, Volume 27, Issue 2, pp 97–105

Endotoxin as modifier of particulate matter toxicity: a review of the literature


    • Laboratory of Experimental Air Pollution, Department of Pathology, Medical SchoolUniversity of Sao Paulo
  • Paulo Hilário Nascimento Saldiva
    • Laboratory of Experimental Air Pollution, Department of Pathology, Medical SchoolUniversity of Sao Paulo
  • Christine Rogers
    • Environmental Health Sciences, School of Public Health and Health SciencesUniversity of Massachusetts
Review Paper

DOI: 10.1007/s10453-010-9179-6

Cite this article as:
Degobbi, C., Saldiva, P.H.N. & Rogers, C. Aerobiologia (2011) 27: 97. doi:10.1007/s10453-010-9179-6


It is well known that particulate matter (PM) and endotoxin are able to trigger inflammatory responses in the lung. Most studies have focused on the components separately and on the identification of chemical components associated with PM. However, since biological components may represent around 20% of airborne PM, and endotoxin may reach concentrations as high as 30 EU/mg, recent studies have focused attention on the characterization of endotoxin present in PM and health effects. Most of the literature has suggested that endotoxin adsorbed in PM is able to elicit immunological responses associated with increase in pro-inflammatory cytokine expression. The aim of this paper is to provide an up to date review of the findings involving toxicity effects of endotoxin associated with PM.


EndotoxinParticulate matterCytokinesHealth effects

1 Introduction

Exposure to ambient particles has been consistently associated with adverse health effects (Alfaro-Moreno et al. 2007; Pope et al. 2004, 2009; Samet et al. 2000). Particle mass is only a crude measure of a complex mixture of carbon (elemental and organic), metals and ions (WHO 2005), whose composition varies with combustion sources as well as photochemical processes (Carlton et al. 2010; Park et al. 2009). In addition, several studies have indicated that adverse health effects are influenced by particle composition (Carvalho-Oliveira et al. 2005; Franklin et al. 2008; Saldiva et al. 2002). The majority of these studies focus on chemical characteristics (Chow et al. 2002; Edgerton et al. 2006; Katsouyanni et al. 2001; Kim et al. 2000; Latha and Badarinath 2005; Ostro et al. 2007; Turnbull and Harrison 2000; Vallius et al. 2005; Wang et al. 2006), with less information on the role of biological component, such as endotoxin, that may be present in ambient particles (Heinrich et al. 2003).

Many factors, including meteorological variations, influence the chemical and biological constitution of particulate matter. Studies have shown that high temperatures are responsible for increasing secondary particle formation (Aw and Kleeman 2003), as well as influencing endotoxin content in particles (Carty et al. 2003). Hence, determining particle toxicity is complex. The present review was designed to compile studies focusing on the role of endotoxin in modulating particle toxicity. The document is structured as follows: (a) a brief description of the nature and toxicity of endotoxin; (b) a summary of the relative content of endotoxin in ambient particles; (c) evidence of the contribution of endotoxin to particle toxicity.

2 Endotoxin

Endotoxin is associated with a constituent of the outer cell wall of gram-negative bacteria (Fig. 1) and is a known trigger of airway inflammation and respiratory symptoms, such as nasal obstruction and discharge, wheeze and sneezing (Dales et al. 2006). Commonly, the term endotoxin describes the biological activity of lipopolysaccharides (LPS), which are comprised of a lipid A moiety, a core oligosaccharide and an O-antigen. Endotoxin activity is specifically associated with the lipid A domain, which activates the innate immune system (Miller et al. 2005; Trent et al. 2006). In contrast to ambient PM studies, most of the studies characterizing endotoxin concentration have been conducted indoors and are described later in the text.
Fig. 1

Outer cell wall of gram-negative bacteria showing LPS structure (O-antigen, Core and Lipid A). Endotoxin activity corresponds to the lipid A region. Source: (Murray et al. 2007)

Endotoxin inhalation has been associated with acute lung injury in murine models (Kabir et al. 2002) and with symptoms such as airway irritation, alterations in breathing, fever (Preas et al. 2001), cough, sneezing (Smid et al. 1994) and progressive airway obstruction in individuals (Kline et al. 1999). It seems that exposure to endotoxin can also result in asthma exacerbations (Michel et al. 1991).

