![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081106005900im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
2008 Progress Report: Assessing Toxicity of Local and Transported Particles Using Animal Models Exposed to CAPs
EPA Grant Number: R832416C003Subproject: this is subproject number 003 , established and managed by the Center Director under grant R832416
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: Harvard Particle Center
Center Director: Koutrakis, Petros
Title: Assessing Toxicity of Local and Transported Particles Using Animal Models Exposed to CAPs
Investigators: Godleski, John J. , Koutrakis, Petros
Institution: Harvard University
EPA Project Officer: Stacey Katz/Gail Robarge,
Project Period: October 1, 2005 through September 30, 2010
Project Period Covered by this Report: October 1, 2007 through September 30,2008
RFA: Particulate Matter Research Centers (2004)
Research Category: Particulate Matter
Description:
Objective:The objective of this project is to differentiate the toxicological effects of locally emitted and transported particles. To do so, short-term 5 hr animal exposures to concentrated ambient fine particles (CAPs) were conducted during the time periods of 5-10 am and 10:30 am-3:30 pm. Starting inhalation exposures at 5 am, before significant vertical mixing takes place, captures particles predominantly from local sources, while, exposures starting about 10:30 am are relatively more enriched in transported particles. Specific biologic outcomes included: breathing patterns, indicators of pulmonary and systemic inflammation, blood pressure, in vivo oxidant responses in the heart and lung, and quantitative morphology of lung and cardiac vessels. To control for circadian variations all outcomes were assessed during both time periods, in relation to those of filtered air (sham) exposures. Animal exposures were characterized using continuous measurements of particle mass, size, number, and black carbon, as well as integrated measurements of particle mass, sulfate, elements, and organics. Strains of rats used include Sprague Dawley (SD), which we have used extensively in previous CAPs studies, Spontaneously Hypertensive Rats (SHR), a sensitive model in many studies, and Wistar Kyoto (WKY) the strain control for SHR rats. Studies of cardiopulmonary mechanisms in relationship to in vivo oxidant responses in the heart and lung were also carried out using CAPs.
Approach:To differentiate the toxicological effects of locally emitted and transported particles on important cardiovascular outcomes, short term animal exposures to CAPs will be conducted during the time periods of 6-10am and 11am-3pm. Starting inhalation exposures at 6am before significant vertical mixing takes place will allow us to capture particles mostly from local sources. In contrast, exposures starting at 11am will be relatively more enriched in transported particles. All outcomes will be assessed in relation to those of filtered air (sham) exposures as well as those of positive controls using particles of known toxicity at both time periods to control for circadian variations. Animal exposures will be characterized using continuous measurements of particle mass, size, number, and black carbon, as well as integrated measurements of particle mass, sulfate, elements and organics. Specific outcome measurements will include: indicators of pulmonary and systemic inflammation, blood pressure, endothelin-1, endothelial nitric oxide synthase, atrial naturetic peptide, in vivo oxidant responses in the heart and lung, and quantitative morphology of lung and cardiac vessels. Statistical analyses will use multi-way ANOVA to assess differences among exposure groups and interactions of exposure and potential effect modifiers. Regression techniques will be used to examine dose-response relationships between measured biological outcomes and particle source contributions as reflected by particle composition. Multiple linear regression using tracer elements will be used to assess the independent effects of multiple pollution sources.
Progress Summary:Early-late experiments. Exposure data from all the experiments are shown in Table 1. CAPs mass concentrations (Early - 505.8 ± 75.8 µg/m3 and Late 407.2 ± 45.7 µg/m3) were slightly higher than previous published studies from our laboratory, and not significantly different from each other using a paired two-tailed t-test. There are significant differences in black carbon and elemental carbon between the early and late exposure, supporting the premise that the early exposure would be more influenced by local (primarily traffic sources) whereas the exposures later in the same day were more likely to contain transported particles. Since these experiments were not done in the summer (when diurnal variation in sulfate production from oxidation of SO2 is greater than in other seasons), there is no significant difference in sulfur between the morning and afternoon. In these studies, in addition to greater black and elemental carbon in the morning, iron, nickel, and copper levels were also significantly higher in the morning. When these data were analyzed using the ratio of a specific component to the total fine mass (or fraction of the component) essentially the same findings were observed with more robust p-values. Thus, elemental carbon, copper, and nickel were significantly higher in the morning. In addition, several components were found to be significantly higher in the afternoon. These include sodium, potassium, magnesium, manganese, and silicon. The sulfur fraction was higher in the afternoon than the morning, but this difference was not significant. Even though statistically significant differences were found between AM and PM exposures, the differences were modest.
