Episodic Exposure to Fine Particulate Air Pollution Decreases Circulating Levels of Endothelial Progenitor Cells

Abstract

Rationale

Acute and chronic exposures to airborne particulate matter (PM) have been linked in epidemiological studies to a wide spectrum of cardiovascular disorders that are characterized by a dysfunctional endothelium. The pathophysiological mechanisms underlying these associations are unclear.

Objective

To examine whether exposure to fine PM with an aerodynamic diameter <2.5 μm (PM_2.5) affects the circulating levels of endothelial progenitors cell (EPC) populations, systemic inflammation and coagulation.

Methods and Results

Phenotypically-distinct EPC populations were quantified by flow cytometry in young (18-25 years), adult humans exposed to episodic increases in PM_2.5 along the Wasatch Mountain Front in Utah. In addition, Sca-1⁺/Flk-1⁺ cells were measured in the peripheral blood of mice exposed to concentrated particles from ambient air in Louisville, Kentucky. In both studies, PM exposure was negatively correlated with circulating EPC levels. In humans, statistically significant associations between PM_2.5 exposure and the plasma levels of platelet-monocyte aggregates, HDL, and non-albumin protein were also observed. Episodic increases in PM_2.5 did not change plasma levels of C-reactive protein, IL-1β, IL-6, fibrinogen or serum amyloid A.

Conclusions

Episodic exposure to PM_2.5 induces reversible vascular injury, reflected in part by depletion of circulating EPC levels, and increases in platelet activation and the plasma level of HDL. These changes were also accompanied by an increase in non-albumin protein and may be related to mechanisms by which exposure to particulate air pollution increases the risk of cardiovascular disease and adverse cardiovascular events.

Acute and chronic exposure to elevated levels of fine airborne particulate matter (PM) is associated with an increase in the incidence of adverse cardiovascular events^1-2, atherogenesis, cardiovascular disease (CVD) risk and cardiovascular mortality. In urban environments, fine PM (PM with aerodynamic diameter < 2.5 μm; PM_2.5) is generated mostly by fossil fuel combustion in automobiles or by industrial processes. Although several mechanisms have been proposed to account for the link between PM exposure and CVD risk, endothelial dysfunction has emerged as a key feature of PM toxicity. Inhalation of concentrated PM_2.5 induces acute conduit artery vasoconstriction in humans and chronic deficits in endothelium-mediated vasodilation in mice^1-2.

The adult endothelium is a differentiated cell layer that provides a non-thrombotic interface between parenchymal cells and peripheral blood. Defects in its function arise due to the upregulated expression of pro-inflammatory and pro-thrombotic molecules or from defective, endogenous repair capacity. Evidence from multiple studies suggests that the endothelium is continually repaired by progenitors cells mobilized from specific niches such as the bone marrow. These cells express both endothelial and stem cell markers and their circulating levels in blood are reflective of CVD risk and burden^3-4. The current study was designed to examine how exposure to PM_2.5 affects endothelial progenitor cell (EPC) populations and whether this is associated with changes in systemic inflammation, coagulation or plasma lipids.

METHODS

For the human study, 16 (8 male and 8 female), young (18-25 years of age), non-smoking, healthy (no existing acute or chronic disease) adults of normal weight (BMI = 19-25) with no reported exposure to second-hand smoke were recruited in Provo, Utah. In the Utah Valley of the Wasatch Front, winter temperature inversion-episodes elevate PM levels as emissions become trapped in a stagnant air mass near the valley floor. These episodes occur under somewhat predictable conditions that include a combination of snow cover, high barometric pressure, and low or falling clearing index⁵. Arrangements were made with the research participants to have their blood drawn 4 times between January and early March of 2009 during a period of high pollution (PM_2.5 > 40 μg/m³), a period of moderate pollution (PM_2.5 approximately 20-40 μg/m³), and two periods of low pollution (PM_2.5 < 10 μg/m³) (Figure 1A). In the murine study, 28 C57BL/6 mice were exposed to either filtered air or PM_2.5, concentrated ~8-fold from ambient downtown Louisville air for 6 h/day for 9 consecutive days (see Online Supplement) at two different times. An expanded Methods section is available in the Online Data Supplement.