The exact doses of endotoxin responsible for immunological mediators leading to TH1 or TH2 responses are not well known. Some studies with animal models suggest that doses of LPS as low as 0.1 μg lead to TH2 type responses (atopic), which may involve release of eosinophils, IL-5 and IL-13 production, activation of TLR4 signalling pathways (Eisenbarth et al. 2002), decrease in IFN-γ levels and an increase in phagocyte function (Alexis et al. 2004). In contrast, high doses of LPS (100 μg) may lead to a TH1 response with an increase in IFN-γ and IL-12 production (Eisenbarth et al. 2002).

The first study that observed endotoxin-induced TH1 response in vivo was conducted by Gereda et al. (2000) with 61 infants 9–24 months old with at least three physician-documented episodes of wheezing. The authors measured endotoxin and allergens in the homes of these infants. In addition, skin prick tests were performed, cytokine production assessed, and the results showed not only that allergen-sensitized infants lived in homes with significantly lower doses of endotoxin (geometric mean: 468 EU/mL), but also higher doses (geometric mean: 1035 EU/mL) were significantly correlated with higher IFN-γ levels. Another study involving 6- to 13-year-old children revealed that endotoxin levels (geometric mean: 37.8 EU/mg of dust) were inversely correlated with the occurrence of hay fever, atopic asthma and atopic sensitization (Braun-Fahrlander et al. 2002). Some studies have suggested that living in farm areas, with high concentrations of endotoxin, may help children to develop a TH1 instead of TH2 response (von Mutius et al. 2000), in agreement with the “hygiene hypothesis”, which states that exposure to high microbe levels in early life leads to TH1 response and protects against atopy (Liu and Redmon 2001).

On the other hand, a study examining atopy in the first 2 years of life has associated high doses of endotoxin exposure during the first months (i.e. mean of 2893 EU/g of dust) with repeated wheeze and atopy to inhalant allergens, especially in high-risk infants (Bolte et al. 2003). Similar results have been found by Gillespie et al. (2006) at 3 months of age and by Park et al. (2001) in high-risk infants in the first year of life. There is some concern about the possible transient characteristic of wheeze symptoms, raising questions about the protective or exacerbating action of endotoxin exposure later in life (Perzanowski et al. 2006). Detailed reviews on the relationship between endotoxin and TH1/TH2 modulatory responses may be found in (Douwes et al. 2002; Liu 2002, 2001).

The complexity of studying different concentrations of endotoxin and their relationship with atopy is aggravated by the fact that endotoxin may act together with other allergens which also modify inflammatory responses. Some studies have found that higher doses of endotoxin may exacerbate symptoms in infants previously sensitized to inhalant allergens (Bolte et al. 2003; Rizzo et al. 1997). Also, other microbiological contaminants, such as fungi, may grow in the same conditions as bacteria in indoor environments and must be considered when carrying out risk assessment (Park et al. 2006). Thus, studies involving possible mechanisms of development of atopy or protection due to exposure to endotoxin must focus not only on the dose of exposure, but also on the time of exposure and concomitant exposure with other agents.

3 Endotoxin content in particulate matter

Since particulate matter and endotoxin by themselves are responsible for a wide range of inflammatory responses, recent studies have focused on the characterization of endotoxin levels associated with PM2.5 and PM10. Biological material derived from microbes may become dissociated from the original particle and associated with and carried by PM. Hence, immunologically relevant molecules can become associated with particles with smaller aerodynamic properties, which enables penetration to deeper lung sites (Monn and Koren 1999). This issue is relatively important because biological particles, such as bacteria, fungal spores, pollen and others may represent around 20% of the total particulate in populated continental areas (Matthias-Maser et al. 2000), and endotoxin may reach high concentrations, such as 30 EU/mg of PM (Mueller-Anneling et al. 2004).

