According to the preliminary data from West Virginia University, consistent statistically significant associations of serum electrolytes with PFAA serum values in a large survey population have been noted (preliminary data available at http://www.hsc.wvu.edu/som/cmed/c8/
). This might result from PFAA-induced vascular endothelium barrier dysfunction. A 30 year medical surveillance has also found an association between elevated uric acid levels and perfluorooctanoic acid (PFOA), another key member of PFAA, concentrations in human serum (Costa et al., 2009
). Hyperuricemia was found to induce endothelial dysfunction (Nakagawa et al., 2006
). Moreover, it was suggested that dysfunction of vascular system may be associated with low birth weight (Holt and Byrne, 2002
; Norman, 2008
). A single human study with unconventional study design raised an association between PFAA exposures and human cardiovascular disease (Anderson-Mahoney et al., 2008
). This study has several limitations, and the question of human cardiovascular disease awaits its first large-scale population evaluation with incident-epidemiologic designs. More fundamentally, PFOS does induce oxidative stress and is known to change surface membrane potential and to alter calcium channels (Harada et al., 2005
). Furthermore, glucose homeostasis and hepatic metabolism appear to be systemically affected, in studies using NHANES data (Lin et al., 2009a
). There is ample physiologic reason to consider endothelial response in cell system and toxicity studies, as well as cardiovascular disease in future incident studies of human populations. Therefore, it is postulated that the identification of PFOS-induced endothelial dysfunction may provide some fundamental information concerning adverse health effects on humans.
In this report, human microvascular endothelial cells (HMVEC), an in vitro
human microvascular endothelium model, were used. This model was generated through immortalization of primary endothelial cells by engineering with a human telomerase catalytic protein (hTERT) (Shao and Guo, 2004
). This cell model resembles the signature of primary human microvascular endothelial cells in phenotype and gene expression profile more than a commercially available human microvascular endothelial cell-1 (HMEC-1) cell line does. HMEC-1 were established by the transduction of SV40 large T antigen. The availability of this model in our laboratory provides us with a unique advantage in the study of PFOA and PFOS-induced effects on endothelium. We have applied this model to successfully identify the toxic effects of iron oxide nanoparticles in vitro (Apopa et al., 2009
The results of this study demonstrate that exposure of HMVEC to both high and low concentrations of PFOS induced the production of ROS. Morphologically, PFOS exposure induced actin filament remodeling and endothelial permeability changes in HMVEC. Furthermore, the production of ROS appeared to play a regulatory role in PFOS-induced actin filament remodeling and endothelial permeability increase. Taken together, the results lead us to hypothesize that PFOS-induced ROS production plays a role in the aberrations of the endothelial permeability barrier. Aberrations of endothelial permeability were reported to play a major role in the pathogenesis of many human diseases, including inflammation, acute lung injury syndromes, and carcinogenesis (Houle and Huot, 2006
; Mehta and Malik, 2006
). Therefore, results from the present study provide insight into potential mechanisms for the effects of PFOS exposure upon humans at the cellular level.
Reactive oxygen species (ROS) refer to a diverse group of reactive, short-lived, oxygen containing species, such as O2·−,
, and ·
OH. Cellular systems are protected from ROS-induced cell injury by an array of defenses composed of various antioxidants with different functions. When ROS in the cellular system overpower the defense systems, ROS produce cell damage or oxidative stress, leading to the development of various diseases. ROS possess five crucial properties of cell injury with capacities: (a) to cause permanent structural changes in DNA, (b) to initiate lipid peroxidation, (c) to induce protein oxidation, (d) to modulate the activity of stress proteins and stress genes that regulate effector genes related to growth, differentiation, and cell death, and (e) to activate cytoplasmic and nuclear signal transduction pathways (Qian et al., 2003a
). ROS have been implicated to be involved in many diseases, ranging from cardiovascular disorders, carcinogenesis, chronic inflammation, and neurodegenerative diseases (Halliwell et al., 1992
; Gutteridge and Halliwell, 2000
). Two separate studies from other research groups suggested that PFOS-induced oxidative stress response was involved in developmental neurotoxicity in PC12 cells, a neuronotypic cell line (Slotkin et al., 2008
), and hepatic fatty acyl-CoA oxidase activity increase in fish (Oakes et al., 2005
). Recently, a report showed that exposure of human hepatoma HepG2 cells to 50–200 μM PFOS induced ROS production, leading to dissipation of mitochondria membrane potential and induction of apoptosis (Hu and Hu, 2009
Wei and colleagues (2009)
found that PFOA and PFOS mixtures affect minnow genes implicated in oxidative stress. Huang and colleagues (2009)
used PFOS to induce oxidative stress in Atlantic salmon. Nakayama and colleagues (2008)
report that PFOS concentration is positively correlated with mRNA levels of glutathione peroxidase 1 and glutathione-s-transferase alpha 3 in cormorant liver, as well as negatively correlated with the response of heat shock protein 8. They inferred that antioxidant enzymes are induced in response to oxidative stress and suppression of molecular chaperones, from PFC exposure, leading to reductions in protein stability (Nakayama et al., 2008
). In human population studies, the consistent and strong association of biomarkers of PFOS with elevated lipids suggests a further reason to examine oxidative stress in humans, especially since humans do not metabolize or excrete longer-chain perfluorocarbons readily (Steenland et al., 2009
; Nelson, et al., 2010
However, a potentially important issue raised from these results is that the concentrations of PFOS applied in these assays were substantially higher than the concentrations relevant to occupational and environmental exposures. Serum PFOS levels in occupational workers are typically between 0.5–2 μg/ml (about 1–5 μM), with the highest level reaching 13 μg/ml (26 μM) (Olsen et al., 1998
; Lau et al., 2007
; Slotkin et al., 2008
). In the general population, serum PFOS levels were reported to be approximately 20 ng/ml (40 nM) (Calafat et al., 2007
; Slotkin et al., 2008
). Therefore, one could argue that identification of PFOS-induced ROS at the low range concentrations may yield more relevant data. In the present study, both macrophages and HMVEC were initially exposed to high-range concentrations of PFOS to determine their capability to produce ROS according to previously published dose ranges (Oakes et al., 2005
; Slotkin et al., 2008
; Hu and Hu, 2009
). Then, HMVEC were exposed to occupationally and environmentally relevant low concentrations of PFOS to detect its capability to produce ROS and HMVEC effects. Results demonstrate that PFOS exposure was able to induce ROS production in an in vitro human endothelial cell system within the occupationally and environmentally exposure-relevant concentration ranges, which implies that human exposure to PFOS at ranges compatible with biomonitoring results might induce the generation of ROS.