Figure 1. PM_2.5 levels are inversely correlated with human EPC number.

(A) Twelve h lagged moving average PM_2.5 levels (solid and dashed black lines) and the daily 24-hr PM_2.5 concentrations (blue and green lines) recorded at two monitoring sites in the Utah Valley between January 1-March 5, 2009. Red dots represent the times of blood draws; (B) Regression analysis of the relationship between CD31⁺/CD34⁺/CD45⁺/CD133⁺ cells (normalized to volume) and previous 24-h PM_2.5 level; and (C) regression analysis of the relationship between non-albumin protein (as a percent of total plasma protein) and average previous 24-h PM_2.5 levels. Individual data points are labeled with the subject numbers and individual-level regression is represented by dotted lines. Solid line represents regression analysis from the pooled fixed-effects regression model.

RESULTS

Ambient levels of PM_2.5 recorded in January to early March of 2009 by two Utah Valley monitoring sites are shown in Figure 1A. A detailed characterization of changes in PM composition during winter inversion has been recently published⁶. A substantial air pollution episode with peak PM_2.5 concentrations occurred around January 22^nd with a more moderate episode about 10-14 days later and a return to baseline levels thereafter. Blood samples were obtained 4 times from each study participant to provide measurements at both high and low PM levels for each individual. EPC populations were identified by a seven color cytometry procedure. The data were regressed on PM_2.5 (average of 24 h prior to blood draw) controlling for subject-specific fixed effects, using heteroskedasticity-consistent covariance matrix estimators.

The most abundant progenitor cell population (CD31⁺/CD34⁺) was negatively correlated with ambient PM_2.5 levels (Table 1). In addition, CD45±CD133 cell populations also demonstrated negative associations with varying statistical strength. The strongest statistical correlation was observed for CD34⁺/CD31⁺/CD45⁺/CD133⁺ cells (Figure 1B). Some data sets showed differences in variability. For example, in Figure 1B at PM_2.5 concentrations of about 35 μg/m³, there is much less variability in some data. These results suggest the need to estimate standard errors and p-values based on heteroskedasticity-consistent covariance matrix estimators. In most cases, these estimators resulted in slightly larger standard errors and corresponding p-values. Nevertheless, similar regression results were observed when these specific measurements were deleted from the regression analysis.

Table 1.

Summary statistics and regression coefficients for study variables regressed on PM_2.5 (average of 24h, before blood draw x 50μg/m³) from models controlling for subject-specific fixed variables

Parameter	Regression coefficient	SE	p-value
CD34+/CD31+ cells	−5.691	1.764	0.002
CD34+/CD31+/ CD45− cells	−4.275	1.569	0.009
CD34+/CD31+/ CD45− /CD133− cells	−4.244	1.559	0.009
CD34+/CD31+/CD45+ cells	−1.416	0.431	0.002
CD34+/CD31+/CD45+ /CD133− cells	−1.074	0.392	0.009
CD34+/CD31+/CD45+ /CD133+ cells	−0.342	0.090	0.0004
platelet-monocyte aggregates	1.387	0.683	0.048
HDL	1.943	0.740	0.012
CRP	−0.020	0.025	0.430
SAA	−14.82	31.32	0.638
Total plasma protein	0.383	0.163	0.023
Plasma albumin	0.033	0.104	0.751
Non-albumin protein	0.349	0.069	<0.0001
Non-albumin protein/total plasma protein	2.848	0.379	<0.0001

