Chemosphere 83 (2011) 618–626

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

In vitro biological effects of airborne PM2.5 and PM10 from a semi-desert cityon the Mexico–US border

Alvaro R. Osornio-Vargas a,g,⇑, Jesús Serrano b, Leonora Rojas-Bracho c, Javier Miranda d,Claudia García-Cuellar a, Marco Antonio Reyna e, Geraldine Flores a, Miriam Zuk c, Margarito Quintero e,Inés Vázquez a, Yesennia Sánchez-Pérez a, Tania López c, Irma Rosas f

a Investigación Básica, Instituto Nacional de Cancerología, México, DF 14080, Mexicob Facultad de Ciencias, Universidad Nacional Autónoma de México, México, DF 04510, Mexicoc Instituto Nacional de Ecología, SEMARNAT, México, DF 04350, Mexicod Instituto de Física, Universidad Nacional Autónoma de México, México, DF 04510, Mexicoe Instituto de Ingeniería, Universidad Autónoma de Baja California, Mexicali, Baja California 21280, Mexicof Centro de Ciencias de la Atmósfera, Universidad Nacional Autónoma de México, México, DF 04510, Mexicog Department of Paediatrics, University of Alberta, Edmonton, Alberta, Canada T6G 2V2

a r t i c l e i n f o a b s t r a c t

Article history:Received 23 July 2010Received in revised form 8 November 2010Accepted 25 November 2010Available online 18 December 2010

Keywords:PM10

PM2.5

MexicaliCytotoxicityPrincipal components analysisElemental composition

0045-6535/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2010.11.073

⇑ Corresponding author at: Department of PaediatDentistry, University of Alberta, 8308 114 St., NW, Ed2V2. Tel.: +1 780 492 7092; fax: +1 780 407 1982.

E-mail address: osornio@ualberta.ca (A.R. Osornio

Compelling evidence indicates that exposure to urban airborne particulate matter (PM) affects health.However, how PM components interact with PM-size to cause adverse health effects needs elucidation,especially when considering soil and anthropogenic sources. We studied PM from Mexicali, Mexico,where soil particles contribute importantly to air pollution, expecting to differentiate in vitro effectsrelated to PM-size and composition. PM samples with mean aerodynamic diameters 62.5 lm (PM2.5)and 610 lm (PM10) were collected in Mexicali (October 2005–March 2006) from a semi-urban (expectedlarger participation of soil sources) and an urban (predominately combustion sources) site. Samples werepooled by site and size, analyzed for elemental composition (particle-induced X-ray emission) and testedin vitro for: induction of human erythrocytes membrane disruption (hemolysis) (colorimetrically); inhi-bition of cell proliferation (ICP) (crystal violet) and TNFa/IL-6 secretion (ELISA) using J774.A1 murinemonocytic cells; and DNA degradation using Balb/c3T3 cell naked DNA (electrophoretically). Results ofPM elemental composition principal component analysis were used in associating cellular effects. Sixteenelements identified in PM grouped in two principal components: Component1 (C1): Mg, Al, Si, P, Cl, K, Ca,Ti, V, Cr, Fe, and Component2 (C2): Cu, Zn. Hemolysis was predominately induced by semi-urban-PM10

(p < 0.05) and was associated with urban-PM10 C1 (r = 0.62, p = 0.003). Major ICP resulted with semi-urban PM2.5 (p < 0.05). TNFa was mainly induced by urban samples regardless of size (p < 0.05) and asso-ciated with urban-PM2.5 C2 (r = 0.48, p = 0.02). Both PM10 samples induced highest DNA degradation(p < 0.05), regardless of location. We conclude that PM-size and PM-related soil or anthropogenic ele-ments trigger specific biological-response patterns.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Human exposure to particulate matter (PM) is associated with arange of diseases, including bronchitis, exacerbation of asthma,and cardiovascular disease (Pope and Dockery, 2006). The impactof PM on health may be associated with its physical and chemicalproperties, including mass, size, number, surface area andcomposition (Schwarze et al., 2006, 2007). PM is a complex

ll rights reserved.

rics, Faculty of Medicine andmonton, Alberta, Canada T6G

-Vargas).

heterogeneous mixture of organic, inorganic and biological com-pounds that result from various sources like primary emissions,secondary photochemical reactions in the atmosphere andre-suspension of particles from natural sources or road dust(Osornio-Vargas et al., 2003; Pope and Dockery, 2006).

PM-size has been the most available metric used in epidemiol-ogy to assess PM-related health effects. However, there is still con-troversy on which PM-size fraction has the stronger potential toimpact health. A recent review of the epidemiological literatureon the health impacts of exposure to PM revealed no major differ-ences between morbidity outcomes from short-term exposures tocoarse particles (PM10–2.5) or fine particles (PM2.5) (Brunekreef andForsberg, 2005). Cohort studies have not identified an association

A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626 619

between exposure to PM10–2.5 and mortality (Pope and Dockery,2006). However, multiple time-series epidemiological studies haveshown associations between mortality and PM2.5, and betweenmortality–morbidity and particles from mobile source emissionsand carbon combustion processes (Dominici et al., 2005; Popeand Dockery, 2006; Franklin et al., 2008; Peng et al., 2008). Othertime-series studies have found stronger associations between total,respiratory, and cardiovascular mortality and PM10–2.5 or inhala-ble-PM (PM10) composed predominantly of geological elements,than with PM2.5 alone or PM10 with �60% of PM2.5 (i.e. mostly asso-ciated with combustion processes) (Ostro et al., 1999; Castillejoset al., 2000; Cifuentes et al., 2000; Mar et al., 2000). Recently, atten-tion has focused on the effects of PM10 and PM10–2.5 from regionswhere anthropogenic and soil PM sources co-exist as a result ofdust storms and synergistically impact more on human health(Cheng et al., 2008; Perez et al., 2008). Experimental studies havealso identified PM10 and PM10–2.5 to be more toxic than PM2.5,probably as a result of PM-size/composition interactions, ratherthan of size alone (Osornio-Vargas et al., 2003; Veranth et al.,2004; Schwarze et al., 2006, 2007; Duvall et al., 2008; Valavanidiset al., 2008; Tong et al., 2010). Nevertheless, more epidemiologicaland toxicological studies are required to fully address the relativeparticipation of PM-size and PM components from various sources,including soil (Viau et al., 2010), as determinants of PM-relatedbiological effects.

