Traffic-Related Air Pollution, Blood Pressure, and Adaptive Response of Mitochondrial Abundance

Abstract

Background

Exposure to black carbon (BC), a tracer of vehicular-traffic-pollution, is associated with increased blood pressure (BP). Identifying biological factors that attenuate BC effects on BP can inform prevention. We evaluated the role of mitochondrial abundance, an adaptive mechanism compensating for cellular-redox-imbalance, in the BC-BP relationship.

Methods and Results

At one or more visits among 675 older men from the Normative Aging Study (observations=1,252), we assessed daily BP and ambient BC levels from a stationary monitor. To determine blood mitochondrial abundance, we used whole blood to analyze mitochondrial-to-nuclear DNA ratio (mtDNA/nDNA) using quantitative polymerase-chain-reaction. Every standard deviation (SD) increase in 28-day BC moving average (MA) was associated with 1.97 mm Hg (95%CI, 1.23–2.72; P<0.0001) and 3.46 mm Hg (95%CI, 2.06–4.87; P<0.0001) higher diastolic and systolic (SBP) BP, respectively. Positive BC-BP associations existed throughout all time windows. BC MAs (5-day to 28-day) were associated with increased mtDNA/nDNA; every SD increase in 28-day BC MA was associated with 0.12 SD (95%CI, 0.03–0.20; P=0.007) higher mtDNA/nDNA. High mtDNA/nDNA significantly attenuated the BC-SBP association throughout all time windows. The estimated effect of 28-day BC MA on SBP was 1.95-fold larger for individuals at the lowest mtDNA/nDNA quartile midpoint (4.68 mm Hg; 95%CI, 3.03–6.33; P<0.0001), compared to the top quartile midpoint (2.40 mm Hg; 95%CI, 0.81–3.99; P=0.003).

Conclusions

In older adults, short- to moderate-term ambient BC levels were associated with increased BP and blood mitochondrial abundance. Our findings indicate that increased blood mitochondrial abundance is a compensatory response and attenuates the cardiac effects of BC.

The American Heart Association (AHA) has identified particulate matter (PM) pollution as a primary contributor to cardiovascular morbidity and mortality, which is estimated to account for ~3.7 million premature deaths per year worldwide.^{1, 2} To facilitate cost-efficient abatement,^{3, 4} recent studies suggested substantial public health benefits could be realized by implementing control for traffic-related PM pollution.^{5, 6} Black carbon (BC) – a combustion byproduct that serves as a proxy for all traffic-related particles – has been associated with adverse cardiovascular events, particularly due to rapid effects within days after exposure peaks.^{2, 7, 8} In particular, short- to moderate-term (days to weeks) increases in ambient BC levels are associated with subsequent increased blood pressure (BP), one of the primary intermediate outcomes contributing to acute air pollution-related cardiovascular disease (CVD).^{2, 9} Determining the impact of ambient BC on BP among vulnerable populations – such as older adults – is key to developing targeted risk-reduction strategies.

Major efforts have been made in identifying the mechanistic pathways linking traffic-related PM and increased BP over the past few decades. Oxidative stress has emerged as key machinery underlying BC-associated cardiovascular events, due to the pro-inflammatory and pro-oxidative properties of BC.⁵ Systemic oxidative stress – which reflects a disturbance in the redox balance of circulating blood cells – can lead to cellular functional and pathologic changes, and play a crucial role in different types of CVD.^10–12 In response to oxidative stress, cells can normally recover their redox balance by utilizing the endogenous antioxidant system.^{13, 14} However, an intensive intracellular reactive oxygen species (ROS) challenge can overwhelm the antioxidant response and, in compensation, trigger mitochondrial over-production – an important adaptive response following environmental challenge to reduce cellular oxidative stress.^{15, 16}

Cellular mitochondrial mass is controlled through biogenesis and degradation, but external oxidative stressors can up-regulate the transcriptional/replication machinery of the mitochondrial genome, resulting in increased mitochondrial abundance. Mitochondrial abundance can be detected as an increased ratio of mitochondrial DNA to nuclear DNA copy number (mtDNA/nDNA).^{15, 17} Elevated mtDNA/nDNA has been shown in peripheral blood cells following oxidative stress and inflammation.^18–20 However, to date, studies on the role of mitochondrial abundance on pollution-related cardiovascular pathogenesis are absent. Understanding how ambient BC might alter blood mitochondrial abundance and how this alteration might impact BC-elicited cardiovascular effects in an aging population can aid the development of preventative strategies.

In this study, we hypothesized that blood mitochondrial abundance is an adaptive response following ambient BC exposure, which might buffer the impact of ambient BC on BP. We utilized the Normative Aging Study to investigate the relationship of ambient BC levels at multiple time windows (2-day to 28-day) with BP. We then examined the association of short- to moderate-term ambient BC levels with blood mitochondrial abundance, and explored its role as an effect modifier for the relationship between ambient BC and BP. In addition, we explored the correlation between blood mitochondrial abundance and plasma inflammatory markers including interleukin 6 (IL6), IL8, IL1β, tumor necrosis factor alpha (TNFα), TNFγ, C-reactive protein (CRP), intercellular adhesion molecule-1 (ICAM-1), and vascular endothelial growth factor (VEGF).