In addition to EPC levels, changes in systemic inflammation, coagulation, and plasma lipids were measured. As listed in Table 1, acute exposure to PM_2.5 was significantly correlated with an increase in platelet (CD41a⁺)-monocyte (CD45⁺) aggregates and HDL levels. No associations with SAA and CRP (Table 1) or LDL, triglycerides, fibrinogen, SDF-1, IL-6, IL-1β, VEGF, PF-4 (data not shown) were observed. PM_2.5 levels were, however, most strongly associated with non-albumin plasma protein (NAP) levels (Table 1). Absolute levels, or levels of these proteins expressed as a percent of total plasma protein, were both highly associated with PM_2.5 (Table 1). Unlike associations of PM with platelet-monocyte aggregates and HDL, which were of marginal significance, there was no anomalous heteroskedasticity observed in the NAP data. Given the incredible strength of the statistical association (p<0.0001) and given the remarkable consistency across all subjects (Table 1, Figure 1C), the association between NAP and PM_2.5 is unlikely to be an artifact of multiple hypothesis testing. Finally, there were no changes in markers of liver or skeletal muscle injury (data not shown) indicating that the effects observed are not symptoms of general tissue injury.

Because several confounding factors can influence the levels of circulating EPCs in humans, we examined PM-induced changes in mice. These mice were exposed to filtered air or concentrated ambient particles (CAPs) from downtown Louisville air for 9 days and blood levels of Sca-1⁺/Flk-1⁺ cells were measured. As shown in Figure 2, CAPs exposure resulted in a ~50 % decrease in Sca-1⁺/Flk-1⁺ cells during the two exposure periods. CAPs exposure was also associated with significant increases in total cholesterol and HDL, but no changes in NAP (Table 2). No significant changes in the number of bone marrow-derived EPCs were observed after 1wk in culture (Online Fig. VI). Collectively, these findings suggest that exposure to particulate air pollution in mice decreases EPC levels.

Figure 2. PM exposure decreases EPC levels in mice.

Flow cytometric analysis of peripheral blood obtained from mice exposed to filtered air or concentrated air particulates. (A) Sca-1⁺/Flk-1⁺ cells (right panels) were quantified in a gated lymphocyte population identified in an SSC vs. FSC dot plot (left panels). (B) EPC levels per μl blood in mice exposed to air or PM_2.5 during the indicated exposure periods (n=4, July 2009, n=8, May 2010; * p<0.05).

Table 2.

Blood and plasma parameters in C57BL/6 mice exposed to air or concentrated ambient particulate matter (PM_2.5) in July 2009

Variable	Air	PM2.5
HCta	0.44. ±0.01	0.44 ± 0.00
Buffy Coata	0.01 ± 0.00	0.01 ± 0.00
Cholesterolb	56.8 ± 1.3	62.4 ± 1.1*
HDLb	41.1 ± 1.2	44.7 ± 1.0*
LDLb	9.8 ± 0.6	10.5 ± 0.6
Triglyceridesb	45.6 ± 3.3	48.2 ± 4.0
TPc	4.07 ± 0.12	4.26 ± 0.07
ALBc	2.99 ± 0.07	3.07 ± 0.06
ALTd	20 ± 2	28 ± 5
ASTd	49 ± 5 (5)	58 ± 4
Creatinineb	0.25 ± 0.01	0.25 ± 0.01
Non-albumin proteinc	1.08 ± 0.08	1.19 ± 0.04

^aUnits: decimal fraction

^b[mg/dl]

^c[g/dL]

^d[U/L]

Values are mean ± SEM

Abbreviations: HCt, hematocrit; HDL, high density lipoprotein cholesterol; LDL, low density lipoprotein cholesterol; TP, total protein; ALB, albumin; ALT, alanine aminotransferase; AST, aspartate aminotransferase

^*p <0.05 from control by One Way ANOVA and Bonferroni post-test. (n=4)

DISCUSSION

The major finding of this study is that exposure to high PM_2.5 levels induces reversible vascular injury as evidenced by a suppression of circulating EPC levels in both humans and mice. In humans, this was also accompanied by an increase in platelet activation and elevated levels of plasma HDL and NAP. No significant changes in markers of systemic inflammation or tissue injury were observed. These results support the notion that suppression of EPC levels in peripheral blood may be an important feature and, perhaps, a significant mechanism of PM-induced cardiovascular injury. Moreover, EPC level may be a sensitive, albeit non-specific, biomarker of endothelial injury due to PM exposure.

