Elevated Lipoprotein-Associated Phospholipase A2 Independently Affects Age-Related Increases in Systolic Blood Pressure: A Nested Case-Control Study in a Prospective Korean Cohort

Inflammatory markers are susceptible to changes over time. Thus, we observed changes in inflammatory markers correlating with age-related increases in blood pressure (BP) through a prospective study. The aim of this study was to investigate changes in inflammatory markers that correlate with age-related increases in BP. The study included 1,500 nondiabetic and normotensive healthy subjects at baseline. Of these, 121 individuals who developed hypertension (defined as systolic BP ≥ 140 mmHg or diastolic BP ≥ 90 mmHg) after 2 years formed the hypertension group. For each incident hypertension case, 2 age- and sex-matched control subjects were selected among those who did not develop hypertension (control group, n = 242). After baseline adjustment, the hypertension group exhibited greater increases in body mass index (BMI), systolic and diastolic BP, triglyceride, total cholesterol, glucose, Lp-PLA2 activity, and urinary 8-epi-prostaglandin F2α (8-epi-PGF2α) levels compared to the control group. In the hypertension group, changes in (Δ) systolic BP correlated positively with Δ Lp-PLA2 activity, which correlated positively with Δ low-density lipoprotein (LDL−) cholesterol and Δ urinary 8-epi-PGF2α levels. Moreover, multiple linear regression revealed baseline systolic BP and Δ Lp-PLA2 activity to be independent predictors of Δ systolic BP in the hypertension group. Our results suggest that age-related increases in systolic BP may correlate strongly with elevated Lp-PLA2 activity and that Lp-PLA2 can be considered a biomarker for systolic BP elevation.


Introduction
Evidence has emerged to support that inflammation may contribute to hypertension development [1,2]. Systemic inflammation is associated with the progression of endothelial dysfunction [3], resulting in changes in the structure and function of the endothelium that are often obvious during the early stage of hypertension development [4,5]. Cross-sectional studies have consistently and significantly shown that inflammatory markers and systolic and diastolic blood pressure (BP) correlate positively [6,7]. According to prospective studies investigating inflammatory markers, high-sensitivity C-reactive protein (hs-CRP) is related to the risk of developing hypertension [8]. In a nested case-control study, interleukin (IL)-6 was less highly related to the risk of developing hypertension than hs-CRP [9]. However, the mechanisms by which inflammatory markers other than hs-CRP and IL-6 are related to hypertension risk remain unclear. Additionally, the analysis of plasma inflammatory markers that are susceptible to changes over time and measured only once at baseline may have introduced arbitrary misclassifications and resulted in an underestimation of the true relative risks [10]. We are also unaware of prospective studies evaluating the relationship between lipoproteinassociated phospholipase A 2 (Lp-PLA 2 ) activity and hypertension risk. erefore, we conducted a nested case-control study to investigate changes in inflammatory markers [hs-CRP, tumor necrosis factor (TNF)-α, IL-1β, IL-6, and Lp-PLA 2 ] with incident hypertension in a mean follow-up 2-year prospective cohort of 1,500 nondiabetic and normotensive healthy subjects aged 20 to 69 years. We also measured urinary excretion of 8-epi-prostaglandin F 2α (8-epi-PGF 2α ), which is a sensitive marker of oxidative stress [11,12].

Subjects.
We performed a nested case-control study with a mean follow-up of 2 years in a prospective cohort that included 1,500 nondiabetic (normal fasting glucose (NFG) < 126 mg/dL) and normotensive (defined as systolic BP < 140 mmHg, diastolic BP < 90 mmHg, and no history of antihypertensive medication use) healthy subjects aged 20 to 69 years who received a biennial medical check-up at the National Health Insurance Corporation Ilsan Hospital in Goyang, Korea (from January 2010 to December 2015). Among the patients, 121 subjects (8.07%) developed hypertension (systolic BP ≥ 140 mmHg or diastolic BP ≥ 90 mmHg) after followup without taking any medication. ese individuals formed a hypertension group (n � 121). For each of the 121 cases of incident hypertension, 2 subjects matched based on age (±1 year) and sex were randomly selected from among the subjects who did not develop hypertension as a control (control group, n � 242). e exclusion criteria were as follows: current and/or a history of diseases including hypertension, cardiovascular disease (CVD), diabetes, dyslipidemia, liver disease, renal disease, pancreatitis, or cancer; pregnancy or lactation; and the use of any medications or supplements. e study purpose was fully explained to all participants, and each of them provided written informed consent.
e Institutional Review Board of Yonsei University and Ilsan Hospital approved the study protocol, which complied with the Declaration of Helsinki.

