Carbonaceous particulate matter on the lung surface from adults living in São Paulo, Brazil

Accumulation of carbonaceous particulate matter (PM) in the lung is associated with chronic disease. The amount of carbonaceous PM in airway macrophages is reported to be associated with exposure to both fossil fuel PM and cigarette smoke. However, the contribution of these exposures to carbonaceous PM at the lung surface is unclear. Objectives We therefore sought to identify the exposures associated with lung surface in long-term residents of São Paulo, Brazil. Methods Lung surface carbon were analyzed in 72 autopsy specimens by image analysis. Smoking history, measured PM10 nearest to the home, distance to main road, and distance-weighted traffic density were used as exposure variables. Data are summarized as median (IQR), and compared by Mann Whitney Test, with correlations done by Spearman’s correlation. Results There was no association between lung surface and age or gender. There was no statistically significant association in lung surface between smokers and non-smokers 6.74 cm2 (3.47 to 10.02) versus 5.20cm2 (2.29 to 7.54), and there was no significant association between lung surface carbon and exposure to environmental PM and markers of traffic exposure. Conclusion We did not find a statistically significant association between lung surface and smokers and non-smokers, and no statistically significant association between lung surface carbon and environmental exposure variables. These results suggest that lung surface carbon in long-term residents of São Paulo may predominately be from environmental PM, but the most appropriate environmental exposure marker remains unclear.


Results
There was no association between lung surface and age or gender. There was no statistically significant association in lung surface between smokers and non-smokers 6.74 cm 2 (3.47 to 10.02) versus 5.20cm 2 (2.29 to 7.54), and there was no significant association between lung surface carbon and exposure to environmental PM and markers of traffic exposure.

Conclusion
We did not find a statistically significant association between lung surface and smokers and non-smokers, and no statistically significant association between lung surface carbon and environmental exposure variables. These results suggest that lung surface carbon in longterm residents of São Paulo may predominately be from environmental PM, but the most appropriate environmental exposure marker remains unclear. PLOS

Introduction
Retention of inhaled carbonaceous particulate matter (PM) in the lung is associated with a wide range of adverse health effects. The association between accumulation of carbonaceous PM in the lung and chronic lung injury was first described over 100 years ago in miners [1], and may be associated with the inhalation of pollutant particles in ambient air among elderly people [2]. Another study carried by Brauer et al [3], reflected that the long-term exposure to PM in adult's autopsied lungs who lived in a region with high levels of ambient particles results in pulmonary retention of large quantities of fine and ultrafine particle aggregates, mostly appearing to be combustion products. By contrast, evidence for the long-term adverse effects of environmental carbonaceous PM, mainly from fossil-fuel combustion in urban areas, has emerged recently. For example, a 2016 report by the Royal College of Physicians (UK), concluded that long-term exposure to carbonaceous fossil fuel derived PM less than 10 micrometers in aerodynamic diameter (PM 10 ) is associated with a wide range of long-term effects including reduced lung function growth in children, accelerated lung function decline in adults, lung cancer, and new onset asthma [4]. In adults, an additional source of carbonaceous PM exposure is cigarette smoke [5].
Assessment of the amount of carbon in airway macrophages using induced sputum in children was previously reported in different studies and has been found to have significant correlations with particulate matter exposure. [6]. More recently, Belli et al [7] reported that both smoking and exposure to environmental PM 2.5 is associated with accumulation of carbonaceous PM in alveolar macrophages. By contrast, little is known about the amount of PM retained in lung tissue after exposure to fossil fuel-and cigarette smoke derived PM, in part because measuring carbonaceous PM in lung tissue is difficult to do in vivo. Although one of the hallmarks of long-term smoking is considered to be blackening of the lung tissue surface, no studies to date have compared surface carbonaceous PM loading in smokers and nonsmoking adults. However, recently, You et al [8] performed high resolution transmission electron microscopy of the residual black material after complete proteolytic digestion of human emphysematous lung from smokers and found 20-50nm spheroid aggregates compatible with carbon black, and nanoparticulate carbon black in dendritic cells from the same lungs. You et al [8] also found that exposure of mice to cigarette smoke, increased black staining both at the lung surface and within dendritic cells. However, this study did not compare lungs from non-smokers, and whether smoking is the major determinant of lung tissue carbon in adults living in areas of high environmental PM therefore remains unclear.
In this study, we hypothesized that smoking is associated with lung tissue carbon, and therefore sought the determine lung surface carbon in long-term residents of São Paulo, a highly polluted city, with also very high levels of cigarette smoking. In São Paulo, air pollution sources are primarily traffic-derived [9], frequently exceeding the levels established by WHO guidelines [10].