To study the relative roles of PM elemental components in pro-ducing health risks, we analyzed the in vitro induced biological ef-fects of PM2.5 and PM10 samples collected from Mexicali, Mexicowhere PM is a mixture produced from anthropogenic and soilsources. Samples were collected at two sites: one influenced pre-dominantly by soil-related sources and another by combustionsources. In vitro biological effects were performed in three differentbiological systems: (1) human red blood cells to evaluate PM-induced membrane disruption potentially induced by geologicalsoil components (Osornio-Vargas et al., 1991); (2) the monocyticJ774A.1 cell line to evaluate PM-induced cytotoxicty and cytokineproduction as previously done by us in other PM studies(Osornio-Vargas et al., 2003) and (3) isolated DNA to explore directPM-induced DNA degradation mainly attributable to metal-relatedoxidative damage (García-Cuellar et al., 2002). Elemental and tox-icological characteristics of PM air samples were analyzed tounderstand PM toxicological profile and to investigate the relativeroles of the soil components.

2. Material and methods

2.1. Study setting

Mexicali is a city of 856 000 inhabitants in Baja California, asemi-desert region bordering the United States that is frequentlyaffected by dust storms. The local emission inventory reports thatmore than 77% of PM10 are generated from fugitive dust (unpavedroads and empty lots). The main components of PM10 are geologi-cal materials (>70% of total mass), followed by direct emissionsfrom vehicles (�12%), field burning (�6%), marine aerosols (<3%)and secondary sulfates. Annual average concentrations of PM10

are 80% higher than the annual standard value (50 lg m�3) andthe 98th percentile of 24 h average PM10 concentrations exceedmore than twice the current 24 h standard (120 lg m�3) (Zuket al., 2007).

2.2. PM Sampling

Airborne PM samples were collected at two monitoring sta-tions: the semi-urban at Colegio Nacional de Educación Profesional

Técnica (CONALEP) and the urban at Universidad Autónoma deBaja California (UABC) (Fig. 1). PM2.5 and PM10 were sampled dur-ing 24 h periods on alternate days at least twice a week fromOctober 2005 to March 2006. High Volume (Tisch TE6070V,Roswell, GA, USA) samplers with modified nitrocellulose mem-branes (Sartorious 11302-131, Satelite, EdeM, Mexico) were usedto obtain adequate amounts of PM (Alfaro-Moreno et al., 2009).PM samples were mechanically recovered from the membranesand pooled by month, site, and particle size for further use andanalysis. Elements were characterized in three sub-samples ofthe pooled sample. Toxicological evaluation was conducted usingaliquots from pooled samples in three separate experiments, eachrun in triplicate. Endotoxins were determined in duplicate inpooled samples from each month.

2.3. Elemental analysis by particle-induced X-ray emission

One milligram samples were ground in an agate mortar and ad-hered to 3.5 lm thick Mylar substrates. Samples were analyzedwith PIXE using a 2.2 MeV proton beam produced by a 9SDH-2Pelletron accelerator (National Electrostatics Corporation, Middle-ton, WI, USA), to identify and quantify trace elements from alumi-num to uranium.

Characteristic X-rays were received with a LEGe detector(Canberra Industries, Meridien, CT, USA), with 150 eV to 5.9 keVof resolution using the device described by Miranda et al. (2000).A 12 lm thick Mylar window separated the chamber vacuum fromthe X-ray detector. The proton beam was 5 mm in diameter, withenergy of 2.2 MeV and a beam current of 15 nA. The total inte-grated charge for each filter spectrum was 5 lC. A PCA3 multichan-nel analyzer (Oxford Industries-Tennelec, Oak Ridge, TN, USA),attached to a personal computer, was used to collect the X-rayspectra that were subsequently analyzed by means of the Quanti-tative X-ray Analysis System (QXAS) (IAEA, 1995). The detectionsystem response was measured using thin film MicroMatter stan-dards (Deer Harbor, WA, USA). A determination of the accuracy forPIXE analyses was performed using a different set of MicroMatterstandards, resulting in an error not larger than 0.7% for the ele-ments detected in this study. Measured elemental concentrationsare expressed in units of lg mg�1.

2.4. Endotoxin determination

Endotoxins were evaluated by a kinetic chromogenic LimulusAmebocyte Lysate (LAL) assay (BioWhittaker, Inc.; Walkersville,MD, USA) on monthly urban and semi-urban-PM10 and PM2.5 sam-ples in duplicate. All samples were resuspended in pyrogen-free50 mM Tris buffer (Cambrex Bio Science Walkersville, Inc.; Walk-ersville, MD, USA) using an ultrasound water bath. Five serial 1:4dilutions from each suspension were prepared. Standard endotoxinEscherichia coli O55:B5 (BioWhittaker, Lot. 2L1850, Walkersville,MD, USA), with a potency of 10 EU ng�1, was used as reference.

Limulus reactions were monitored in a kinetic micro-plate read-er (Kinetic-QCL Reader; BioWhittaker, Inc., Walkerville, MD, USA)at 37 �C. The optical density (OD) of each well was recorded at awavelength of 405 nm every 150 s. The micro-plate reader wascontrolled and a Gateway 450 PC-XT computer recorded data.The results were expressed as lg mg�1 (Rosas Perez et al., 2007).

2.5. In vitro evaluation of PM-induced biological effects

Various tests were conducted to determine in vitro biological ef-fects of PM collected in Mexicali. Hemolysis was used as a toxicityindicator of mineral particle-induced cell membrane disruption;inhibition of cell proliferation (ICP) as an indicator of cell death; tu-mor necrosis factor-alpha (TNFa) and interleukin-6 (IL-6) secretion

Fig. 1. The locations of the six operating air monitoring stations in Mexicali, Mexico. Locations circled in pink represent the sites where PM10 and PM2.5 used in this studywere collected. CONALEP = semi-urban and UABC = urban.

620 A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626

as indicators of pro-inflammatory processes, and DNA degradationas an indicator of genotoxicity.

2.5.1. HemolysisHuman red blood cells were obtained from three ‘healthy do-

nors’ whole blood by centrifugation at 475g/10 min at 4 �C in130 mM potassium chloride (KCl), 20 mM Tris buffer, pH 7.4. Cellsuspensions (0.6%) were incubated with 20–160 lg mL�1 of PMsamples, using the same buffer. After 60 min at 37 �C, sampleswere centrifuged at 475g/10 min at 4 �C to read in a spectropho-tometer for released hemoglobin in the supernatant at 540 nm. Re-sults are expressed as percent hemolysis in relation to thehemoglobin released by the same amount of red blood cells fromeach donor incubated with double-distilled water.