Table 1. CAPs mass and component concentrations during the early and late exposure periods
(concentrations are expressed in µg/m3 and fractions in %)
|
Early mean±SE
|
Late mean±SE
|
p =
|
|
|
CAPs Mass
|
505.8 ± 75.8
|
407.2 ± 45.7
|
0.083
|
|
*Black Carbon Mass
|
10.5 ± 0.9
|
7.3 ± 1.1
|
0.002
|
|
*Elemental Carbon
|
22.5 ± 2.6
|
16.5 ± 2.8
|
0.032
|
|
Organic Carbon
|
72.2 ± 6.9
|
67.3 ± 6.9
|
0.505
|
|
Total Carbon
|
94.9 ± 9.0
|
83.8 ± 9.5
|
0.261
|
|
Sodium
|
8.9 ± 2.6
|
10.4 ± 2.9
|
0.181
|
|
Chlorine
|
9.7 ± 3.6
|
13.8 ± 6.2
|
0.174
|
|
Silicon
|
9.5 ± 1.6
|
8.8 ± 1.1
|
0.448
|
|
Aluminum
|
3.4 ± 0.6
|
3.1 ± 0.4
|
0.437
|
|
Sulfur
|
37.0 ± 5.9
|
35.7 ± 5.3
|
0.832
|
|
Calcium
|
6.3 ± 0.9
|
6.1 ± 0.8
|
0.761
|
|
Titanium
|
0.33 ± 0.05
|
0.26 ± 0.04
|
0.108
|
|
Potassium
|
2.8 ± 0.3
|
2.7 ± 0.3
|
0.465
|
|
*Iron
|
13.3 ± 1.9
|
9.7 ± 1.0
|
0.035
|
|
Zinc
|
1.0 ± 0.1
|
1.1 ± 0.1
|
0.857
|
|
*Nickel
|
0.07 ± 0.015
|
0.04 ± .007
|
0.033
|
|
Vanadium
|
0.03 ± 0.01
|
0.01 ± .009
|
0.144
|
|
Magnesium
|
1.2 ± 0.3
|
1.4 ± 0.3
|
0.293
|
|
*Copper
|
0.4 ± 0.06
|
0.2 ± 0.03
|
0.019
|
|
Manganese
|
0.2 ± 0.03
|
0.3 ± 0.04
|
0.349
|
|
*EC Percent of Mass
|
5.3 ± 0.5
|
3.9 ± 0.4
|
0.007
|
|
OC Percent of Mass
|
17.8 ± 1.7
|
18.7 ± 1.5
|
0.586
|
|
TC Percent of Mass
|
23.1 ± 2.1
|
22.5 ± 1.8
|
0.753
|
|
*Sodium Percent of Mass
|
2.3 ± 0.7
|
2.7 ± 0.7
|
0.026
|
|
Chlorine Percent of Mass
|
2.3 ± 0.9
|
3.1 ± 1.3
|
0.180
|
|
*Silicon Percent of Mass
|
2.5 ± 0.4
|
2.9 ± 0.5
|
0.040
|
|
Aluminum Percent of Mass
|
0.89 ± 0.17
|
1.03 ± 0.19
|
0.107
|
|
Sulfur Percent of Mass
|
7.5 ± 0.7
|
8.6 ± 0.8
|
0.123
|
|
Calcium Percent of Mass
|
1.6 ±0.3
|
2.0 ± 0.4
|
0.056
|
|
Titanium Percent of Mass
|
0.09 ± 0.01
|
0.09 ± 0.02
|
0.814
|
|
*Potassium Percent of Mass
|
0.68 ± 0.07
|
0.77 ±0.08
|
0.017
|
|
Iron Percent of Mass
|
3.3 ± 0.4
|
2.9 ± 0.4
|
0.226
|
|
Zinc Percent of Mass
|
0.25 ± 0.03
|
0.29 ± 0.04
|
0.429
|
|
*Nickel Percent of Mass
|
0.019 ± 0.004
|
0.009 ± 0.001
|
0.025
|
|
Vanadium Percent of Mass
|
0.007 ± 0.002
|
0.003 ± 0.002
|
0.074
|
|
*Magnesium Percent of Mass
|
0.29 ± 0.07
|
0.39 ± 0.07
|
0.048
|
|
*Copper Percent of Mass
|
0.09 ± 0.01
|
0.07 ± 0.01
|
0.043
|
|
*Manganese Percent of Mass
|
0.06 ± 0.01
|
0.08 ± 0.01
|
0.040
|
These data can also be assessed by strain for each study carried out. These analyses are shown for selected elements in Table 2. In this analysis, it can be seen that for specific experiments there are greater differences between early and late exposures than are apparent in the overall averages. Different levels of exposures received by the different strains and seasonal groups were due to the random variations in composition of the ambient pollution in Boston for different exposure days.