Anthropometric Measurements.
e body weight (UM0703581; Tanita, Tokyo, Japan) and height (GL-150; G-tech International, Uijeongbu, Korea) of each subject were measured while they wore lightweight clothes with no shoes, and corresponding body mass index (BMI) values were calculated (kg/m 2 ). A random-zero sphygmomanometer (HM-1101, Hico Medical Co., Ltd., Chiba, Japan) with appropriate-sized cuffs was used to measure systolic and diastolic BP after at least 20 min of rest in a seated position. BP was assessed three times; as differences among the three measurements of systolic BP were always below 2 mmHg, the BPs were deemed stable. e participants were instructed that they should not smoke or drink alcohol before each BP measurement. e mean BP measurement value was used for statistical analysis.

Blood and Urine Sample Collection.
Plasma and serum specimens were obtained from the participants. Specifically, venous blood was collected after at least 12 h of an overnight fast in EDTA-treated and serum tubes (BD Vacutainer; Becton, Dickinson and Company, Franklin Lakes, NJ, USA). e tubes were centrifuged (1,200 rpm, 20 min, 4°C) to separate the plasma and serum, and aliquots were stored at −80°C prior to analysis.
Urine specimens were collected after at least 12 h of an overnight fast in a polyethylene tube containing 1% butylated hydroxytoluene. e tubes were immediately protected from light and stored at −20°C prior to analysis.

Statistical
Analysis. SPSS version 23.0 (IBM, Chicago, IL, USA) was utilized for statistical analyses. Skewed variables were transformed to a logarithmic form for normalization. For descriptive purposes, the results are presented as means ± standard error (SE) without logarithmic transformation. Statistical significance was considered at a two-tailed p-value less than 0.05. Independent t-tests were used to compare continuous variables between the control and hypertension groups. Comparisons of nominal variables between the two groups were performed using a Chisquared test. General linear model analysis was also conducted with adjustment for potential confounding factors. Paired t-tests were conducted to verify differences between the values at baseline and 2-year follow-up in each group. Multiple linear regression was applied to verify significant independent predictors of changes in systolic and diastolic BP. Pearson's correlation coefficients were calculated to determine associations between variables. A heat map was produced to visualize and evaluate associations among variables in the study population (MeV v.4.9.0; http://mev. tm4.org).

Clinical Characteristics and Inflammatory Markers at
Baseline and 2-Year Follow-Up. No significant differences were found between the control (normal BP at both baseline and follow-up) and hypertension (normal BP at baseline but development of hypertension during follow-up) groups with regard to baseline characteristics, including sex (51.2% males and 48.8% females), age (controls, 48.7 ± 0.65 years; hypertension group, 49.3 ± 0.97 years), smoking status (controls, 20.2% current smokers; hypertension group, 13.2% current smokers), and alcohol consumption (controls, 59.1% current drinkers; hypertension group, 65.3% current drinkers). As shown in Table 1, subjects who developed hypertension were significantly heavier and had higher systolic and diastolic BP, hs-CRP levels, and IL-1β levels at baseline than did subjects who did not develop hypertension during the mean follow-up of 2 years. After 2 years, the hypertension group showed significant increases in BMI, systolic and diastolic BP, total and LDL-cholesterol levels, glucose levels, and HOMA-IR index; these increases were significantly greater than the increases observed in the controls (with the exception of LDL-cholesterol levels and HOMA-IR index) after adjusting for baseline values. Meanwhile, after 2 years, the controls displayed a significant increase in LDL-cholesterol levels. At the 2-year follow-up, higher systolic and diastolic BP, and IL-1β levels were found in the hypertension group compared to the controls after adjusting for follow-up BMI (Table 1).