2.5.2. Inhibition of cell proliferationCrystal violet method was used to assess inhibition of cell pro-

liferation using J774A.1 murine macrophage cells (ATCC, VirginiaUSA) under log-phase growth. Cells were plated in 96-well platesat 5000 cells/well in 10% fetal bovine serum (FBS) – Dulbecco Mod-ified Eagle Medium (DMEM) and exposed to PM (20, 40, 80 and160 lg cm�2) in DMEM for 72 h. Effects were determined by mea-suring the residual cell number by crystal violet staining with anELISA plate reader at 595 nm (Multiscan MS 352; Lab-Systems,Helsinki, Finland). The percentage of inhibition of cell proliferationwas expressed as percentage of the retained color by remaining at-tached cells in relation to non-exposed cells (Rosas Perez et al.,2007).

2.5.3. Secretion of cytokines by monocytic cells in cultureSecretion of TNFa and IL-6 was measured in the supernatants of

confluent J774A.1 cells exposed to non-lethal concentrations of PM

(10, 20 and 40 lg cm�2) using ELISA kits (DY410E and DY406EDuoset ELISA Development System, R&D Systems). Cells weregrown to confluence in 10% FBS-DMEM, and then changed toserum-free DMEM for 24 h before exposure to PM in serum-free DMEM. After 24 h cell supernatants were collected, centri-fuged at 14 000g for 15 min and frozen at �70 �C until analyzedby ELISA (Osornio-Vargas et al., 2003; Rosas Perez et al., 2007).Unexposed cells were used as negative controls and cells exposedto 10 lg mL�1 lipopolysaccharide (LPS) from Escherichia coli055:B5 (Sigma), as positive controls. Similar experiments weredone in the presence of 2 lg mL�1 recombinant neutralizing endo-toxin protein (rENP) (Saikagaku America). Results are expressed inpg mL�1.

2.5.4. Induction of DNA degradationDegradation of ‘‘naked’’ Balb/c3T3 cell DNA was determined

after obtaining DNA with a commercially available DNA extractionkit (Roche 11814770001). These sets of experiments were designedto explore biological effects mediated by metal induced oxidativestress damage since they were done in the presence of H2O2 andin the presence or absence of Deferoxamine (DFO). Four hundrednanograms of isolated DNA was exposed to 40, 80 and 160 lg mL�1

of PM in the presence or absence of 1 mM hydrogen peroxide(H2O2) and in the presence or absence of 1 mM DFO during 24 hat 37 �C. The analysis was performed by electrophoresis (Horizon58; GIBCO-BRL Gaithersburg, MD, USA) in 1.5% agarose gels at100 V for 3 h, stained with ethidium bromide (1.2 lg mL�1) andphotographed under ultraviolet light in a Kodak Gel logic 200imaging system (New Haven, CT, USA). Controls included: (1)phage k/Hind III (1 lg), (2) 400 ng of DNA alone, (3) 400 ng ofDNA with 1 mM H2O2, (4) 400 ng of DNA with 1 mM DFO and (5)‘‘Fenton Reaction’’ using 400 ng of DNA plus 5 mM copper sulfate

A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626 621

(CuSO4) with 1 mM H2O2. Quantitative analysis of degradation wasdone using the Kodak Molecular Imaging software version 4.X CD(New Haven, CT, USA). DNA band optical density was determinedusing a fixed rectangular area of 200 pixels. Results are expressedas ng of DNA after comparing densities from the PM-treated DNAbands with the densities measured from the correspondent un-treated DNA bands.

3. Data analyses

3.1. PM composition

Descriptive statistical analyses were conducted for elementsidentified in PM samples by Site and PM-size: semi-urban-PM2.5,semi-urban-PM10, urban-PM2.5 and urban-PM10 (Site � PM-sizeinteractions). Principal component analysis (PCA) was used to sim-plify the description of the complex elemental composition and forfurther inferential analysis, except for endotoxins, which were ana-lyzed separately. Variables were transformed to reduce skew. Afirst PCA analysis was done using the data from all samples. Theresultant scored components were used in an exploratory factorialanalysis of variance (F-test) to evaluate differences in concentra-tions by Site (semi-urban vs. urban; Site effect), by PM-size (PM2.5

vs. PM10; size effect), and by Site � PM-size interactions (interac-tion effect). A second PCA analysis was applied to evaluate inmore detail the elemental composition structure according toSite � PM-size data and explore associations with cell effects. Com-ponents with eigenvalues over 1 were extracted (Rosas Perez et al.,2007).

3.2. Enrichment factors

Enrichment factors (EF) were computed to estimate the origin(soil vs. anthropogenic) of the elements present in each of the ex-tracted principal components (Paredes-Gutiérrez et al., 1997). TheEF for every element and sample was determined using referencedcomposition of geological PM10 from the same region (Watson andChow, 2001). Si was chosen as a reference element as its presenceis expected in all soil-derived particles. The EF were defined usingthe following equation:

EF ¼ ðCz=CSiÞS=ðCZ=CSiÞRef ð1Þ

Here, CZ is the concentration of element Z, while the subscripts Sand Ref refer to the sample and to the reference composition,respectively. When the EF is the order of unity it is expected thatthe element is related to a soil source; otherwise, its origin maybe anthropogenic.

3.3. Biological effect tests

Biological effects associated to Site, PM-size and interactions ofSite � PM-size, were analyzed by an exploratory factorial analysisof variance (transformations were applied when necessary to correctconventional assumptions with analysis of variance). Further com-parative tests were done accordingly to obtained results (t-test orcontrasts). Associations between biological effect variables and prin-cipal component scores extracted from each Site� PM-size subset ofdata were assessed by Spearman’s correlation coefficients. Correla-tion analysis was focused on responses obtained with: 80 lg mL�1

for hemolysis; 80 lg cm�2 for ICP; 40 lg cm�2 for TNFa and IL-6 pro-duction; and 160 lg mL�1 for DNA degradation.

