP using expression levels of LTF, PGLYRP1, GZMH, and/or IL18RAP.

Table S6 Demographics Table for Validation Cohort.

Table S7 Demographics of hypertensive individuals from MH‐GRID.

Table S8 Linear Regression of IL18RAP expression versus MAP or eGFR with and without adjustment for age, gender and BMI.

Table S9 Demographics of case and control individuals in BioVU.

Table S10 Logistic regression for LTF SNP rs1126478 with hypertension including covariates.

Click here for additional data file. (259.7KB, pdf)

Articles from British Journal of Pharmacology are provided here courtesy of The British Pharmacological Society

RESOURCES