Lp-PLA 2 Activity and 8-Epi-PGF 2α at Baseline and 2-Year
Follow-Up. No significant differences were observed between the control and hypertension groups in baseline levels of plasma Lp-PLA 2 activity and urinary 8-epi-PGF 2α (Figure 1). After 2 years, levels of plasma Lp-PLA 2 activity and urinary 8-epi-PGF 2α were significantly increased in the hypertension group; these increases were significantly greater than those in controls after adjustment for the baseline values. At the 2-year follow-up, the hypertension group exhibited higher Lp-PLA 2 activity and 8-epi-PGF 2α levels than the control group after adjusting for follow-up BMI ( Figure 1). Figure 2 shows the relationships between changes in (differences from baseline, Δ) Lp-PLA 2 activity with Δ systolic BP, Δ LDL-cholesterol levels, and Δ urinary 8-epi-PGF 2α levels in both groups. In the hypertension group, Δ Lp-PLA 2 activity correlated positively and strongly with Δ systolic BP (r � 0.793, p < 0.001), Δ LDL-cholesterol (r � 0.695, p < 0.001) and Δ urinary 8-epi-PGF 2α (r � 0.731, p < 0.001) (Figure 2). A correlation matrix of changes in major clinical characteristics, biochemical parameters, and inflammatory markers in all subjects (n � 363) is depicted in Figure 3.

Correlations between Changes in Clinical Variables, Inflammatory Markers, and Oxidative Stress Markers.
Multiple linear regression was conducted within the hypertension group to determine the major clinical factors related to Δ systolic BP or Δ diastolic BP (dependent variables). For Δ systolic BP, baseline values for systolic BP, BMI, hs-CRP, and IL-1β and Δ BMI, Δ triglyceride, Δ total cholesterol, Δ glucose, and Δ Lp-PLA 2 activity were defined as the independent variables. For Δ diastolic BP, baseline values for diastolic BP, BMI, hs-CRP, and IL-1β and Δ BMI, Δ triglyceride, Δ total cholesterol, Δ glucose, and Δ Lp-PLA 2 activity were defined as the independent variables. Baseline systolic BP (standardized β � −0.286, p � 0.005) and Δ Lp-PLA 2 activity (standardized β � 0.604, p < 0.001) emerged as independent predictors of Δ systolic BP. e baseline diastolic BP emerged as an independent predictor of Δ diastolic BP (standardized β � −0.461, p � 0.001).

Discussion
In this study, the hypertension group (normal BP at baseline but development of hypertension during follow-up) experienced an 11.7% increase in systolic BP and a 13.6% increase in diastolic BP during the 2-year period. We investigated age-related alterations in inflammatory markers (hs-CRP, Lp-PLA 2 activity, TNF-α, IL-1β, and IL-6) and found that only Δ Lp-PLA 2 activity was independently associated with age-related increases in systolic BP. is result is consistent with a recent report of prehypertension-associated elevation in circulating Lp-PLA 2 activity [15]. e results suggest that age-related increases in systolic BP correlate strongly with elevated Lp-PLA 2 activity.
Prospective studies regarding inflammatory markers and the risk of hypertension have shown contradictory results. For example, hs-CRP has been associated with hypertension risk after adjustment for abdominal obesity in 2 cohorts of middle-aged men and women [16,17]. In contrast, no association was observed between hs-CRP and IL-6 in age-and sex-adjusted analyses of 795 initially normotensive diabetic men and women [18]. Recently, elevated plasma inflammatory markers, including hs-CRP and IL-6, have been reported to be nonsignificantly related to a higher risk of hypertension [10]. In the present study, baseline hs-CRP and IL-1β levels in the hypertension group were significantly higher than those in the control group (normal BP at both baseline and follow-up), whereas baseline IL-6 levels showed an increasing trend. However, Δ hs-CRP, Δ TNF-α, Δ IL-1β, and Δ IL-6 levels were not significantly different between the groups. Indeed, only Δ Lp-PLA 2 activity was substantially higher in the hypertension group than in the control group.
is report is the first prospective study to evaluate the association between Δ Lp-PLA 2 activity and Δ BP. e hypertension group exhibited an 18.8% increase in Lp-PLA 2 activity for the 2-year period, whereas significant alteration in Lp-PLA 2 activity did not occur in the control group. Lp-PLA 2 is primarily generated by macrophages but is also produced by monocytes, T lymphocytes, mast cells, and hepatic cells [19,20]. In the bloodstream, Lp-PLA 2 primarily circulates bound to LDL-cholesterol [20]. e current study also showed a strong positive correlation between Δ LDL-cholesterol levels and Δ Lp-PLA 2 activity in the hypertension group. Elevated Lp-PLA 2 activity has been demonstrated to be associated with endothelial dysfunction, and endothelial dysfunction caused by Lp-PLA 2 activity plays an important role in BP [21,22]. Similarly, the hypertension group in this study exhibited significant increases in both Lp-PLA 2 activity and BP, with a strong positive correlation between Δ Lp-PLA 2 activity and Δ systolic BP.
Increases in Lp-PLA 2 activity in the hypertension group occurred in conjunction with enhanced urinary excretion of 8-epi-PGF 2α , which is a sensitive marker of oxidative stress [11,12]. Lp-PLA 2 , which is bound to LDL-cholesterol, hydrolyzes the ester bond at the sn-2 position and generates International Journal of Hypertension Mean ± SE. tested by logarithmic transformation. p a -values derived from an independent t-test and adjusted for BMI at baseline. p b -values derived from an independent t-test and adjusted for BMI at follow-up. p c -values derived from an independent t-test and adjusted for baseline values at the changed value. * p < 0.05, * * p < 0.01, and * * * p < 0.001 are derived from a paired t-test. . Mean ± SE. tested by logarithmic transformation. † p < 0.05, † † p < 0.01, † † † p < 0.001 derived from an independent t-test between the groups at baseline and follow-up and adjusted for the baseline BMI and follow-up BMI, respectively. * p < 0.05, * * p < 0.01, * * * p < 0.001 derived from a paired t-test in each group. ‡ p < 0.05, ‡ ‡ p < 0.01, ‡ ‡ ‡ p < 0.0011 derived from an independent t-test at the changed value and adjusted for the baseline values.