Table 2. CAPs mass and component concentrations during the early and late exposure periods by time
of year and by strain (concentrations are reportedin µg/m3 , practicle count is in 1,000
particles per cm3)
Twenty one repetitions of the early-late experiments were carried out, and animals were studied for breathing pattern, BAL, and chemiluminescence outcomes, in each treatment group for each strain. The numbers of animals studied provided sufficient power to examine differences between groups, based on our previous studies. Table 3 shows the difference between all early and late sham exposed animals for respiratory parameters. Since there are significant differences between these animals, the need to control for diurnal variation in analyses of respiratory parameters is emphasized. All Data reported in this study are presented as continuous differences in parameter measurements between CAPs and Sham animal exposures to compensate for this diurnal variability.
Table 3. Early-late differences in sham animals indicating significant diurnal effects may influence
these exposures.
For all rat strains, breathing pattern and in vivo chemiluminescence studies show significant differences between CAPs and Sham exposures. With breathing pattern there is an increase in respiratory frequency with concomitant shortening of the time of inspiration and expiration. Statistical modeling was used to assess the size and strength of association between CAPs or Sham exposure and each respiratory outcome. Additive mixed models were applied to 10-minute averaged data collected from all CAPs and Sham animals during AM or PM exposure to estimate overall, time -and species-specific effects of exposure. The models represent an extension of linear regression models that make it possible to estimate exposure effects while (1) controlling for potentially non-linear effects of time within exposure period, and (2) including random animal effects to account for correlation among repeated measurements taken on the same animal during the exposure period. For each outcome, four models using increasing level of detail for exposure effects were run. These models corresponded to estimating overall, time-specific, species specific effects, and time-specific effects for each species. Random forests methods were used to rank the importance of each measured CAPs component in predicting differences between CAPs and filtered air outcome means. This approach represents an extension of classification and regression trees (CART), identifying predictors that yield the largest gains in prediction accuracy of the response (in this case the CAPs-Sham mean difference in outcome for a given exposure) achieved by CART when applied to multiple bootstrap resamples of the original data. This analysis was supplemented by plotting raw CAPs - Sham mean differences for each outcome versus the components identified in the random forest analysis.
With breathing pattern, there are substantial differences between the spring and the fall. The overall pattern tends to follow the fall pattern. There was much greater variability in spring. Overall, there is an increase in respiratory frequency with concomitant shortening of the time of inspiration and expiration. An overall increase of frequency was seen in all groups/strains with CAPs exposure compared to controls, (13.081±3.139 p<0.0001). This increase was also significant in the early morning exposure for the 3 strains (SD, SHR and WKY; p≤0.01). Inspiratory and expiratory times were decreased overall (-0.007±0.003 and -0.013±0.005 respectively; p≤0.05). Figure 1 shows the change in respiratory parameters for early morning and late day exposures by strain and overall. Although responses for animals exposed in the morning were often greater than those of animals exposed in the afternoon, this difference was only significant for frequency in the SHR rats (p=0.0393). Tidal volumes tended to increase in the overall analyses and there were no early-late differences that were significant. Inspiratory and expiratory flows had some statistically significant differences, but these tended to be increases. With increases in tidal volumes and flows in these experiments, there does not appear to be toxicologically significant changes in pulmonary function, even though there is a definite change in breathing pattern. The rapid shallow pattern suggests that the morning exposure was sensed by the animals to be sufficiently irritative to change breathing pattern more so than in the afternoon. However, change in breathing pattern was not accompanied by decrease in volume or flow suggesting that there was not either a bronchoconstrictive change or inflammatory change in the airways. This lack of change in these parameters was corroborated by BAL and histological findings which did not show any significant changes.
