oxidant-induced cell death is a mediator of pathologies associated with ischemia-reperfusion, toxicants, and inflammation in many different organs. Reactive oxygen species (ROS) react with several intracellular targets including DNA, proteins, and lipids. Lipid peroxidation changes the biophysical properties of membrane phospholipids, which can affect ion permeability (17) and membrane fluidity (8). Glycerophospholipids in cell and organelle membranes contain two fatty acyl chains linked to the glycerol backbone at the sn-1 and sn-2 positions. The sn-2 fatty acids have a higher degree of unsaturation than sn-1 fatty acids making them a primary target of ROS in biological membranes. Phospholipase A₂ enzymes specifically catalyze the cleavage of the sn-2 bond of phospholipids and have been hypothesized to selectively cleave oxidized fatty acids from the sn-2 position to preserve membrane integrity (41).

There are more than 20 isoforms of PLA₂ with different characteristics including Ca²⁺ requirement and subcellular localization. Six and Dennis (35) classified PLA₂ enzymes into 14 groups based on their nucleotide sequence. Most PLA₂ in these groups are relatively small proteins (12–19 kDa), require millimole amounts of Ca²⁺ for activity, and use a histidine for catalysis. These groups (I-III, V, and IX-XIV) have historically been called secreted PLA₂ (sPLA₂). The remaining groups consist of Ca²⁺-dependent cytosolic PLA₂ (cPLA₂; group IVA and B), platelet-activating factor acetylhydrolases (PAF-AH; groups VII and VIII), and Ca²⁺-independent PLA₂ (iPLA₂, group VI and IVC).

cPLA₂γ (group IVC) contains sufficient nucleotide homology to cPLA₂α to be classified as a group IV enzyme, but lacks the Ca²⁺-dependent lipid binding domain and is anchored to endoplasmic reticulum, golgi apparatus, and mitochondrial membranes (38, 40). iPLA₂β (group VIA) is predominately cytosolic, but one of the several splice variants of iPLA₂β associates with crude membrane fractions (22). iPLA₂γ (group VIB) appears to be exclusively membrane bound (20, 24, 25, 44). The 63-kDa isoform of iPLA₂γ has been localized in peroxisomal membranes of rat liver cells (44) and higher molecular weight isoforms have been identified in microsomes of rabbit renal proximal tubular cells (RPTC) and ventricular myocytes (20) and in rat heart mitochondria (25).

One method used to identify the isoform(s) of PLA₂ responsible for an observed activity is to determine the sensitivity of the activity to different PLA₂ inhibitors (10, 15, 20). iPLA₂β is sensitive to methyl arachidonyl fluorophosphonate (MAFP), arachidonyl trifluoromethylketone (AACOCF₃), and S-BEL, but not R-BEL at low micromolar concentrations (2, 15, 23). iPLA₂γ is sensitive to R-BEL, but not MAFP or S-BEL at low micromolar concentrations (15, 20). cPLA₂γ is inhibited by MAFP and AACOCF₃, but not BEL (38). Finally, PAF-AH is potently inhibited by MAFP and not inhibited by BEL at concentrations up to 20 μM (18).

Recent studies showed that mitochondrial PLA₂ activity in several different tissues and species is Ca²⁺ independent (6, 7, 13, 25, 40, 43). In rat liver mitochondria, the iPLA₂ activity was completely inhibited by BEL but not by inhibitors of sPLA₂ or cPLA₂ (6). In rabbit ventricular myocytes, mitochondrial iPLA₂ expression was demonstrated in the inner mitochondrial membrane (43). Several studies have utilized a commercially available iPLA₂β antibody to demonstrate iPLA₂ expression in mitochondria (6, 7, 43). Recently, Gross et al. (25) reported that iPLA₂γ is expressed in rat heart mitochondria based on immunoblot analysis with an iPLA₂γ antibody. Finally, immunohistochemical analysis revealed that cPLA₂γ colocalizes with mitochondria in immortalized mouse lung fibroblasts (40). In summary, iPLA₂γ, iPLA₂β, and cPLA₂γ are reportedly localized to mitochondria in different tissues, but the presence, identity, and role of a mitochondrial iPLA₂ in the kidney has not been examined.

Several studies have implicated PKC-mediated phosphorylation as a mechanism for activation of membrane-bound iPLA₂ in the heart and kidney (11, 28, 37). Membrane-associated iPLA₂ activity in ventricular myocytes and human coronary artery endothelial cells is increased in a diacylglycerol-dependent, Ca²⁺-independent fashion suggesting the involvement of a novel PKC isoform (28, 37). In ventricular myocytes, PKCε is the only novel isoform detected in the membrane fraction (37). Endoplasmic reticulum iPLA₂γ activity in RPTC is increased by phorbol 12-myristate 13-acetate (PMA) treatment, which mimics the effect of diacylglycerol (11). To date, no studies have examined the regulation of mitochondrial iPLA₂ activity by PKC.