METHODS

Study Population

The Normative Aging Study is a prospective cohort of older men established in Eastern Massachusetts by the United States Veterans Administration. The present study included 675 community-dwelling men with available data on ambient BC (at any of the time windows), BP (systolic or diastolic), and blood mitochondrial abundance (Table 1). Participants were recalled for visits every three to five years. We considered visits from April 1999 (i.e., the earliest date with available mitochondrial DNA data) to December 2012, for a total of 1,252 visits (1–4 visits per participant; average 1.9 visits). The study was approved by the institutional review boards of all participating institutions. All subjects provided informed consent.

Table 1.

Baseline characteristics, black carbon (BC) levels, blood mitochondrial DNA to nuclear DNA copy number ratio (mtDNA/nDNA), and blood pressure (BP) in the Normative Aging Study (N=675), 1999–2012.

Characteristic	N (%)	BC (μg/m3) Mean (SD)	SBP (mm Hg) Mean (SD)	DBP (mm Hg) Mean (SD)	mtDNA/nDNA* Mean (SD)
Age
55–69 years	215 (31.8)	1.29 (0.40)	129.8 (14.9)	79.2 (8.3)	1.04 (0.24)
70–79 years	340 (50.4)	1.23 (0.41)	131.0 (17.7)	74.8 (9.6)	1.00 (0.24)
80–89 years	117 (17.3)	1.17 (0.42)	132.3 (20.8)	72.4 (10.0)	1.01 (0.34)
> 90 years	3 (0.4)	0.74 (--)	122.7 (15.3)	77.0 (6.6)	1.08 (0.09)
Physical Activity
<12 MET-hrs/wk	434 (64.3)	1.25 (0.41)	131.1 (17.7)	75.7 (9.6)	1.01 (0.27)
12–29 MET-hrs/wk	146 (21.6)	1.22 (0.41)	127.8 (15.8)	75.6 (10.1)	1.01 (0.22)
≥30 MET-hrs/wk	95 (14.1)	1.21 (0.42)	134.2 (18.2)	76.5 (9.0)	1.02 (0.29)
Alcohol Use
<2 drinks/day	537 (79.6)	1.24 (0.41)	130.3 (17.4)	76.1 (9.7)	1.02 (0.26)
≥2 drinks/day	138 (20.4)	1.20 (0.42)	132.7 (17.5)	74.8 (9.3)	1.00 (0.25)
Diabetes
No	549 (81.3)	1.25 (0.41)	130.1 (17.1)	76.2 (9.5)	1.01 (0.27)
Yes	126 (18.7)	1.19 (0.41)	133.9 (18.3)	74.3 (10.0)	1.01 (0.22)
Race
Non-white	18 (2.7)	1.32 (0.47)	128.7 (13.5)	75.4 (9.5)	0.65 (0.44)
White	657 (97.3)	1.23 (0.41)	130.8 (17.5)	75.8 (9.6)	1.02 (0.25)
BMI
<25 kg/m2	130 (19.3)	1.26 (0.41)	130.6 (19.0)	73.9 (9.0)	1.04 (0.32)
≥25 kg/m2	545 (80.7)	1.23 (0.41)	130.8 (17.1)	76.3 (9.7)	1.01 (0.24)
Smoking Status
Never	199 (29.4)	1.24 (0.41)	130.7 (17.4)	76.5 (9.4)	1.02 (0.24)
Current	29 (4.3)	1.21 (0.41)	132.2 (15.7)	76.0 (6.6)	1.02 (0.31)
Former	447 (66.2)	1.24 (0.41)	130.7 (17.6)	75.5 (9.9)	1.01 (0.27)
Education
≤High School	177 (26.2)	1.26 (0.40)	131.3 (18.4)	75.5 (9.8)	1.00 (0.23)
College	337 (49.9)	1.23 (0.41)	130.6 (17.5)	75.7 (9.6)	1.01 (0.28)
Graduate School	161 (23.9)	1.21 (0.43)	130.5 (16.1)	76.4 (9.4)	1.04 (0.25)

N indicates the number of participants; SBP and DBP indicate systolic and diastolic BP, respectively; SD indicates standard deviation; MET indicates metabolic equivalent of task; BMI indicates body mass index.

^*mtDNA/nDNA measurements reflects the average cellular mitochondrial abundance, calculated based on the ratio of copy number estimates of a mitochondrial gene to those of a nuclear gene. The ratio was scaled to a standard DNA sample presenting the relative mitochondrial content within a cell.

Blood Pressure

A physician measured the participant’s resting seated systolic BP (SBP) and diastolic BP (DBP) in the morning at each visit using a standard mercury sphygmomanometer, following a AHA-recommended protocol, as previously described.²¹ The means of the right and left arm measurements were used. Three readings were taken and the average of the second and third readings was used for statistical analysis.