Due to the relevance of this topic, studies have been conducted in order to identify parameters that influence endotoxin aerosolization and its relationship to particulate matter. Temperature and relative humidity seem to be related to airborne endotoxin. Carty et al. (2003) have assessed the endotoxin content of PM2.5 in outdoor air of Munich, Germany and found that the levels were associated with higher temperatures (with the highest levels during summer and spring) and a decrease in per cent relative humidity. These results partially agree with studies conducted by our group, where endotoxin was positively correlated with temperature in the atmosphere of Sao Paulo, Brazil. It seems that the endotoxin content in PM is higher in the warmer seasons and is more associated with PM10 than with PM2.5 (Osornio-Vargas et al. 2003; Schins et al. 2004), with concentrations up to ten times higher in PM10 when compared to the fine fraction (Heinrich et al. 2003). Samples collected from Southern California showed that rural agricultural sites had the lowest PM10 (mean-20 μg/m3), but mid endotoxin concentrations (mean-18.9 EU/mg of PM) when compared to desert and mountain sites. In desert and mountains’ sites, despite the low value for PM10 (mean-21.1 μg/m3), the endotoxin content was the highest (mean-30 EU/mg). For the highest reported PM10 concentrations, the endotoxin levels tended to be in the upper quartile. Endotoxin and PM were most correlated in the summer, and no other pollutants (ozone, nitrogen dioxide, PM2.5 and total acids) showed the same pattern (Mueller-Anneling et al. 2004). In another study comparing rural and industrialized areas in spring time, the endotoxin content of coarse particles was higher in rural areas, which maybe explained by the high influence of vegetation sources (Schins et al. 2004).

4 Contribution of endotoxin to PM toxicity

The contribution of endotoxin to PM toxicity is a recent topic in the literature and has been investigated mostly through experiments involving inhalation of PM, tracheal instillation, or macrophage direct stimulation using animal models. The main studies are summarized in Table 1.
Table 1

Results of studies on synergic effects of endotoxin and particulate matter



Effects of endotoxin and particulate matter

 Dybing et al.

Enhanced TNFα and IL-6 response to PM was not due to endotoxin

 Huang et al.

Endotoxin was responsible for 36% of TNF-α production associated with PM2.5 and 24% for PM10

 Steerenberg et al.

Endotoxin content of PM was not correlated with immunoglobulin release, cytokines, histopathological changes in the lungs or total cell numbers

 Osornio-Vargas et al.

Endotoxin associated with PM increased the release of TNF-α and IL-6

 Schins et al.

The highest endotoxin content coincided with the highest IL-8 and TNF-α release

 Imrich et al.

Macrophages primed with LPS had an amplified TNF-α release response to PM

 Ning et al.

Endotoxin plus PM caused 30-fold more macrophage inflammatory protein-2 (MIP-2) production

 Bonner et al.

Metal and endotoxin present in PM activated IL-1 β and PDGF

 Monn et al. and Becker et al.

Endotoxin plus PM-induced TNF-α, IL-6 and IL-8 production

 Alexis et al.

Biological contaminants plus PM enhanced TNF-α mRNA, mCD14, CD11b/CR3 and HLA-DR expression

 Dong et al.

Endotoxin enhanced the TNF-α response to PM

 Becker et al.

PM had the cytokine production inhibited by antibody to CD14

 Long et al.

TNF-α response amplified after priming the macrophages with LPS

Imrich et al. (1999) found that macrophages primed with LPS had an amplified TNF-α release response to concentrated ambient particles with diameters less than 2.5 μm (CAPS) collected in Boston, USA, and urban air particles, not concentrated, collected in Washington, USA. Another study has shown that trace endotoxin associated with CAPs caused a nearly 30-fold greater macrophage inflammatory protein-2 (MIP-2) production than predicted by endotoxin alone (Ning et al. 2000). Studies that took place in three regions of Mexico City have shown that metals and endotoxin together are able to induce platelet-derived growth factor [(PDGF)-which helps to regulate mesenchymal cell proliferation] by activating IL-1β, and in another pathway, endotoxin associated with PM may be responsible for the direct activation of PDGF (Bonner et al. 1998). Later, the same group found differences in metal and endotoxin content associated with PM10 in the same regions. The results might point to metal as an inducer of toxic effects, and endotoxin of most of the proinflammatory responses (Alfaro-Moreno et al. 2002), and these results may be similar to those achieved by Monn and Becker (1999). The role of endotoxin became clearer when the authors found that the south-eastern PM10 induced the highest TNF-α and IL-6 releases and part of the effect was due to endotoxin since the addition of rENP (an endotoxin inhibitor) reduced the response. Also, the endotoxin content in the south-eastern samples was two-fold higher than the endotoxin content from PM10 from the northern region (Osornio-Vargas et al. 2003). Another study suggested that endotoxin and metals are responsible for inflammatory responses associated with coarse fraction (PM2.5–10), but triggered by different pathways. Some metals were correlated with aracdonic acid release, and endotoxin was associated with higher concentrations of TNF-α released by purified macrophages (Mudway et al. 2010). In another study, the endotoxin content in the coarse fraction was up to ten times higher in rural sites, which represented the major IL-8 and TNF-α release after intra-tracheal instillation in rat lung. No such association was found for the fine fraction (Schins et al. 2004).