In this study, three different techniques were used to measure PFOS-induced ROS production in both macrophages and HMVEC to take advantage of the unique features of each individual technique. To evaluate the production of PFOS-induced ROS in macrophages, ESR spin trapping was used. Spin trapping is the method of choice for detection and identification of free radical generation in biological systems due to its specificity and sensitivity (Qian et al., 2001
). Confocal microscopy imaging analysis was also used to examine PFOS-induced ROS in HMVEC. In recent years, the molecular staining techniques have been significantly improved to serve as additional methods to identify and characterize specific ROS. The confocal microscopy imagining analysis technique provides the physical images of ROS production in the cell system, as well as information about whether the free radicals are produced intracellularly or extracellularly. To further quantify the production of ROS, a fluorescence microplate reader was used to measure PFOS-induced ROS production in HMVEC. With these three different techniques, results indicate that PFOS exposure induced significant amounts of superoxide radical and hydroxyl radical production intracellularly in both human endothelial cells and mouse macrophages.
Colorimetric changes reflect the production of ROS in HMVEC ( and ). In , reflecting a concentration range of PFOS between 0 to100 μM, the difference of ROS production was large and the change in red color staining substantial, which was easily detected by a confocal microscope. In contrast, in , the concentration range of PFOS was between 0 to 5 μM, and the difference of ROS production was small. The change in red color staining was not detected by a confocal microscope. However, the changes in the blue color staining, which measure how much the dye was degraded by ROS, were substantial (), indirectly reflecting the amount of PFOS-induced ROS production in HMVEC.
The results of this study indicate that PFOS exposure induced an alteration in endothelial permeability. Mechanistically, data show that PFOS-induced endothelial permeability was modulated by production of ROS. These results regarding the role of ROS in endothelial cell permeability are consistent with several published observations by other researchers that ROS-induced oxidant stress directly increases endothelial permeability (Lum and Roebuck, 2001
; Houle and Huot, 2006
). It was found that exposure of endothelial cell monolayer to ROS generators, xanthine/xanthine oxidase or glucose/glucose oxidase, increased endothelial cell permeability in a concentration-dependent manner (Shasby et al., 1985
; Holman and Maier, 1990
). Furthermore, the direct stimulation of endothelial cells with H2
induced an increase in monolayer permeability (Siflinger-Birnboim et al., 1996
). Our published results showed that ROS production mediates iron oxide nanoparticle-induced permeability in HMVEC (Apopa et al., 2009
In this study, it was also found that PFOS-induced endothelial permeability may be associated with remodeling of actin filaments. Actin is one of the most abundant proteins in eukaryotic cells. There are two different forms of actin, monomeric actin (G-actin) and actin filaments (F-actin). Actin filaments are involved in a wide variety of cellular processes, including cell permeability, cell motility, cell cycle control, cellular structure, and cell signaling (Schmidt and Hall, 1998
). They function in cellular processes by undergoing a dynamic structural remodeling as well as continuing polymerization and depolymerization, leading to the formation of discrete structures at the cell periphery for attachment to the substratum in response to extracellular signal transduction (Qian et al., 2002
). Compelling evidence showed that actin filament remodeling plays a major role in determining the structural integrity of the confluent endothelial monolayer (Lee and Gotlieb, 2003
). In endothelial cells, actin filaments are the key components of cell junctions and adhesions. Actin filaments undergo dynamic remodeling to form DPB in maintaining cell-cell adhesion, central stress fibers to develop strong substrate anchorage, and lamellipodia and filopodia for cell spreading and motility (Lee and Gotlieb, 2002
). Mehta and Malik (2006)
demonstrated that the disruption of actin filaments is directly related to an increase in endothelial cell permeability. Studies also found that both the reduction of DPB and the induction of stress fibers are associated with agonist-induced cell permeability changes (Vyalov et al., 1996
). The results from demonstrate that exposure of HMVEC to PFOS disrupted the integrity of DPB and induced the formation of lamellipodia, filopodia and stress fibers, which may generate the force to pull monolayer HMVEC apart. The results also demonstrate that the removal of ROS by catalase abolished PFOS-induced actin filament remodeling and HMVEC monolayer separation, resulting in an inhibition of PFOS-induced increase of endothelial permeability. Furthermore, PFOS-induced actin filament remodeling was concurrent with PFOS-induced endothelial cell permeability. Therefore, our results suggest that PFOS increases endothelial monolayer permeability through the production of ROS, which may in turn stimulate remodeling of actin filaments. This scheme is summarized in .
Schematic representation of signal transduction from PFOS stimulation to cell permeability
In the current study, dihydroethidium dye was used to determine the PFOS-induced production of ROS in HMVEC. Future studies are needed to examine a) the source of ROS production and b) the mechanism by which ROS produce actin filament remodeling and endothelial cell permeability.