Blood Mitochondrial Abundance and Plasma Inflammatory Markers

We measured blood mitochondrial abundance through mtDNA/nDNA, a widely used biomarker representing the mitochondrial DNA copy number versus the nuclear DNA copy number.^{15, 17} At every visit, mtDNA copy number was analyzed on whole blood samples, collected after overnight fasting. We adapted a multiplex quantitative real-time polymerase chain reaction (RT-PCR) method with minor modifications.²² To measure mtDNA copy number, we used the mtDNA 12S ribosomal ribonucleic acid (RNA) TaqMan (Applied Biosystems, Waltham, Massachusetts) probe (6FAM-5′ TGCCAGCCACCGCG 3′-MGB). The sequences of primers used for amplification in mtDNA were mtF805 (5′CCACGGGAAACAGCAGTGATT3′) and mtR927 (5′CTATTGACTTGGGTTAATCGTGTGA3′). The quantity of mtDNA was corrected by simultaneous measurement of a single copy nuclear Ribonuclease P gene. We used a commercial kit to quantify nuclear DNA (nDNA) (TaqMan® RNase P Control Reagents Kit, Applied Biosystems). RT-PCR assays were performed following a published protocol,²² using Bio-Rad CFX384 Touch^™ Real-Time PCR Detection System (Bio-Rad, Hercules, California). All samples were run in triplicate. The mean of the three measurements was used for statistical analysis. The within-run and between-run coefficients of variation of this assay were 3.35% and 3.26%, respectively. A laboratory reference DNA sample – which was a pool of 300 test samples (20 μL taken from each sample, final concentration: 40 ng/μL) – was used to construct standard curves (mtDNA and nDNA R² ≥ 0.99). The standard curves were used to quantify mtDNA and nDNA copy numbers to standardize the mtDNA/nDNA obtained from all test samples in all reactions.²³ We used mtDNA/nDNA in the statistical analysis. A ratio value of 1 indicates that the mtDNA/nDNA of the test sample is equal to the mtDNA/nDNA in the reference DNA pool used in the assay.

Plasma inflammatory markers (IL6, IL8, IL1β, TNFα, TNFγ, CRP, ICAM-1, and VEGF) were measured using previously described methods.^{24, 25}

Air Pollution and Weather Data

Daily ambient BC and PM_2.5 levels were measured at a Harvard School of Public Health monitoring site (Boston, MA), 1 km from the clinical examination site, using previously described methods.²¹ We evaluated the average pollutant concentrations during different time windows (2-day, 5-day, 7-day, 14-day, 21-day, and 28-day, before the time of BP measurement and blood draw). We assigned exposure index to participants based on the visit date. Outdoor air temperature, relative humidity, dew point temperature, and wind speed were obtained from the Boston Logan airport weather station. The corresponding moving averages of outdoor apparent temperature and absolute humidity were calculated using previous reported methods.^{26, 27}

Statistical Methods

We adjusted for potential confounders, selected based on literature evidence, i.e., age (continuous), race (white/non-white), outdoor apparent temperature (continuous), absolute humidity (continuous), season (winter/spring-fall/summer), highest education (≤high school/college/graduate school), weekday of the visit (Monday/Tuesday/Wednesday/Thursday/Friday), and the visit date (continuous). To increase efficiency, we adjusted for major risk factors for increased BP: Body Mass Index (BMI) (continuous), smoking status (in pack years, continuous), physical activity (metabolic equivalent of task-hours per week, continuous), alcohol use (<2 drinks/day/≥2 drinks/day), CRP>10 mg/L (yes/no); use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitors (yes/no). Due to high cell-type specificity of mitochondria contents, mtDNA/nDNA in whole blood might be affected by differences in the proportions of cell types. We adjusted for the percentage of major cell types containing mitochondria (i.e., lymphocytes, neutrophils, and platelets).

Linear relationships were examined using graphical diagnostics and spline regression models, no deviation from linearity was observed. We utilized the likelihood ratio test to determine whether it is adequate to model the independent variables or the effect modifier with a linear trend.

In our dataset, BC moving average at different time windows exhibited different variability. To compare the magnitude of BC-BP associations across multiple time windows, we presented the change in BP and mtDNA/nDNA per standard deviation (SD) increase in ambient BC levels.

To account for repeated assessments for many participants, we utilized linear mixed effect model with random intercept (compound symmetry covariance structure) to examine the association between ambient BC levels and BP, as well as the association between ambient BC levels and mtDNA/nDNA:

$${\mathrm{Y}}_{\text{ij}}={\mathrm{\beta }}_0+{\mathrm{\beta }}_1{\text{BC}}_{\text{ij}}+{\mathrm{\beta }}_2{\mathrm{X}}_{2\text{ij}}+\dots +{\mathrm{\beta }}_{\mathrm{p}}{\mathrm{X}}_{\text{pij}}+{\mathrm{b}}_{\mathrm{i}}+{\mathrm{\epsilon }}_{\text{ij}}$$ (Model 1)

Where Y_ij was the BP or mtDNA/nDNA for participant i at visit j, β₀ was the overall intercept, b_i was the separate random intercept for subject i, and b_i ~ N(0, θ),ε_ij ~ N(0, σ²). X_2ij–X_pij were the covariates including confounders and risk factors for increased BP, for participant i at visit j. We also examined effect modification by mtDNA/nDNA by including additionally the main effect of mtDNA/nDNA and an interaction term between mtDNA/nDNA and BC. All analyses were performed using SAS 9.4 (SAS Institute, Cary, NC).