One study has evaluated the immunological response of healthy subjects exposed to inhaled PM10 (~0.65 mg/subject) (Alexis et al. 2006). The results showed that TNF-α mRNA expression in purified sputum macrophages after PM2.5–10+ exposure (PM associated with biological material) was enhanced compared to the control (saline group), but no such association was observed after PM2.5–10− inhalation (with PM heat treated to inactivate biological contaminants). Also, PM2.5–10+ was correlated with the monocyte phagocytosis and different cell phenotypes with a significant increase in macrophage and monocyte mCD14 expression (associated with LPS-mediated responses), macrophage CD11b/CR3 expression (associated with host defence) and HLA-DR macrophage expression (associated with antigen presentation). The exact mechanism of activation is reviewed elsewhere (Kitchens 2000). No differences in neutrophil influx to the airways have been observed, highlighting the importance of microbial components in macrophage activation, but not in neutrophil release (Alexis et al. 2006). These results are complemented by Huang et al. (2002), who attributed 36% of TNF-α production of mouse monocyte–macrophage cells to endotoxin in PM2.5 and 24% (when considering direct endotoxin measures) in PM10. Previously, Dong et al. (1996) have shown that the utilization of polymyxin B sulphate (an endotoxin neutralizer) was capable of completely inhibiting cytokine release (including TNF-α) by rat alveolar macrophages after exposure to treated urban air particles. Interestingly, no endotoxin levels or cytokine release was found after exposure to treated diesel particles either. Altogether, the data suggest that all of the cytokine response was due to endotoxin attached to urban particles. Similar results with pro-inflammatory cytokine (TNF-α, IL-6 and IL-8) inhibition after complete (Monn and Becker 1999) or partial (Becker et al. 1996) endotoxin neutralization associated with coarse and urban air particles have been found.

In addition to studies with purified LPS, studies involving intact bacterial cells have found that gram-positive and gram-negative bacteria are responsible for PM-induced stimulation of alveolar macrophages (Becker et al. 2002). Although only 30% of the bacteria isolated were gram negative, there was a preferential response of alveolar macrophages to the group, which plays a major role in the response at lower particle levels. Only at high particle concentrations was there sufficient gram-positive cell numbers to invoke a cytokine response. Indeed, only one gram-negative bacterium was necessary to activate one hundred alveolar macrophages resulting in IL-6 release in vitro, but three times more gram-positive bacteria was necessary to generate the same effect (Becker et al. 2002). Later, the same group has found that human alveolar macrophages exposed mainly to coarse particle fraction (PM2.5–10) had cytokine production inhibited by antibody to CD14, suggesting that endotoxin was associated with inflammatory responses. Also, phagocytosis of opsonized yeast and yeast-induced oxidative burst were inhibited. The authors concluded that the results suggested the response of human macrophages to PM involves receptors to microbial components, which may be responsible for inflammatory outcomes (Becker et al. 2003). In fact, adsorbed endotoxin on high-density polyethylene and Co-Cr–Mo alloy has been demonstrated to enhance macrophage-mediated cytokine release (IL-1β, TNFα and IL-6) in vitro (Daniels et al. 2000). It has also been demonstrated that the macrophage cytokine activity in vitro is mostly mediated by insoluble materials present in concentrated ambient air particles (CAPs) and adsorbed endotoxin, but the activation state of the cells determined the response to a specific component (i.e. purified macrophages were highly activated by endotoxin associated with CAPs, but macrophages primed with endotoxin were activated by other components, not endotoxin, present in CAPs) (Imrich et al. 2000). These results are similar to others that demonstrated the LPS is most associated with the particle suspension (or insoluble material) and coarse particles than to the supernatant of CAPs samples and is responsible for a 50-fold increase in IL-6, TNFα, and MCP-1 release and macrophage phagocytosis inhibition (Soukup and Becker 2001).