4. Results

4.1. PM composition

PM from Mexicali was characterized by sixteen elements: Mg, Al,Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, and Zn. These elements wereidentified in all monthly samples from both sampling sites and PM-size fractions, except for Ni that was absent in some monthly samplesand was therefore eliminated from consideration in subsequentanalyses. Elements with highest concentrations were Si, Ca, Al, Feand K (means > 1 lg mg�1) (�84% of elemental mass); in contrast,Mg, Cl, Ti, V, P, Cr, Mn, Ni, Cu and Zn presented lower concentrations(means < 1 lg mg�1) (�16%). Average total element concentrations(lg mg�1) were: 13.13 (semi-urban-PM10), 22.03 (semi-urban-PM2.5),17.2 (urban-PM10) and 14.19 (urban-PM2.5) (Table 1). Averageendotoxin concentrations (EU mg�1) were: 63.99 ± 29.02 (semi-ur-ban-PM10), 111.41 ± 150.59 (semi-urban-PM2.5), 71.04 ± 59.05 (urban-PM10) and 36.93 ± 35.69 (urban-PM2.5). No statistically significantdifferences in endotoxin concentration by Site, PM-size, or Site� PM-size interaction were found (F-test; p > 0.05).

Two principal components (C1, C2) were extracted which to-gether explained 87% of total variance. Elements present in C1

were: Mg, Al, Si, P, Cl, K, Ca, Ti, V, Cr, Mn, and Fe, while C2 includedCu and Zn. Interestingly, S presented a high correlation with bothcomponents.

Averages of C1 scores were not significantly different betweenSites, PM-size fractions, or for Site� PM-size interaction. By contrast,C2 scores showed a significant Site effect, with lower mean levels atthe semi-urban site, which indicates lower Cu and Zn concentra-tions in the semi-urban environment.

It was found that most of the elements in C1 have EF close tounity, indicating a geological origin. However, V (urban: 34.6 ±4.9; semi-urban: 31.5 ± 4.2) and Cr (urban: 44.0 ± 10.0; semi-ur-ban: 42.8 ± 8.6) also present within C1 had larger EFs, suggestingthe participation of both geologic and anthropogenic sources. Cuand Zn present in C2 had EFs > 1, indicating an anthropogenic ori-gin. Sulfur EFs varied from sample to sample suggesting a mixedparticipation of soil and anthropogenic sources, probably explain-ing S correlations in C1 and C2 (Table 2).

The second PCA, conducted with subsets of data by Site �PM-size, generally corroborated that the components structuredescribed above was retained among samples. The components ex-tracted from semi-urban-PM10, semi-urban-PM2.5, and urban-PM10

samples were very similar to the general pattern extracted previ-ously (Fig. 2A, B and D, respectively). However the urban-PM2.5

components structure showed two differences: (1) C2 includedmore elements (P, S, K, Cu, and Zn) and (2) a third component(C3) associated to Cl was also identified (Fig. 2C).

4.2. Biological effects

Biological responses to varying PM concentrations consistentlygave positive dose–response curves. IL-6 results obtained with sam-ples collected in November are presented in Fig. 3 as an example.Maximal response doses used in subsequent experiments were:80 lg mL�1 for hemolysis; 80 lg cm�2 for ICP; 40 lg cm�2 for TNFaand IL-6 production; and 160 lg mL�1 for DNA degradation.

Results are presented as distribution plots using the values ob-tained with monthly-pooled samples by site and size as previouslymentioned.

4.2.1. HemolysisHemolysis outcomes were significantly different by Site, PM-Size

and for the Site � PM-size interaction, with highest esponse afterexposure to semi-urban-PM10 samples (semi-urban-PM10 >

Table 2PM mean enrichment factors (EF) for both sampling sitesa.

Element Urban Semi-urban

Mean SDb Mean SDb

Mg 2.9 0.09 2.9 0.22Al 1.1 0.01 1.1 0.01Si 1.0 0.00 1.0 0.00P 34.1 4.50 41 12.00S 7.8 1.90 5.1 0.97Cl 2.6 0.30 2.4 0.13K 3.6 0.42 3.3 0.29Ca 2.3 0.03 2.3 0.07Ti 3.1 0.18 2.8 0.15V 34.6 4.90 31.5 4.20Cr 44 10.00 42.8 8.60Mn 10.1 3.80 3.5 0.50Fe 3.1 0.39 2.20 0.19Ni 61 12.00 53 11.00Cu 203 68.00 91 25.00Zn 175 82.00 35.5 10.00

a Elemental composition of Mexicali ‘‘Bulk Dirt Composite’’ from Watson andChow (2001) was used as reference.

b SD = standard deviation.

Table 1Descriptive statistics for elemental concentration in PM samples.A

Element SUa-PM2.5b SU-PM10

c Ud-PM2.5 U-PM10

Mean % eSD fCOV Mean % SD COV Mean % SD COV Mean % SD COV

Elemental concentrations (lg mg�1)Si 7.93 35.98 6.12 77.26 4.53 34.47 3.75 82.85 3.88 27.31 3.21 82.74 5.55 32.30 5.67 102.05Ca 4.72 21.44 3.83 81.12 2.92 22.28 2.36 80.72 2.59 18.22 1.98 76.55 3.91 22.74 4.21 107.56Al 2.83 12.84 2.16 76.51 1.65 12.56 1.33 80.80 1.41 9.96 1.15 81.11 2.04 11.89 2.07 101.36Fe 1.94 8.83 1.41 72.76 1.13 8.57 0.93 82.85 1.57 11.04 1.29 82.64 1.60 9.33 1.48 92.15K 1.72 7.83 1.18 68.63 1.05 8.03 0.75 71.08 1.88 13.25 0.98 52.25 1.33 7.71 1.22 91.89S 0.67 3.04 0.35 53.07 0.44 3.36 0.25 55.73 1.13 7.97 0.67 59.64 0.72 4.20 0.56 77.01Mg 0.65 2.95 0.47 72.22 0.41 3.14 0.27 65.74 0.31 2.18 0.24 77.14 0.57 3.32 0.53 92.10Cl 0.69 3.13 0.49 71.45 0.39 2.94 0.34 88.43 0.54 3.83 0.45 83.17 0.40 2.32 0.36 90.99Ti 0.29 1.30 0.27 95.46 0.18 1.37 0.18 99.78 0.19 1.34 0.15 79.67 0.24 1.41 0.28 117.80P 0.22 1.02 0.11 49.85 0.15 1.14 0.06 40.44 0.22 1.53 0.07 32.44 0.21 1.19 0.17 83.25V 0.13 0.60 0.17 128.18 0.10 0.73 0.11 114.42 0.10 0.71 0.08 84.05 0.13 0.78 0.20 151.30Cr 0.09 0.42 0.15 159.25 0.08 0.60 0.10 121.86 0.07 0.50 0.08 106.70 0.12 0.68 0.21 178.48Zn 0.05 0.23 0.04 87.56 0.04 0.29 0.03 89.75 0.16 1.16 0.12 72.63 0.18 1.04 0.26 145.36Mn 0.04 0.17 0.05 130.00 0.03 0.25 0.03 98.70 0.05 0.36 0.03 65.45 0.07 0.39 0.09 126.55Cu 0.04 0.16 0.04 109.57 0.02 0.19 0.02 89.93 0.08 0.58 0.06 70.70 0.09 0.51 0.08 89.46Ni 0.01 0.06 0.01 115.56 0.01 0.09 0.01 69.58 0.01 0.05 0.01 146.93 0.03 0.18 0.06 178.51