RESULTS

Cohort Characteristics and Ambient BC Levels

The NAS is a cohort of aging men, between 55 and 100 years old at the first visit of the present study. We acquired data on ambient BC levels, BP, and mtDNA/nDNA among 675 participants. The baseline characteristics of the participants included in the final analysis are given in Table 1. All participants were male, and 97.3% were white. Eighty-one percent of the participants were overweight and 18.7% were diabetics. The study population included 4.3% current smokers and 20.4% heavy alcohol drinkers (≥ 2 drinks/day) (Table 1). During the study period (April 1999 – December 2012), the 7-day BC moving average varied between 0.20 μg/m³ to 2.18 μg/m³, with an average level of 1.02 μg/m³. The baseline average SBP and DBP was 131 mm Hg and 76 mm Hg, respectively. Participants included or excluded from each analysis are similar in key baseline characteristics (age, physical activity, alcohol use, diabetes, BMI, smoking status, SBP, and DBP) (data not shown).

Ambient BC Level and BP

Short- to moderate-term ambient BC levels were consistently associated with significantly increased SBP and DBP (Figure 1). Every SD increase in the 2-day, 5-day, 7-day, 14-day, 21-day, and 28-day BC moving average was associated with 1.79 mm Hg (95% confidence interval (CI), 0.65, 2.92; P=0.002), 2.35 mm Hg (95% CI, 1.14, 3.55; P=0.0001), 2.83 mm Hg (95% CI, 1.57, 4.09; P<0.0001), 3.02 mm Hg (95% CI, 1.68, 4.35; P<0.0001); 3.10 mm Hg (95% CI, 1.71, 4.49; P<0.0001); and 3.46 mm Hg (95% CI, 2.06, 4.87; P<0.0001) increase in SBP, respectively. Likewise, every SD increase in the 2-day, 5-day, 7-day, 14-day, 21-day, and 28-day BC moving average was associated with 0.65 mm Hg (95% CI, 0.05, 1.24; P=0.03), 1.40 mm Hg (95% CI, 0.77, 2.03; P<0.0001), 1.73 mm Hg (95% CI, 1.07, 2.39; P<0.0001), 1.81 mm Hg (95% CI, 1.11, 2.52; P<0.0001); 1.91 mm Hg (95% CI, 1.17, 2.65; P<0.0001); and 1.97 mm Hg (95% CI, 1.23, 2.72; P<0.0001) increase in DBP, respectively (Table 2).

Figure 1.

Association of ambient black carbon (BC) levels with blood pressure (BP), Normative Aging Study, 1999–2012. Estimates represent the increase in BP per standard deviation increase in BC. Results were adjusted for: apparent temperature and absolute humidity at the matching time window; race; percentage of major cell types (lymphocytes, neutrophils, and platelets); age; weekday; date of visit; body mass index; C-reactive protein>10 mg/L; smoking; alcohol use; season; physical activity; education level; use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitor.

Table 2.

Association of ambient black carbon (BC) levels with blood pressure (BP).

	Time Window	Estimate*	95% CI	P	N	n
Original Model†
SBP (mm Hg)	2-day	1.79	0.65 to 2.92	0.002	656	1243
	5-day	2.35	1.14 to 3.55	0.0001	657	1252
	7-day	2.83	1.57 to 4.09	<0.0001	657	1252
	14-day	3.02	1.68 to 4.35	<0.0001	654	1234
	21-day	3.10	1.71 to 4.49	<0.0001	649	1206
	28-day	3.46	2.06 to 4.87	<0.0001	650	1211
DBP (mm Hg)	2-day	0.65	0.05 to 1.24	0.03	656	1243
	5-day	1.40	0.77 to 2.03	<0.0001	657	1252
	7-day	1.73	1.07 to 2.39	<0.0001	657	1252
	14-day	1.81	1.11 to 2.52	<0.0001	654	1234
	21-day	1.91	1.17 to 2.65	<0.0001	649	1206
	28-day	1.97	1.23 to 2.72	<0.0001	650	1211
Sensitivity Analysis Adjusting for PM2.5‡
SBP (mm Hg)	2-day	3.72	2.27 to 5.16	<0.0001	606	1048
	5-day	4.03	2.58 to 5.47	<0.0001	610	1058
	7-day	4.29	2.85 to 5.73	<0.0001	610	1052
	14-day	4.36	2.90 to 5.82	<0.0001	600	1021
	21-day	4.19	2.68 to 5.70	<0.0001	585	984
	28-day	4.32	2.78 to 5.86	<0.0001	583	985
DBP (mm Hg)	2-day	1.64	0.89 to 2.38	<0.0001	606	1048
	5-day	2.40	1.66 to 3.14	<0.0001	610	1058
	7-day	2.59	1.85 to 3.33	<0.0001	610	1052
	14-day	2.61	1.86 to 3.36	<0.0001	600	1021
	21-day	2.54	1.77 to 3.31	<0.0001	585	984
	28-day	2.49	1.71 to 3.26	<0.0001	583	985

SBP and DBP indicate systolic and diastolic BP, respectively; N and n indicate the number of participants and observations, respectively; PM_2.5 indicates particulate matter up to 2.5 micrometers in size.