Changes in the pause parameters also showed little change between the early and late exposures. Penh showed a significant increase overall in the morning exposures which appeared to be largely related to the changes in this parameter for SD rats in the fall exposure group. Univariate analyses indicated very few significant relationships. The random forest analysis for frequency and the predicted ranking of elements associated with that response are shown in figure 2. The relationship between CAPs vs. Sham differences in frequency and concentrations of the first 4 ranked components in the random forest are shown in figure 3 along with similar plots for BC and total sulfur. Zinc, lead, calcium top the random forest list and all have significant positive relationships (increasing change in frequency with increasing concentration of the component). Tin (Sn) has a negative relationship. Sulfur has a weak positive relationship, but there is no relationship with black carbon.
Figure 2: Dose-respiratory frequency response of all exposure components using the random forest
as an approach to multifariate analysis
Figure 3: Scatter plots of changes in respiratory frequency vs exposure concentrations of components
identified yb random foreset ranking.
In summary, diurnal differences in breathing pattern were found, and these pattern differences were enhanced by CAPs exposure. Responses of animals exposed early in the day were greater than and different from the responses of animals exposed at mid day. Responses of animals exposed during the spring were very different from the responses of animals exposed during the fall. The basis of this difference is not explained by differences in exposure composition, but differences in meteorological conditions such as temperature and hours of light need to be further explored. Changes in respiratory parameters were not strongly correlated with expected differences in black carbon and sulfur concentrations. Zinc, an element usually regarded as an emission of diesel engines, had similar concentrations in the morning and mid-day, but was consistently ranked as one of the components most strongly associated with changes in respiratory parameters. Sprague Dawley and SHR rats had comparable responses to the exposures, while WKY rats consistently had a lesser response to the same exposures. These results were presented at the 2008 American Thoracic Society meeting in Toronto by Dr. Edgar Diaz.
Data from the in vivo chemiluminescence studies show that the lung had significant effects of CAPs exposures whereas chemiluminescence changes in the heart did not reach significance in these exposures. There were significant changes with both early and late exposures, and analyses showed that early and late responses were not significantly different. Because the CAPs effects from the different rat strains were not significantly different from one another, and the CAPs effects during early and late exposures were also not significantly different, component analyses with the heart and lung data estimated overall concentration slopes not segregated by strain or early/late exposures. In these analyses, no cardiac effects were found using univariate or multivariate analyses. Table 4 illustrates many components with significant univariate relationships to lung chemiluminescence, but none of these were significant in multivariate analyses.
Table 4. Effects of specific exposure components on in vivo lung chemiluminescence using univariate
analyses.
|
Element/
Component
|
Estimate ± SE
|
P value
|
|
CAPs Mass
|
0.012±0.005
|
0.027
|
|
Organic Carbon
|
0.106±0.038
|
0.0064
|
|
Elemental Carbon
|
0.298±0.121
|
0.016
|
|
Al
|
1.573±0.739
|
0.036
|
|
Si
|
0.577±0.267
|
0.034
|
|
S
|
0.132±0.061
|
0.033
|
|
Fe
|
0.509±0.218
|
0.022
|
Overall, the analyses largely confirm our earlier studies and findings with CAPs exposures in Boston. It is of particular interest that, apparently, despite statistically significant differences in the composition of early and late exposures on given days, there is no significant difference in toxicity. It seems that our studies have adequate statistical power, since we were able to detect significant diurnal differences with respiratory patterns, significant strain differences, as well as CAPs vs Sham differences. Since both early and late exposures show significant toxicity, with no significant difference in the biological outcomes between the two exposure periods, these results do not suggest any difference in the toxic potential of local and transported sources.
Exposures of WKY and SH rats, monitoring blood pressure, electrocardiogram, and blood parameters have been completed, and data analyses are in progress.