Sum 22.03 100.00 13.13 100.00 14.19 100.00 17.20 100.00

A Elements are listed in decreasing concentrations; n = 18 by Site and PM-size.a SU = semi-urban.b PM2.5 = particulate matter with aerodynamic diameter equal or less than 2.5 lm.c PM10 = particulate matter with aerodynamic diameter equal or less than 10 lm.d U = urban.e SD = standard deviation.f COV = coefficient of variation.

622 A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626

urban-PM10; t-test, p < 0.05). In general, PM10 induced strongereffects (Fig. 4A). Hemolysis was positively associated with the C1

from urban-PM10 (r = 0.623, p = 0.003).

4.2.2. Inhibition of cell proliferationPM-size and the interaction of Site � PM-size were significantly

different (F-test, p < 0.05), with a stronger inhibition for exposuresto semi-urban-PM2.5 than to the others (semi-urban-PM2.5 > urban-PM10 and semi-urban-PM10 contrast, p < 0.05) (Fig. 4B). No signifi-cant associations between principal components and inhibitionof cell proliferation were detected (p > 0.05).

4.2.3. Secretion of cytokines by monocytic cells in cultureFor TNFa, a significantly stronger response was found for cells

exposed to PM from the urban site in contrast to those collected

at the semi-urban site (Site effect; F-test, p < 0.05). There were nosignificant differences of TNFa secretion between PM-size fractionsor for the interaction Site� PM-size (Fig. 4C). In addition, TNFa wasassociated with C2 from urban-PM2.5 (r = 0.48, p = 0.02). Secretionof TNFa in the presence of the endotoxin inhibitor rENP inhibitedTNFa secretion (�22%), with no changes in the pattern attainedin its absence (F-test, p > 0.05).

For IL-6, high variation was observed for all samples and no sig-nificant differences were found by Site, PM-size or the interactionSite� PM-size (Fig. 4D). Secretion of IL-6 was negatively associatedwith C1 from semi-urban-PM10 (r = �0.697; p = 0.001). rENP inhib-ited IL-6 secretion (�9%), with no changes in the pattern attainedin its absence (F-test, p > 0.05).

No cytotoxic effects were observed with any of the sampledoses used in these experiments.

4.2.4. Induction of DNA degradationPM10 samples were significantly stronger inducers of DNA deg-

radation than PM2.5 (PM-size effect; F-test, p < 0.05) (Fig. 4E). Ingeneral, DFO inhibited DNA degradation (F-test, p < 0.05). Therewere no significant associations between DNA degradation andany of the principal components.

A summary of major PM-induced toxic effects and their associ-ations with PM characteristics can be seen in Fig. 5.

5. Discussion

Airborne PM collected at two sites in Mexicali contains ele-ments from soil mixed with elements from anthropogenic combus-tion sources. These elements are grouped into two principalcomponents. The first component had elements from the soil (Al,Si, Ca) (EF � 1) and elements from anthropogenic sources (V, Cr)(EF > 1). Elements in this component did not vary significantly bySite (urban vs. semi-urban) or PM-size fraction (PM10 or PM2.5).PM10 and PM2.5 samples from the urban site presented higher con-centrations of Cu and Zn, suggesting the influence of emissionsfrom anthropogenic sources (Miranda et al., 1994; Moreno et al.,

Fig. 2. Principal component analysis for Site � PM-size showed a similar structure for: (A) semi-urban PM2.5 (SU-PM2.5), (B) semi-urban-PM10 (SU-PM10), and (D) urban-PM10

(U-PM10). In the case of urban PM2.5 (U-PM2.5) (C). Component2 included P, S, K, Cu, and Zn, and a third component was identified composed by Cl.

Fig. 3. Dose response curves were observed for all biological effects after exposureto various concentrations of PM10 and PM2.5 from semi-urban and urban sites. Weare presenting the example of IL-6 production by J774A.1 cells exposed to 10, 20and 40 lg cm�2 of PM samples collected in November 2005.

A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626 623

2008)). Furthermore, urban-PM2.5 had a unique and more complexC2, and the existence of a third component, C3, was revealed afterPCA.

Our results identify different biological effect potentials forPM10 and PM2.5 samples, and also different profiles betweensampling sites (cell membrane disruption by semi-urban vs. pro-inflammatory by urban). These results indicate that the soil

components could explain the hemolytic effects and that theanthropogenic components could explain the pro-inflammatoryand ICP effects. The results also suggest that some effects, likeendotoxin-mediated cytokine production and DNA degradation,depend on complex interactions among PM components and PM-size.

Previous work from our group have also identified characteris-tic patterns of in vitro biological response specifically related to PMsamples from various regions of Mexico City (Alfaro-Moreno et al.,2002; Osornio-Vargas et al., 2003; Rosas Perez et al., 2007). The re-sults suggested that biological responses were driven by the over-all complexity of PM collected in specific sites rather than by aspecific component of the mixture, as also observed by others(Viau et al., 2010). However, the lack of information on the vari-ability of PM composition prevented further statistical exploration.In the present paper, extensive PM sampling gave us the opportu-nity to explore by PCA the relative participation of PM componentson specific site PM mixtures and related patterns of PM-inducedbiological effects, using the same cell systems as before. Compara-tive results using Mexicali soil samples, Mexico City’s PM10 andPM2.5 and current Mexicali airborne PM showed three clearly de-fined patterns: (1) a stronger hemolysis response was associatedwith soil samples; (2) lowest induction of ICP and higher DNA deg-radation were associated with Mexico City’s airborne PM samplesand (3) Mexicali samples induced intermediate hemolysis, strongcytokine production and ICO and, intermediate DNA degradation(data not shown).