^*Estimate presents the increase in BP (mm Hg) per standard deviation (SD) increase in BC.

^†Results were adjusted for: apparent temperature and absolute humidity at the matching time window; race; percentage of major cell types (lymphocytes, neutrophils, and platelets); age; weekday; date of visit; body mass index; C-reactive protein>10 mg/L; smoking; alcohol use; season; physical activity; education level; use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitor.

^‡Results were adjusted for all the covariates listed above, and PM_2.5 at the matching time window.

Ambient BC Level and Mitochondrial Abundance

The average ambient BC levels over the preceding time windows (2-day, 5-day, 7-day, 14-day, 21-day, and 28-day) were consistently positively associated with blood mitochondrial abundance (Figure 2). Every SD increase in the 2-day, 5-day, 7-day, 14-day, 21-day, and 28-day BC moving average was associated with 0.05 SD (95% CI, −0.01, 0.12; P=0.11), 0.12 SD (95% CI, 0.06, 0.19; P=0.0003), 0.14 SD (95% CI, 0.07, 0.21; P=0.0002), 0.15 SD (95% CI, 0.07, 0.23; P=0.0002), 0.14 SD (95% CI, 0.06, 0.22; P=0.001), and 0.12 SD (95% CI, 0.03, 0.20; P=0.007) increase in blood mtDNA/nDNA, respectively (Table 3).

Figure 2.

Association of ambient black carbon (BC) levels with blood mitochondrial DNA to nuclear DNA copy number ratio (mtDNA/nDNA), Normative Aging Study, 1999–2012. Estimates represent the increase in standard deviation (SD) for mtDNA/nDNA per SD increase in BC. Results were adjusted for: apparent temperature and absolute humidity at the matching time window; race; percentage of major cell types (lymphocytes, neutrophils, and platelets); age; weekday; date of visit; body mass index; C-reactive protein>10 mg/L; smoking; alcohol use; season; physical activity; education level; use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitor.

Table 3.

Association of ambient black carbon (BC) levels with blood mitochondrial DNA to nuclear DNA copy number ratio (mtDNA/nDNA).

Time Window	Estimate*	95% CI	P	N	n
Original Model†
2-day	0.05	−0.01 to 0.12	0.11	656	1241
5-day	0.12	0.06 to 0.19	0.0003	657	1250
7-day	0.14	0.07 to 0.21	0.0002	657	1250
14-day	0.15	0.07 to 0.23	0.0002	654	1232
21-day	0.14	0.06 to 0.22	0.001	649	1204
28-day	0.12	0.03 to 0.20	0.007	650	1209
Sensitivity Analysis Adjusting for PM2.5‡
2-day	0.12	0.04 to 0.20	0.005	606	1046
5-day	0.17	0.09 to 0.25	<0.0001	610	1056
7-day	0.15	0.07 to 0.23	0.0002	610	1050
14-day	0.14	0.06 to 0.23	0.001	600	1019
21-day	0.12	0.03 to 0.21	0.008	585	982
28-day	0.09	0.00 to 0.17	0.05	583	983

N and n indicate the number of participants and observations, respectively; PM_2.5 indicates particulate matter up to 2.5 micrometers in size.

^*Estimate presents the increase in standard deviation (SD) for mtDNA/nDNA per SD increase in BC.

^†Results were adjusted for: apparent temperature and absolute humidity at the matching time window; race; percentage of major cell types (lymphocytes, neutrophils, and platelets); age; weekday; date of visit; body mass index; C-reactive protein>10 mg/L; smoking; alcohol use; season; physical activity; education level; use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitor.

^‡Results were adjusted for all the covariates listed above and PM_2.5 at the matching time window.

BC and BP Association: Effect Modification by Blood Mitochondrial Abundance

The association of ambient BC levels with SBP was modified by blood mtDNA/nDNA across multiple time windows (P=0.05; P=0.04; P=0.03; P=0.01; P=0.01, for 5-day, 7-day, 14-day, 21-day, and 28-day, respectively) (Table 4). Ambient BC levels had a relatively smaller impact on SBP in individuals with higher blood mitochondrial abundance, compared to those with lower blood mitochondrial abundance (Figure 3 and Supplemental Figure 1). For example, for individuals whose mtDNA/nDNA was at the midpoint of the lowest quartile, the 28-day BC moving average was estimated to be associated with 4.68 mm Hg (95%CI, 3.03, 6.33; P<0.0001) increase in SBP. In comparison, the 28-day BC moving average was estimated to be associated with 2.40 mm Hg (95%CI, 0.81, 3.99; P=0.003) increase in SBP among individuals whose mtDNA/nDNA was at the midpoint of the top quartile. We did not observe significant effect modification by blood mitochondrial abundance on 2-day BC-SBP association, and on the association between short-term ambient BC levels and DBP (Table 4). In addition, increased blood mtDNA/nDNA was negatively correlated with plasma IL6, TNFγ, CRP, and ICAM-1 levels (P=0.03, P=0.01, P=0.003, P=0.03, respectively) (Supplemental Table 1).