Although the mechanisms by which PM exposure decreases the circulating EPC levels remain unclear, concurrent increases in thrombosis, and HDL indicate that the loss of EPCs from peripheral blood may be due to endothelial injury. Consistent with this idea, the increase in NAP is likely reflective of an increase in globulin levels. After albumin, globulins are the second most abundant proteins in the plasma and an increase in their levels may be reflective of a mild systemic immune response.

Vascular dysfunction and endothelial injury are well described effects of PM exposure^1-2. Our observation that PM decreased EPC levels in humans and mice exposed to similar doses (25μg/kg per 24h day and 54μg/kg per 6h day, respectively) and despite potential compositional differences between Provo and Louisville air, suggests that suppression of EPC levels is a robust response to PM exposure. Previous studies show that chronic exposure to tobacco smoke, which contains high levels of PM_2.5 and other pollutants, is also associated with low EPC levels⁷. Thus, exposure to air particulates and/or their co-pollutants from several sources may have the general property of reducing circulating EPC levels. Further studies are required to identify specific PM components that might be related to EPC suppression.

The CD34⁺ cell population, which was significantly correlated with PM_2.5 levels, has been shown previously to be associated with CVD risk⁴. Moreover, stronger association with CD133⁺ than CD133⁻ cells suggests that PM exposure affects the immature, early EPC population. This population shows a 5-fold reduction in patients with coronary artery disease and is predictive of adverse cardiovascular events in patients with pre-existing CVD⁴. Significantly, the levels of non-monocytic EPC population (CD45⁻) cells were also suppressed upon PM exposure. Therefore, depletion of multiple EPC populations could contribute to CVD risk imposed by PM_2.5 exposure by inducing deficits in endothelial repair and angiogenesis. However, we could not study functional changes because the effects of PM on EPC levels in humans were reversible. Moreover, ex vivo assays to assess EPC function or proliferation require prolonged (7-21 days) culture, during which time the PM-induced changes are likely to be lost. Nevertheless, reversible suppression of EPC levels suggests that exposure to PM induces a transient mismatch between EPC utilization and recruitment. Given that this gap is robustly associated with several cardiovascular diseases^3-4, it appears likely that depletion of circulating EPCs, along with changes in blood coagulation, lipids and non-albumin proteins is reflective of vascular injury induced by PM even in the absence of overt cardiovascular disease.

NOVELTY AND SIGNIFICANCE.

What is known?

• Several epidemiological studies show that acute exposure to elevated levels of fine airborne particulate matter is associated with an increase in the risk of adverse cardiovascular events.

• Controlled laboratory exposure to particulate matter has been reported to induce acute conduit artery vasoconstriction in humans and chronic deficits in endothelium-mediated vasodilation in mice.

• Endothelial progenitor cells (EPC) in peripheral blood contribute to post-embryonic endothelial repair and regeneration and a decrease in circulating EPC levels is reflective of cardiovascular disease risk and burden.

What new information does this article contribute?

• An increase in air borne particulate matter due to winter temperature inversion-episode in Utah Valley of the Wasatch Front was associated with a reversible decrease in circulating levels of EPC in a cohort of young (18-25 years), healthy adults.

• The increase in particulate matter was also accompanied by an increase in plasma levels of platelet-monocyte aggregates, HDL, and non- albumin protein. No changes in C-reactive protein, IL-1β, IL-6, fibrinogen or serum amyloid A were observed.

• Circulating levels of EPC were also decreased in mice exposed to concentrated air borne particles from downtown Louisville, Kentucky.

Particulate air pollution contributes to cardiovascular dysfunction and mortality but mechanisms for this remain unclear. Here we show that episodic exposure to high levels of particulate matter decreased circulating EPCs in young adults and that this effect was reversible. These effects were accompanied by an increase in markers of thrombosis, but no change in systemic inflammation. Exposure to concentrated PM also decreased circulating EPCs in mice. Consistent data between humans and mice at two locales suggests that depletion of circulating EPCs is a characteristic feature of PM exposure and may be one mechanism by which PM contributes to cardiovascular disease.

Non-standard Abbreviations and Acronyms

NAP

non-albumin protein

EPC

endothelial progenitor cell

PM

particulate matter

CAPS

concentrated ambient air particles

CVD

cardiovascular disease