In contrast to studies that found linear relationships between endotoxin content in PM and pro-inflammatory cytokine release, one study (Hetland et al. 2005) that has collected PM2.5 and PM10 in European cities has found that rat macrophage release of TNFα and IL-6 was higher after PM10 challenge (which also had a higher endotoxin content) but the results were not related to endotoxin, since the addition of polymyxin B sulphate (an endotoxin neutralizer) did not change the cytokine release pattern. It should be noted that the authors found that higher levels of endotoxin were related to higher levels of IL-6 release, but no direct proportionality was observed. Also, the author did not exclude the possibility of endotoxin interaction with other proinflammatory components (Hetland et al. 2005). These results may partially agree with a previous study where endotoxin content of PM2.5 and PM10 was not correlated with immunoglobulin release (IgE, IgG1 and IgG2a), cytokines (IL-4, IL-5, IFN-γ and TNF-α), histopathological changes in the lungs and total cell numbers, after ovalbumin challenge in mice (Steerenberg et al. 2004).

Indoor studies characterizing particulate matter and endotoxin effects are less often done, but they highlight the importance of such measurement since people spend about 90% of their time indoors. It is worthwhile to note that there are differences in particle constitution, with indoor samples being derived from internal sources and by outdoor particles that penetrate to the interior. A small in vitro study in the Boston area utilized paired indoor and outdoor PM2.5 samples to stimulate rat alveolar macrophages. There was an increase in TNF-α release after exposure to both, indoor and outdoor particles. The response was amplified after priming the macrophages with LPS, suggesting that preexisting proinflammatory conditions may exacerbate the PM2.5 effects. Also, indoor particles led to a higher TNF-α release, slightly enhanced after the data were normalized for endotoxin. However, the endotoxin amounts found indoors and outdoors did not differ significantly, which suggests that other compounds may act synergistically with endotoxin, amplifying the inflammatory response (Long et al. 2001). Another study did not find significant cytokine release by human monocytes exposed to PM2.5 and PM10 from indoors, but found cytokine release for PM10 from outdoor air, with pro-inflammatory activity related to endotoxin content (Monn and Becker 1999).

In addition to endotoxin responses associated with PM, other studies have focused on the enhancement of inflammatory effects of endotoxin exposure concomitantly with other agents, such as fungi (Rylander and Holt 1998; Wan and Li 1999), allergens (Braga et al. 2004; Eldridge and Peden 2000) and ozone (Harkema and Wagner 2005; Wagner et al. 2001).

5 Conclusions

Particulate matter and endotoxin are responsible for a wide range of health effects, including inflammatory and toxic outcomes. Since biological contaminants may be carried by the particles and represent nearly 20% of PM total mass, recent studies have focused on the characterization of endotoxin content in PM and inflammatory responses, mainly, in vitro. Most of the studies have shown that PM10 has the highest endotoxin content when compared to PM2.5. Also, the highest concentrations in outdoor air may be associated with warmer seasons and the insoluble fraction.

Immunological responses to PM have provided interesting results. Some authors have shown an increase in pro-inflammatory cytokine release after priming with LPS, highlighting the importance of the modulatory response. Controversial results related to inflammatory responses have been found, but most of the articles have attributed the responses mainly to endotoxin content associated with PM, especially PM10. Comparative indoor/outdoor studies have shown endotoxin-dependent proinflammatory responses, but have provided conflicting results about the PM potency found indoors and outdoors.

This review has focused on the importance of endotoxin as modulator of toxicity in PM. Further research is needed to confirm the results listed previously and to better understand synergistic effects and immunological responses associated not only with chemical, but also with endotoxin content of particulate matter.

Copyright information

© Springer Science+Business Media B.V. 2010