Lp
atherogenic byproducts (oxidized free fatty acids and lysophosphatidylcholines) [23,24]. Stafforini et al. [23] demonstrated that secreted Lp-PLA 2 produces F 2 -isoprostanes from the phosphatidylcholine at the sn-2 position with high affinity. Additionally, intracellular Lp-PLA 2 , which shows homology with plasma Lp-PLA 2 , has been reported to be involved in the metabolism of esterified 8-epi-PGF 2α [25]. In the current study, a positive relationship between Δ Lp-PLA 2 activity and Δ8-epi-PGF 2α levels was observed for the hypertension group. e age-related increases in Lp-PLA 2 activity in the hypertension group may be, at least partially, explained by the higher level of oxidative stress compared with the controls at the 2-year follow-up.
Additionally, the hypertension group showed higher mean changes in fasting glucose than the controls. However, no correlations were found between Δ BP and Δ fasting  International Journal of Hypertension 5 glucose or between Δ BP and Δ HOMA-IR index. ese results suggest that elevated Lp-PLA 2 activity precedes changes in insulin resistance during a span of elevated systolic BP and occurs in conjunction with progressed oxidative stress. Accordingly, this observation certainly indicates that individuals with newly diagnosed mild hypertension should be given appropriate lifestyle advice, including dietary recommendations to reduce LDL-cholesterol, Lp-PLA 2 activity, and oxidative stress, to reduce the risk of cardiovascular disease. e limitation of this study was its small sample size. Additionally, the generalizability of the observed associations with Lp-PLA 2 activity and hypertension to other populations with different ages, ethnicities, and socioeconomic statuses is unknown. Finally, we could not demonstrate why Lp-PLA 2 activity only showed a correlation with an increase in systolic BP but not in diastolic BP. In addition to our study, several others have reported a positive association between Lp-PLA 2 activity and systolic BP [26][27][28]. To the best of our knowledge, the underlying mechanisms that explain why only systolic BP mostly correlates with Lp-PLA 2 activity have not been elucidated.
erefore, attempts to verify the relationship between systolic BP and Lp-PLA 2 activity and between diastolic BP and Lp-PLA 2 activity are needed in the future.
Despite these limitations, the present study demonstrates that age-related increases in systolic BP correlate strongly with elevated Lp-PLA 2 activity. is suggests that Lp-PLA 2 should be carefully observed to prevent future risk of BP elevation.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.