Table 4.

Effect modification by blood mitochondrial DNA to nuclear DNA copy number ratio (mtDNA/nDNA), on the association of ambient black carbon (BC) with blood pressure (BP).

		SBP (mm Hg)			DBP (mm Hg)
	mtDNA/nDNA	Estimate	95% CI	P	Estimate	95% CI	P
2-day	Q1 Midpoint	2.36	0.99 to 3.73	0.001	0.80	0.08 to 1.51	0.03
	Q2 Midpoint	1.98	0.82 to 3.15	0.001	0.67	0.06 to 1.28	0.03
	Q3 Midpoint	1.65	0.49 to 2.80	0.01	0.56	−0.05 to 1.17	0.07
	Q4 Midpoint	1.09	−0.38 to 2.56	0.15	0.38	−0.40 to 1.15	0.34
	Pinteraction			0.14			0.35
5-day	Q1 Midpoint	3.19	1.71 to 4.66	<0.0001	1.50	0.73 to 2.27	0.0001
	Q2 Midpoint	2.69	1.43 to 3.95	<0.0001	1.44	0.79 to 2.10	<0.0001
	Q3 Midpoint	2.24	1.02 to 3.46	0.0003	1.39	0.76 to 2.03	<0.0001
	Q4 Midpoint	1.51	0.04 to 2.99	0.04	1.31	0.54 to 2.08	0.001
	Pinteraction			0.05			0.67
7-day	Q1 Midpoint	3.76	2.23 to 5.3	<0.0001	1.93	1.13 to 2.73	<0.0001
	Q2 Midpoint	3.25	1.92 to 4.57	<0.0001	1.82	1.13 to 2.51	<0.0001
	Q3 Midpoint	2.79	1.51 to 4.06	<0.0001	1.72	1.06 to 2.39	<0.0001
	Q4 Midpoint	2.03	0.55 to 3.51	0.01	1.56	0.79 to 2.33	<0.0001
	Pinteraction			0.04			0.39
14-day	Q1 Midpoint	4.01	2.39 to 5.62	<0.0001	2.00	1.15 to 2.85	<0.0001
	Q2 Midpoint	3.48	2.07 to 4.89	<0.0001	1.89	1.15 to 2.63	<0.0001
	Q3 Midpoint	3.02	1.68 to 4.37	<0.0001	1.80	1.09 to 2.50	<0.0001
	Q4 Midpoint	2.26	0.74 to 3.78	0.004	1.64	0.84 to 2.44	<0.0001
	Pinteraction			0.03			0.40
21-day	Q1 Midpoint	4.22	2.56 to 5.88	<0.0001	2.17	1.29 to 3.05	<0.0001
	Q2 Midpoint	3.62	2.15 to 5.08	<0.0001	2.05	1.27 to 2.82	<0.0001
	Q3 Midpoint	3.08	1.68 to 4.48	<0.0001	1.94	1.20 to 2.68	<0.0001
	Q4 Midpoint	2.19	0.62 to 3.76	0.01	1.76	0.92 to 2.59	<0.0001
	Pinteraction			0.01			0.33
28-day	Q1 Midpoint	4.68	3.03 to 6.33	<0.0001	2.21	1.34 to 3.09	<0.0001
	Q2 Midpoint	4.00	2.54 to 5.46	<0.0001	2.10	1.33 to 2.87	<0.0001
	Q3 Midpoint	3.40	1.99 to 4.80	<0.0001	1.99	1.25 to 2.74	<0.0001
	Q4 Midpoint	2.40	0.81 to 3.99	0.003	1.82	0.98 to 2.66	<0.0001
	Pinteraction			0.01			0.36

N and n indicate the number of participants and observations, respectively; SBP and DBP indicate systolic and diastolic BP, respectively.

Estimate represents that per standard deviation (SD) increase in mtDNA/nDNA, the lessened increase in BP (mm Hg) per SD increase in BC.

Results were adjusted for: apparent temperature and absolute humidity at the matching time window; race; percentage of major cell types (lymphocytes, neutrophils, and platelets); age; weekday; date of visit; body mass index; C-reactive protein>10 mg/L; smoking; alcohol use; season; physical activity; education level; use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitor.

Figure 3.