Study population
Post mortem lungs were obtained from São Paulo Autopsy Centre from January to August in the year 2014 and informed consent was obtained from next of kin after death. The Autopsy Centre performs post mortem services for all of the deaths in the São Paulo metropolitan area. Only lungs from adults who had lived within the city limits of São Paulo for at least 10 years were included, and were classified by a pathologist as free of acute respiratory disease at the gross examination. Smoking history was determined from a questionnaire completed by

Analysis of carbon
Lung surface carbon. A digital image of lung surface was obtained from the right upper lobe from each specimen. Images were taken using a Nikon digital camera (Nikon D-3300) in a professional light box with blue background. A 7 x 7 cm cropped image from the right upper lobe was then obtained from the area with the least surface indentations. Lung surface carbon was analyzed using Image J software (National Institute of Health, MD, USA), with the operator blinded to the information about the deceased. By comparing to the original color digital image, an operator adjusted the "threshold" command in Image J, to best capture black areas. Lung surface carbon was expressed as cm 2 black carbon /49 cm 2 lung surface, Fig 1A and 1B. Cigarette smoking. Smoking status was determined from the questionnaire of relatives used by the São Paulo Autopsy Centre. In this questionnaire, relatives are asked if the subjects were smokers or non-smokers at the time of the death. Pack-years was not assessed in the questionnaire due to the large information bias when applied to the family. Cause of death was obtained from medical records.
Measured environmental PM exposure. We obtained measurements of environmental PM 10 using data from monitoring stations across São Paulo metropolitan area, which was supplied by the São Paulo Environmental Agency (CETESB) [9]. Monitoring stations are situated throughout the Sao Paulo metropolitan area, and the whole population therefore lives within 10 km of a station. For each subject, PM 10 measurements was obtained from the nearest monitoring station based on their home addresses. Data was recorded 24 h before death, and mean annual exposures at the home address for 2013, 2012, and 2011 respectively. Home road distance and distance weighted traffic density. Road and traffic emissions near to the home address was assessed using a method previous described by Habermann and Gouveia [11]. Exposure to local traffic was determined using the distance from home to heavy traffic roads (m), and the distance-weighted traffic density-DWTD (vehicles/hour).
Histopathology. To assess the distribution of carbon in the lung, an additional right lung specimen was randomly selected at the Autopsy Centre after the pathologist's final evaluation. The lung was infused with 10% neutral buffered formalin with a non-specific pressure and fixed for 48 hr. A set of transverse sections was performed after that in order to observe the deep layers and the carbon intake. Sections were stained with hematoxylin and imaged with a Nikon D-3300 camera.

Statistical analysis
Data are described as median (IQR). Comparisons between groups were done by Mann Whitney test. Correlations were done by Spearman rank test. A P value <0.05 was considered significant. Statistical analysis was done using SPSS version 23 for Windows software (SPSS Inc., Chicago, IL, USA).

Results
Lung surface carbon was assessed in all 72 specimens analyzed. Demographics and cause of death are given in Table 1. Environmental PM exposure data were obtained for all subjects ( Table 2). There was a marked heterogeneity in Lung surface carbon between individuals. In some specimens, the lung surface was mainly pink with fissures relatively free of carbonaceous PM. In other specimens, large amounts of carbonaceous PM accumulated both in fissures with and on the lung surface, Fig 2. There was no statistically significant association between age and lung surface carbon (Rs = -0.018; P = 0.119), and there was no statistically significant difference in lung surface carbon between males and females (4.24cm 2 (1.9 to 8.45) vs. 6.18cm 2 (4.69 to 9.95) P =). There was no statistically significant difference in lung surface carbon between smokers and nonsmokers 6.74 cm 2 (3.47 to 10.02) versus 5.20cm 2 (2.29 to 7.54) Fig 3, and there was no statistically significant difference between lung surface carbon and exposure to environmental PM at the nearest monitoring station to the home address (Table 2). There was no statistically significant difference in lung surface carbon and distance to a major road categorized by <150 m and ! 150m (4.26cm 2 (1.69 to 6.77) vs. 6.11cm 2 (3.5 to 9.55) P = 0.058). The lung randomly selected for histological analysis was from a smoker. Carbonaceous PM was present in alveolar macrophages, in tissue at the lung surface, and in non-surface lung tissue, Fig 4A and 4B.