PCA has been recently used to explore relationships betweenPM composition and in vitro PM-induced cell responses. Mostly,pro-inflammatory effects have been linked to specific groups of

Fig. 4. Distribution plots of biological effects attained with PM samples from Mexicali. (A) Hemolysis was predominantly induced by PM10, and more importantly by SU-PM10

samples (SU-PM10 > U-PM10; t-test, p < 0.05). (B) ICP was importantly induced by SU-PM2.5 (SU-PM2.5 > U-PM10 and SU-PM10, contrast, p < 0.05). (C) Samples from the urbansite induced larger TNFa release than PM from the semi-urban site (Site effect; F-test, p < 0.05). (D) No differences in IL-6 secretion were observed with any of the samples. (E)DNA degradation was induced only by PM10 (PM-size effect; F-test, p < 0.05). All shown values are before transforming and represent the distribution of biological responsesobtained with pooled monthly samples by size and site of collection. Figures represent minimal value, lower quartile, median, upper quartile and maximal value. SU = semi-urban; U = urban; � = significant differences (p < 0.05), s = outliers.

624 A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626

Fig. 5. Statistically significant differences by Site, PM-size, and Site � PM-sizeinteractions.

A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626 625

PM components. For example, PM10–2.5 from Chapel Hill, NC,showed stronger seasonal effects on IL-6/IL-8 production by nor-mal human epithelial cells related to Cr/Al/Si/Ti/Fe/Cu content(Becker et al., 2005). PM2.5 isolated from various soil samples alsoshowed important variability in association with IL-6 release byBEAS-2B cells. Effects were associated with PM2.5 content of pyro-lyzable carbon fractions and crustal elements defined by their alu-minum to silicon ratio (Veranth et al., 2006). A previous studyconducted in our laboratory using airborne PM from Mexico Cityalso indicated that the three PM-identified principal componentscould induce a defined cell response pattern. Our study suggestedthat pro-inflammatory effects were the result of complex interac-tions between these three principal components, whereas cyto-toxic effects were linked to S/K/Ca/Ti/Mn/Fe/Zn/Pb content (RosasPerez et al., 2007).

Although these studies provide some evidence regarding therole of PM components and their associated cell effects, there isstill no consensus within the scientific community as to which spe-cific components are the most significant determinants of the cel-lular response. High variability of PM components is associatedwith PM sources, climate conditions (seasonality) and to the meth-ods used in each study (Becker et al., 2005; Lippmann et al., 2006;Duvall et al., 2008; Perrone et al., 2010). Also, these studies may beaddressing a fraction of the total PM components or componentsthat may be serving as indicators/proxies of the other componentsnot studied in a complex mixture, given that comprehensive char-acterization of PM composition is a major task. For example, theobservation by Veranth et al., 2006 that the soil derived PM2.5 frac-tion induced IL-6 release related to its carbon content and ourobservation that Mexicali PM induced IL-6/TNFa secretion wasstrongly related to Cu/Zn content, suggests a major contributionby elements from anthropogenic sources to the PM2.5-related ef-fects. In addition, the stronger effects induced by PM10–2.5 fromChapel Hill, PM10 from Mexico City (Osornio-Vargas et al., 2003),or even by fine particles from Milan (Perrone et al., 2010), seemto be related to PM components in which elements from the soiland from anthropogenic sources co-exist, probably as an indicationof particle re-suspension from mixed sources. The co-variation ofelements like S, V, and Cr in the C1 identified in the present study,support the latter. Additional PM constituent characterization (e.g.organics) is required before complex PM-composition/PM-effectsproblems can be fully understood.

In the case of Mexicali, soil particle re-suspension seems to af-fect both PM fractions, with major impacts on PM10. This is impor-tant since in 1986, a young woman in Mexicali contracted and diedfrom pulmonary fibrosis as a result of inhalation of soil particles(aluminum-silicates). Experimental studies demonstrated soil-related toxic effects, including the induction of hemolysis in vitroand the generation of pulmonary lesions in exposed animals, sim-ilar to the ones originally described in the Mexicali patient(Osornio-Vargas et al., 1991).

Our current results indicate that PM10 from Mexicali retained animportant in vitro hemolytic potential, but hemolysis effects werenot observed in Mexicali PM2.5. Hemolytic activity has been linkedto particle surface activity and pulmonary fibrogenic potential(Warheit et al., 2007), corroborating our original observation onthe potential of particles from Mexicali to induce pulmonary fibro-sis. The lack of hemolytic potential of PM2.5 samples provides animportant contrast in the biological effects pattern, giving supportto recent observations about the probable underestimation of theinfluence of coarse fraction particles (PM10–2.5) on pulmonary mor-bidity (Brunekreef and Forsberg, 2005; Alexis et al., 2006). The ob-served differences in the effects induced by PM2.5 and PM10

provides further indication that each fraction retains a specificcomposition profile and toxic potential in spite of the factthat PM2.5 is contained in PM10, as also previously reported(Osornio-Vargas et al., 2003; Mendoza et al., 2010). Taken togetherthese results provide further evidence that PM of different compo-sition could lead to varying health outcomes as has been proposedelsewhere (Steerenberg et al., 2006; Franklin et al., 2008; Mannet al., 2010).

In particular, our results support the hypothesis that PM withhigher participation of elements from the soil is predominantlyassociated with cell membrane disruption that in combinationwith pro-inflammatory effects could participate in severe healthprocesses like pulmonary fibrosis. In contrast, PM with largeranthropogenic participation may have a higher impact on lungand systemic cardiovascular diseases, due to its stronger potentialto induce the production of pro-inflammatory cytokines. However,in vitro tests alone provide limited insight into mechanisms ofpotential health outcomes, requiring confirmation from in vivotests and epidemiological research. Furthermore, while the twogroupings of 16 elements provided useful information about thedifferential effects of PM composition and sources, additional com-pounds need to be analyzed (e.g. organics, ions, elemental and or-ganic carbon, etc.) to enrich the understanding of the toxicologicaleffects of PM composition. Nonetheless, a recent epidemiologicalstudy using a more comprehensive PM composition analysis, sug-gests the existence of complex interactions between soil particlecomponents and those from anthropogenic sources that result inhigher mortality risk (Perez et al., 2009).

Our study indicates that local air quality management effortsshould aim to reduce not only PM2.5 and PM10 from anthropogenicbut from soil sources as well, since soil components participateimportantly in observed biological effects. Special interest shouldbe placed on fugitive dust since no regulations exist in Mexico.

This study supports growing evidence that PM-size, composi-tion and origin interact in a complex manner to determine theimpact of PM on human health. Further toxicological and epidemi-ological research on the impacts of PM-size and composition isessential to better understand these phenomena and direct andprioritize pollution and dust control efforts.