Mechanistic studies using In vivo Chemilumenescence: A number of important mechanistic studies in this area have been completed in the past year supported within this project (Rhoden et al 2008; Ghelfi et al 2008). These animal studies are important correlates to the studies of oxidant stress and autonomic function in Project 1. Studies (Rhoden et al 2005) supported in Project 4 suggest that ambient particles modulate autonomic tone leading to cardiovascular oxidant stress and dysfunction. The relationship between inhaled CAPs, neurogenic signals from the lung, and effects on cardiac oxidative stress, has now been directly investigated (Ghelfi et al 2008). In these studies, the effect of blockade of vanilloid receptor 1 (TRPV1) was assessed in relationship to CAPs-induced cardiac oxidative stress and dysfunction in a rat model of inhalation exposure. Capsazepine (CPZ), a selective antagonist of TRPV1, was given intraperitoneally or as an aerosol immediately before exposure to CAPs. Control and CPZ-treated rats were exposed to filtered air or CAPs aerosols for 5 hours using the Harvard Ambient Particle Concentrator (mean PM2.5 mass concentration: 218 ± 23 µg/m3). At the end of the exposure, cardiac oxidative stress was measured using in situ chemiluminescence: CL), lipid peroxidation (TBARS), and tissue edema. Cardiac function was monitored throughout the exposure, and measures of heart rate, heart rate variability, and cardiac conduction times were assessed. As shown in Figure 4, CPZ (aerosol) decreased CAPs-induced CL. Lipid TBARS and edema in the heart were similarly affected indicating that blocking TRP receptors, systemically or locally, decreases heart CL.
Figure 4. CPZ aerosolization prevents oxidative stress in the heart of rats exposed to CAPs as
measuresd with in vivo chemiluminescence. Values represent the mean of eight
independent determinations ± SEM. *p < 0.05.
CAPs exposure led to significant decreases in HR (CAPs 350±32 bpm, control: 370±29), and in the length of the RT, Pdur, QT and Tpe intervals. These changes were observable immediately upon exposure and were maintained throughout the 5 hours of CAPs inhalation. Changes in cardiac rhythm and ECG morphology were prevented by CPZ. These data are shown in Table 5 below.
Table 5: Change in ECG Parameters during Acute Exposure to either Filtered Air or CAPs with CPZ
blocade of Vanilloid receptor 1 in the lung.
The findings in Table 5 suggest that cardiac conduction current abnormalities in CAPs-exposed rats alter action potentials leading to changes in conduction velocity and ventricular repolarization. Taken together these results suggest that inhaled CAPs stimulate TRVP1 (and possibly other pulmonary irritant receptors) and thereby activate autonomic nervous system reflexes. The end result of this reflex activation is increased cardiac oxidative stress, and functional cardiac electrophysiologic changes including increased P-wave duration and QT interval and decreased QRS and Tpe durations. Thus, CAPs exposure results in cardiac current abnormalities leading to changes in conduction velocity and ventricular repolarization, and that triggering of TRPV1-mediated autonomic reflexes in the lung is essential for the observed changes in conduction, repolarization and cardiac rhythms.
In other mechanistic studies, Rhoden et al (2008) used a specific blocker of superoxide anion to assess the role of this oxidant in induction of pulmonary inflammation with instilled ambient particles. The findings of this study indicate that superoxide anion plays a significant role in the development of inflammation. Investigators of this project have also carried out additional analyses of studies completed in the previous center and published these studies (Wellenius et al 2006). Important reviews of the cardiovascular effects of air particulate on the cardiovascular system (Godleski 2006) as well as gave several invited presentations at national meetings on this topic were given by Dr. Godleski. Finally, in collaboration with the statistical core of the Center, a number of statistical methods papers have been published by members of core and this project (Nikolov et at 2007, 2008)
This proposal offers the unique application of novel techniques to improve understanding effects of specific sources of particles and mechanisms of health effects. This combination of exposure scenarios and pulmonary and cardiovascular outcomes will provide new data to assess the effect of specific particle sources on specific mechanistic pathways by which ambient air particles produce adverse health effects.
Supplemental Keywords:Air Particulates, Inhalation Exposure, Cardiovascular, Pulmonary Mechanisms, concentrated air particles, acute cardiovascular effects, coarse particles, fine particles, vascular dysfunction
Relevant Websites:
http://www.hsph.harvard.edu/epacenter
Progress and Final Reports:
2006 Progress Report
2007 Progress Report
Original Abstract
Main Center Abstract and Reports:
R832416 Harvard Particle Center
Subprojects under this Center:
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R832416C001 Cardiovascular Responses in the Normative Aging Study: Exploring the Pathways of Particle Toxicity
R832416C002 Cardiovascular Toxicity of Concentrated Ambient Fine, Ultrafine and Coarse Particles in Controlled Human Exposures
R832416C003 Assessing Toxicity of Local and Transported Particles Using Animal Models Exposed to CAPs
R832416C004 Cardiovascular Effects of Mobile Source Exposures: Effects of Particles and Gaseous Co-pollutants
R832416C005 Toxicological Evaluation of Realistic Emission Source Aerosol (TERESA): Investigation of Vehicular Emissions