Discussion
In this study, we, for the first time, used image analysis to measure the amount of carbonaceous PM at the lung surface of long-term residents of São Paulo. Since Saxena et al [12] reported high levels of extracted elemental carbon deposits in autopsied lungs of smokers, we hypothesized that smoking is the major determinant of carbonaceous PM at the lung surface. However, despite significant amounts of lung surface carbon in many lung specimens, we found no statistically significant difference in lung surface carbon between smokers and nonsmokers.
Since we found no difference in lung surface carbon between smokers and non-smokers, our hypothesis that smoking is the major determinant of black staining at the lung surface is not supported. However, other reported findings in airway macrophages provide clear evidence that smoking does increase the carbon PM burden. Why this is not reflected by increased translocation of PM to the lung surface is to date unclear. A putative explanation is that in residents of São Paulo, the contribution of smoking is overwhelmed by chronic inhalation of very high levels of ambient PM. Indeed, previous studies suggest that environmental PM is a major driver of accumulation of carbonaceous PM within lung tissue. For example, Saxena et al [12] reported extracted elemental carbon PM from the lungs of non-smokers, and Pinkerton et al [13], reported carbonaceous PM in lymphatics in the sub pleural interlobular septa of non-smoking Californian residents. Furthermore, animal models suggest an association between lung surface carbon and inhaled dose of environmental PM. First, in cats exposed to 20h diesel exhaust PM for 28 days, Pepelko et al [14] reported increased charcoal grey staining of the lung surface compared control animals exposed to air. Second, in rats, Kato et al [15] reported that exposure for 24 to 60 weeks to urban roadside filtered air, resulted in a dosedependent increase in carbonaceous staining of the lung surface. Third, in rats with repeated inhalation of biodiesel PM for 13 weeks, Finch et al [16] reported that increased inhaled dose resulted in increased grey discoloration of the lung surface. Although our study is the first to quantify the amount of carbon at the lung surface of human smokers and non-smokers, evidence of a major role of environmental PM on black staining is provided by the 1971 study of Pratt and Kilburn [17], who in an analysis of 250 non-emphysematous adults lungs, concluded that pigmentation may be an indicator of exposure to environmental PM.
Although we found that smoking is not a major determinant of lung surface carbon, we found no association between lung surface carbon and markers of environmental PM-either mean annual PM 10 at the nearest monitoring station to the home address, or distanceweighted traffic density, or distance of home to the nearest heavily used road. Although these exposure variables are reported to be associated with adverse health effects in urban populations [18,19,20,21,22], in our population these markers do not capture exposures to environmental PM associated with deposition of carbon PM at the lung surface. We did not determine the effect of accumulation of carbonaceous PM at the lung surface on clinically relevant pathological changes. However, studies have previously reported pathological abnormalities in the lung adjacent to carbon PM aggregates. For example, alveolar macrophages are more numerous in smokers than in non-smokers according to a study reported by Wallace et al. [23]. In another study, Pinkerton et al [13] reported an association between retained tissue carbon and fibrosis in adjacent respiratory bronchioles. We did however, confirm by histological analysis of a lung from a single adult, significant accumulation of interstitial carbon at the pleural surface associated with significant lung surface carbon.

Conclusion
This study found that image analysis quantifies accumulation of carbon PM at the lung surface. We disproved our hypothesis that smoking is a major determinate of lung surface carbon. Although the most likely source of lung surface carbon is environmental PM, we could not identify an association with markers of exposure.
Supporting information S1 Table. Full table of