Acknowledgments

This work was supported by InterGen-LASPAU (Border OzoneReduction and Air Quality Improvement Program); National Insti-tute of Ecology, and the Mexican Science and Technology Council

626 A.R. Osornio-Vargas et al. / Chemosphere 83 (2011) 618–626

[CONACyT-43138M]. We thank J. Landeros for PM-sampling;E. Salinas/L. Martínez for LPS determination; K. López/F. Jaimesfor PIXE analysis; L. Sevilla for graphical assistance; I. Buka,P. Schwarze, F. Balleste, J.C. Bonner, M.S. O’Neill and Liu Wei forcritical comments and reviews.

References

Alexis, N.E., Lay, J.C., Zeman, K., Bennett, W.E., Peden, D.B., Soukup, J.M., Devlin, R.B.,Becker, S., 2006. Biological material on inhaled coarse fraction particulatematter activates airway phagocytes in vivo in healthy volunteers. J. Allergy Clin.Immunol. 117 (6), 1396–1403.

Alfaro-Moreno, E., Martínez, L., García-Cuellar, C., Bonner, J.C., Murray, J.C., Rosas, I.,Ponce de León Rosales, S., Osornio-Vargas, A.R., 2002. Biologic effects inducedin vitro by PM10 from three different zones of Mexico City. Environ. HealthPerspect. 110 (7), 715–720.

Alfaro-Moreno, E., Torres, V., Miranda, J., Martinez, L., Garcia-Cuellar, C., Nawrot,T.S., Vanaudenaerde, B., Hoet, P., Ramirez-Lopez, P., Rosas, I., Nemery, B.,Osornio-Vargas, A.R., 2009. Induction of IL-6 and inhibition of IL-8 secretion inthe human airway cell line Calu-3 by urban particulate matter collected with amodified method of PM sampling. Environ. Res. 109 (5), 528–535.

Becker, S., Dailey, L.A., Soukup, J.M., Grambow, S.C., Devlin, R.B., Huang, Y.C., 2005.Seasonal variations in air pollution particle-induced inflammatory mediatorrelease and oxidative stress. Environ. Health Perspect. 113 (8), 1032–1038.

Brunekreef, B., Forsberg, B., 2005. Epidemiological evidence of effects of coarseairborne particles on health. Eur. Respir. J. 26 (2), 309–318.

Castillejos, M., Borja-Aburto, V.H., Dockery, D.W., Gold, D.R., Loomis, D., 2000.Coarse particles and mortality in Mexico City. Inhalation Toxicol. 12 (s1), 61–72.

Cheng, M.F., Ho, S.C., Chiu, H.F., Wu, T.N., Chen, P.S., Yang, C.Y., 2008. Consequencesof exposure to Asian dust storm events on daily pneumonia hospital admissionsin Taipei, Taiwan. J. Toxicol. Environ. Health, Part A 71 (19), 1295–1299.

Cifuentes, L.A., Vega, J., Kopfer, K., Lave, L.B., 2000. Effect of the fine fraction ofparticulate matter versus the coarse mass and other pollutants on dailymortality in Santiago, Chile. J. Air Waste Manage. Assoc. 50 (8), 1287–1298.

Dominici, F., McDermott, A., Daniels, M., Zeger, S.L., Samet, J.M., 2005. Revisedanalyses of the National Morbidity, Mortality, and Air Pollution Study: mortalityamong residents of 90 cities. J. Toxicol. Environ. Health, Part A 68 (13–14),1071–1092.

Duvall, R.M., Norris, G.A., Dailey, L.A., Burke, J.M., McGee, J.K., Gilmour, M.I., Gordon,T., Devlin, R.B., 2008. Source apportionment of particulate matter in the US andassociations with lung inflammatory markers. Inhalation Toxicol. 20 (7), 671–683.

Franklin, M., Koutrakis, P., Schwartz, P., 2008. The role of particle composition onthe association between PM2.5 and mortality. Epidemiology 19 (5), 680–689.

García-Cuellar, C., Alfaro-Moreno, E., Martínez-Romero, F., Ponce de León Rosales, S.,Rosas, I., Pérez-Cárdenas, E., Osornio-Vargas, A.R., 2002. DNA damage inducedby PM10 from different zones of Mexico City. Ann. Occup. Hyg. 46 (Suppl. 1),425–428.

IAEA, 1995.Manual for QXAS. International Atomic Energy Agency, Vienna, Austria.Lippmann, M., Ito, K., Hwang, J.S., Maciejczyk, P., Chen, L.C., 2006. Cardiovascular

effects of nickel in ambient air. Environ. Health Perspect. 114 (11), 1662–1669.Mann, J.K., Balmes, J.R., Bruckner, T.A., Mortimer, K.M., Margolis, H.G., Pratt, B.,

Hammond, S.K., Lurmann, F.W., Tager, I.B., 2010. Short-term effects of airpollution on wheeze in asthmatic children in Fresno, California. Environ. HealthPerspect. 118 (10), 1497–1502.

Mar, T.F., Norris, G.A., Koenig, J.Q., Larson, T.V., 2000. Associations between airpollution and mortality in Phoenix, 1995–1997. Environ . Health Perspect. 108(4), 347–353.

Mendoza, A., Pardo, E.I., Gutierrez, A.A., 2010. Chemical characterization andpreliminary source contribution of fine particulate matter in the Mexicali/Imperial Valley border area. J. Air Waste Manage Assoc. 60, 258–270.

Miranda, J., Cahill, T.A., Morales, J.R., Aldape, F., Flores, M.J., 1994. Determination ofelemental concentrations in atmospheric aerosols in Mexico City using particleinduced X-ray emission, proton elastic scattering and laser absorption. Atmos.Environ. 28 (14), 2299–2306.

Miranda, J., Rodríguez-Fernández, L., De Lucio, O.G., López, K., Harada, J.A., 2000. Anew beamline for characteristic X-ray experiments at the Pelletron accelerator.Rev. Mex. Fís. 46, 367–372.

Moreno, T., Querol, X., Pey, J., Minguillón, M.C., Pérez, N., Alastuey, A., Bernabé, R.M.,Blanco, S., Cárdenas, B., Eichinger, W., Salcido, A., Gibbons, W., 2008. Spatial andtemporal variations in inhalable CuZnPb aerosols within the Mexico Citypollution plume. J. Environ. Monit. 10 (3), 370–378.