Association of ambient black carbon (BC) levels with systolic blood pressure (SBP) at different blood mitochondrial DNA to nuclear DNA copy number ratio (mtDNA/nDNA) levels, Normative Aging Study, 1999–2012. Results were adjusted for: apparent temperature and absolute humidity at the matching time window; race; percentage of major cell types (lymphocytes, neutrophils, and platelets); age; weekday; date of visit; body mass index; C-reactive protein>10 mg/L; smoking; alcohol use; season; physical activity; education level; use of calcium channel blockers, β-blockers, and angiotensin-converting enzyme inhibitor.

Sensitivity Analyses

BC is the major component of fine particles (PM_2.5); PM_2.5 itself is a contributor to increased BP and enhanced oxidative damage. In sensitivity analysis, we adjusted for the ambient PM_2.5 levels in the matching time window – to rule out the possibility that the observed effect was partially due to confounding by ambient PM_2.5. This adjustment did not affect our conclusion (Tables 2 and 3). In addition, we adjusted for the platelet/lymphocyte and platelet/neutrophil ratios to minimize the influence from disproportional increase in platelets, and our conclusions remained the same (data not shown).

DISCUSSION

This study on a cohort of aging male Boston-area residents confirmed that short- to moderate-term ambient BC levels are associated with increased BP. We showed that short- to moderate-term ambient BC levels also increased peripheral blood mitochondrial abundance in older adults. Further, individuals with higher blood mitochondrial abundance appeared less susceptible to the impact of ambient BC levels on BP, compared to those with less blood mitochondrial abundance.

In the United States, the proportion of the population over the age of 65 is projected to increase from 12.7% in 2000 to 20.3% in 2050.²⁸ The AHA has emphasized that air pollution is an ubiquitous public health threat, particularly for older people, who are especially susceptible to air pollution-triggered CVD.² BC, the byproduct of incomplete combustion of fossil fuels, biofuels, and biomass, is the most effective form of PM (by mass) at absorbing solar energy and is considered a potent trigger for systemic oxidative stress upon exposure.²⁹ Increased BP related to ambient BC levels is an alarming physiological change that may represent intermediate mechanisms contributing to acute adverse cardiovascular events.^{2, 9}

Consistent with our previous study, we confirmed the significant association between BP and the average ambient BC level in the past 7 days.²¹ The present study presented a comprehensive analysis of the short- to moderate-term cardiovascular effects of BC, ranging from 2 days to 28 days. The estimated effect size of ambient BC levels increased gradually as the time window of exposure increased, and the strongest estimated effect was observed for the 28-day time window. The BP increase associated with ambient BC levels reflects the physiologic effect of traffic-related air pollution that may contribute to the pathophysiological changes in the cardiovascular system – which could be a potential mechanism linking air pollution and cardiac morbidity and mortality in the aging population. In addition, the observed effect estimates for ambient BC were increased with a larger magnitude after adjusting for ambient PM_2.5 levels, suggesting that the proportion of BC content in PM_2.5 has a stronger effect on BP increases than the actual BC mass.

Recent studies suggest that particle-induced oxidative stress initiates systemic inflammatory response,³⁰ which is considered a major mechanism underlying the cardiac responses to air pollution. Mitochondria, the specialized cellular organelles that play a pivotal role in energy production, are the main immediate target of ROS.^{11, 12} The role of mitochondrial biogenesis and mitochondria DNA copy number maintenance in the determination of cell survival and function under oxidative stress has been increasingly appreciated in recent years. However, most studies have focused on inflammatory markers such as cytokines, and the role of mitochondria in air pollution-related cardiovascular pathogenesis remains a major gap in current knowledge. The amount of mitochondrial genomic content within a cell is normally controlled,³¹ but upon exposure to environmental stress, mitochondria biogenesis can be enhanced to increase energy supply for eliminating damage to cellular components.^{12, 15} Malik and Czajka suggested that mtDNA/nDNA increases when the cell’s endogenous antioxidant response fails to recover its redox balance.¹⁵ For example, exposure to air pollutants such as benzene was associated with increases in blood mitochondrial genomic contents.³² Systemic oxidative stress stimulated by the activation of PI3K/Akt signaling pathway could also result in increased mitochondrial DNA copy number in leukocytes.²⁰ Mitochondria proliferation and mitochondrial DNA amplification under oxidative stress was observed even when overall cell division was under arrest.³³ The present study for the first time provides epidemiological evidence showing that ambient BC level is linked with increased blood mitochondrial abundance, as reflected in elevated mtDNA/nDNA in peripheral blood.

In addition, we demonstrated that high blood mitochondrial abundance attenuates the association of ambient BC levels with BP. In in vivo studies, increased mitochondrial biogenesis induced by over-expressing recombinant human mitochondrial transcription factor A resulted in a 50% reduction in the oxidative damage to proteins following lipopolysaccharide endotoxin challenge.¹⁶ Under the challenge of endogenous or exogenous ROS, increased mitochondrial DNA copy number in aging tissues is a feedback response that compensates for insufficient ROS cleavage and upholds the energy metabolism to rescue the cell.^{12, 34} In our study, increased blood mitochondria abundance was linked with reduced inflammatory responses. Collectively, evidence support our hypothesis that blood mitochondrial abundance is an adaptive process following ambient BC exposure and reduces susceptibility to the cardiovascular effect of BC, possibly by scavenging radical and suppressing inflammation. Further elucidation of the changes in blood mitochondrial abundance in the process of stress response to render cardiovascular tissue’s survival under oxidative stress is of prime importance for a better understanding of cardio-pathogenesis related to air pollution.