Osornio-Vargas, A.R., Hernandez-Rodriguez, N.A., Yanez-Buruel, A.G., Ussler, W.,Overby, L.H., Brody, A.R., 1991. Lung cell toxicity experimentally induced by amixed dust from Mexicali, Baja California, Mexico. Environ. Res. 56 (1), 31–47.

Osornio-Vargas, A.R., Bonner, J.C., Alfaro-Moreno, E., Martinez, L., Garcia-Cuellar, C.,Ponce-de-Leon Rosales, S., Miranda, J., Rosas, I., 2003. Proinflammatory andcytotoxic effects of Mexico City air pollution particulate matter in vitro aredependent on particle size and composition. Environ. Health Perspect. 111 (10),1289–1293.

Ostro, B.D., Hurley, S., Lipsett, M.J., 1999. Air pollution and daily mortality in theCoachella Valley, California: a study of PM10 dominated by coarse particles.Environ. Res. 81 (3), 231–238.

Paredes-Gutiérrez, R., López-Suárez, A., Miranda, J., Andrade, A., González, J.A., 1997.Comparative study of elemental contents in atmospheric aerosols from threesites in Mexico City using PIXE. Rev. Int. Contam. Ambient. 13 (2), 81–85.

Peng, R.D., Chang, H.H., Bell, M.L., McDermott, A., Zeger, S.L., Samet, J.M., Dominici,F., 2008. Coarse particulate matter air pollution and hospital admissions forcardiovascular and respiratory diseases among Medicare patients. JAMA 299(18), 2172–2179.

Perez, L., Tobias, A., Querol, X., Kunzli, N., Pey, J., Alastuey, A., Viana, M., Valero, N.,Gonzalez-Cabre, M., Sunyer, J., 2008. Coarse particles from Saharan dust anddaily mortality. Epidemiology 19 (6), 800–807.

Perez, L., Medina-Ramon, M., Kunzli, N., Alastuey, A., Pey, J., Perez, N., Garcia, R.,Tobias, A., Querol, X., Sunyer, J., 2009. Size fractionate particulate matter,vehicle traffic, and case-specific daily mortality in Barcelona, Spain. Environ. Sci.Technol. 43 (13), 4707–4714.

Perrone, M.G., Gualtieri, M., Ferrero, L., Lo Porto, C., Udisti, R., Bolzacchini, E.,Camatini, M., 2010. Seasonal variations in chemical composition and in vitrobiological effects of fine PM from Milan. Chemosphere 78 (11), 1368–1377.

Pope, C.A., Dockery, D.W., 2006. Health effects of fine particulate air pollution: linesthat connect. J. Air Waste Manage. Assoc. 56 (6), 709–742.

Rosas Perez, I., Serrano, J., Alfaro-Moreno, E., Baumgardner, D., Garcia-Cuellar, C.,Martin, Miranda., del Campo, J.M., Raga, G.B., Castillejos, M., Colin, R.D., OsornioVargas, A.R., 2007. Relations between PM10 composition and cell toxicity: amultivariate and graphical approach. Chemosphere 67 (6), 1218–1228.

Schwarze, P.E., Ovrevik, J., Lag, M., Refsnes, M., Nafstad, P., Hetland, R.B., Dybing, E.,2006. Particulate matter properties and health effects: consistency ofepidemiological and toxicological studies. Hum. Exp. Toxicol. 25 (10), 559–579.

Schwarze, P.E., Ovrevik, J., Hetland, R.B., Becher, R., Cassee, F.R., Lag, M., Lovik, M.,Dybing, E., Refsnes, M., 2007. Importance of size and composition of particles foreffects on cells in vitro. Inhalation Toxicol. 19 (Suppl. 1), 17–22.

Steerenberg, P.A., van Amelsvoort, L., Lovik, M., Hetland, R.B., Alberg, T., Halatek, T.,Bloemen, H.J., Rydzynski, K., Swaen, G., Schwarze, P., Dybing, E., Cassee, F.R.,2006. Relation between sources of particulate air pollution and biological effectparameters in samples from four European cities: an exploratory study.Inhalation Toxicol. 18 (5), 333–346.

Tong, H., Cheng, W.Y., Samet, J.M., Gilmour, M.I., Devlin, R.B., 2010. Differentialcardiopulmonary effects of size-fractionated ambient particulate matter inmice. Cardiovasc. Toxicol.. doi:10.1007/s12012-010-9082-y.

Valavanidis, A., Fiotakis, K., Vlachogianni, T., 2008. Airborne particulate matter andhuman health: toxicological assessment and importance of size andcomposition of particles for oxidative damage and carcinogenic mechanisms.J. Environ. Sci. Health, Part C: Environ. Carcinog. Ecotoxicol. Rev. 26 (4), 339–362.

Veranth, J.M., Reilly, C.A., Veranth, M.M., Moss, T.A., Langelier, C.R., Lanza, D.L., Yost,G.S., 2004. Inflammatory cytokines and cell death in BEAS-2B lung cells treatedwith soil dust, lipopolysaccharide, and surface-modified particles. Toxicol. Sci.82 (1), 88–96.

Veranth, J.M., Moss, T.A., Chow, J.C., Labban, R., Nichols, W.K., Walton, J.C., Watson,J.G., Yost, G.S., 2006. Correlation of in vitro cytokine responses with thechemical composition of soil-derived particulate matter. Environ. HealthPerspect. 114 (3), 341–349.

Viau, E., Levi-Schaffer, F., Peccia, J., 2010. Respiratory toxicity and inflammatoryresponse in human bronchial epithelial cells exposed to biosolids, animalmanure, and agricultural soil particulate matter. Environ. Sci. Technol. 44 (8),3142–3148.

Warheit, D.B., Webb, T.R., Colvin, V.L., Reed, K.L., Sayes, C.M., 2007. Pulmonarybioassay studies with nanoscale and fine-quartz particles in rats: toxicity is notdependent upon particle size but on surface characteristics. Toxicol. Sci. 95 (1),270–280.

Watson, J.G., Chow, J.C., 2001. Source characterization of major emission sources inthe imperial and Mexicali Valleys along the US/Mexico border. Sci. TotalEnviron. 276 (1–3), 33–47.

Zuk, M., Tzintzun-Cervantes, M.G., Rojas Bracho, L., 2007. Tercer almanaque dedatos y tendencias de la calida del aire en nueve ciudades mexicanas. Secretaríade Medio Ambiente y Recursos Naturales, Instituto Nacional de Ecología,Mexico, DF, pp. 49–61. ISNB 978-968-817-840-9 (Chapter 4).