Nevertheless, the stress response of mitochondria to oxidative stress may be a double-edged sword. On one hand, the compensatory increase in blood mitochondrial proliferation supplies energy to meet the need for blood cell survival.¹² On the other hand, mitochondria are the main intracellular source of ROS and play an important role in biological oxidation.^{11, 12} Excess ROS generated from the increased mitochondria can be pathogenic, initiating a vicious cycle that leads to chronic oxidative damage and inflammation.^{11, 12, 15, 35, 36} Unfortunately, we only measured blood mtDNA/nDNA and BP concurrently at each visit; therefore, we are not able to assess the effect of continuously higher blood mitochondrial abundance on BP due to lack of temporality. Future studies are warranted to shed light on the cardiovascular effect of chronically elevated blood mitochondrial abundance due to traffic-related pollution.

This study has several strengths, including its repeated-measure design and relatively large sample size. In addition, we utilized a highly reproducible RT-PCR method to quantify the mtDNA content from easy-to-obtain blood samples, which has been widely adapted in epidemiological studies.³⁷ To measure ambient traffic-related PM, we used daily BC monitor data as a proxy, because variation in short- to moderate-term ambient air pollution was mostly due to temporal variation rather than spatial variation.³⁸ Measurement error, which is inherent in air pollution epidemiological studies, cannot be completely avoided due to lack of personal exposure assessment. However, misclassification is expected to be non-differential and bias towards the null, because it was unlikely to be associated with participants’ BP status or mitochondrial abundance. To limit confounding, we included in regression models an extensive list of covariates. We conducted further analysis to evaluate the sensitivity of our results to covariate specification, and our results were stable and robust. In addition, all BP measures were conducted at the same time of the day to eliminate confounding due to diurnal variation. While residual confounding due to unmeasured variables is possible, chances that the observed association and effect modification reflected bias resulting from residual confounding are minimized.

The present study has several limitations. Some of the eligible participants were excluded from analysis due to missing BC data. They were similar in baseline characteristics (including BP) with participants included in the analysis. Moreover, it is reasonable to assume that the missingness was independent of ambient BC levels and participants’ mitochondrial abundance based on study operation. Therefore, selection bias due to informative missingness was unlikely. In addition, we standardized mtDNA/nDNA measurements to a lab reference DNA sample; therefore, the absolute value of mtDNA/nDNA from the present study might not be directly comparable to other studies. Finally, our findings are limited to male older individuals who were residing in a lightly-polluted urban area. Our conclusion might not be generalizable to young adults, females, or people living in other areas due to differential environmental and physiological factors.

In summary, our results showed that, in older adults, levels of short- to moderate-term ambient traffic-related PM were associated with increases in BP and blood mitochondrial abundance. Individuals with higher blood mitochondrial abundance showed significantly weaker association of ambient BC with BP. Our study provided a novel mechanistic perspective – the compensatory increase in blood mitochondrial abundance following exposure to ambient BC might be an adaptive response that attenuates individual susceptibility to post-exposure BP increase in older people. Further investigation of the mitochondria-mediated oxidative responses is warranted to provide insights into the physiologic effects of traffic-related pollution and may help guide targeted prevention and novel pharmaceutical interventions.

Clinical Perspective.

Traffic-related pollution (black carbon, BC) is a pervasive environmental threat, particularly for older individuals, who are especially susceptible to pollution-triggered cardiovascular morbidity and mortality. Systemic oxidative stress is a key mechanism linking BC pollution and cardiovascular events, and the role of mitochondria abundance in compensating for cellular-redox-imbalance under oxidative stress has been increasingly appreciated. However, the response of mitochondria in the peripheral blood upon ambient BC exposure remains unknown. We utilized the Normative Aging Study to address a mechanistic question about cardiovascular effects of air pollution: how does short- to moderate-term ambient BC alter blood mitochondrial abundance; and how this alteration might impact BC-elicited cardiovascular effects. Our results indicated that, in older adults, levels of short- to moderate-term ambient BC were associated with increased blood pressure (BP) and blood mitochondrial abundance. In addition, individuals with lower blood mitochondrial abundance were estimated to exhibit 1.95-fold larger effect of BC with systolic BP, suggesting that increased blood mitochondrial abundance buffered the effect of ambient BC on BP. The impact of our study is in the investigation of compensatory increase in blood mitochondrial abundance as an adaptive response to reduce the cardiovascular effects of ambient BC. The present study provides insights into the physiologic effects of traffic-related pollution and may aid targeted risk-reduction and novel pharmaceutical interventions. Because of the central role of oxidative stress in the pathogenesis of cardiovascular diseases, the study results may provide novel perspective to the cardiovascular effect of other environmental risk factors.