We suggested that iPLA₂γ acts to repair or prevent lipid peroxidation which protects cells from oxidative stress. This hypothesis is supported by the finding that inhibition of iPLA₂γ activity in RPTC potentiates oxidant-induced lipid peroxidation and necrotic cell death (10). Furthermore, doxorubicin- and tert-butyl hydroperoxide (TBHP)-induced cardiomyocyte death is enhanced by inhibition of iPLA₂ with BEL (27, 39). In this study, we determined that RPTC possess mitochondrial iPLA₂ activity and identified the isoform responsible for the activity. Several potential mechanisms for regulation of mitochondrial iPLA₂ activity were examined and the role for mitochondrial iPLA₂ in oxidant-induced mitochondrial lipid peroxidation and swelling was investigated.

MATERIALS AND METHODS

Materials.

Female New Zealand White rabbits (1.5–2.0 kg) were purchased from Myrtle’s Rabbitry (Thompson Station, TN). R- and S-BEL were generously provided by B. S. Cummings (University of Georgia, Athens, GA) (20) or purchased from Cayman Chemical (Ann Arbor, MI). The iPLA₂β antibody was purchased from Cayman Chemical. The PKCε-specific inhibitor PKCε V1–2 was purchased from Biomol (Plymouth Meeting, PA). All other chemicals and materials were obtained from Sigma (St. Louis, MO) or reported previously (10, 11, 23).

Isolation of renal cortical mitochondria.

Rabbits were euthanized by intravenous injection of 75 mg/kg pentobarbital sodium and kidneys were removed by blunt dissection using a protocol approved by the Medical University of South Carolina IACUC. Human cadaveric kidneys (International Bioresearch Solution, Tucson, AZ) that were rejected for transplant were used within 24 h of removal from the donor according to the Medical University of South Carolina IRB guidelines. Kidney cortex tissue was collected and placed on ice in mitochondrial isolation buffer containing (in mM): 270 sucrose, 5 Tris·HCl, 1 EGTA (pH 7.4). Rabbit and human renal cortical mitochondria (RCM) were isolated by differential centrifugation and purified by Percoll density gradient separation where noted (3). Mitochondrial inner membrane fractions were isolated from Percoll-purified mitochondria as described by our laboratory (3).

Immunoblot analysis.

We contracted Aves Labs (Tigard, OR) to generate the anti-rabbit iPLA₂γ antibody using the rabbit-iPLA₂γ-specific peptide sequence, CENIPLDESRNEKLDQ (20). Resultant antiserum was affinity purified and used as the primary antibody. Equal amounts of mitochondrial and cytosolic protein (25 μg) were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were incubated with the iPLA₂γ or iPLA₂β antisera at a dilution of 1:1,000. Bound antibodies were visualized by chemiluminescence detection on a ChemiImager 5500 imager (Alpha Innotech, San Leandro, CA).

Mitochondrial swelling.

After incubation with cis-parinaric acid and washing, isolated mitochondria were suspended at a concentration of 0.4 mg protein/ml in swelling buffer (150 mM KCl, 20 mM Tris·HCl, pH 7.4) in a 96-well plate and incubated with diluent, inhibitors, or antioxidants at room temperature for 10 min. Ferrous sulfate heptahydrate (Sigma) to achieve a final Fe²⁺ concentration of 10 μM or diluent (swelling buffer) was then added to initiate oxidative stress (14). Mitochondrial swelling was measured using a SpectraMax 190 spectrophotometric plate reader (Molecular Devices, Sunnyvale, CA) as the loss of optical density at 540 nm over time as previously described (14). Swelling buffer was deoxygenated immediately before use to prevent immediate oxidation of Fe²⁺ to Fe³⁺ (14).

Measurement of cis-parinaric acid oxidation.

Lipid peroxidation in isolated mitochondria was measured using the fluorescent lipid, cis-parinaric acid as described previously (21, 29), with modifications. Isolated mitochondria were suspended at a concentration of 1 mg protein/ml in swelling buffer and incubated on ice with cis-parinaric acid (6.4 μM) for 10 min. The mitochondria were pelleted by centrifugation, the supernatant discarded, and the mitochondria were resuspended (0.4 mg protein/ml) in deoxygenated swelling buffer. Mitochondria were added to a 96-well plate and incubated with diluent, inhibitors, or antioxidants at room temperature for 10 min and then treated with 10 μM Fe²⁺. Lipid peroxidation was measured as the loss of cis-parinaric acid fluorescence (excitation 320 nm, emission 405 nm) over time using a Fluoroskan Ascent fluorescent plate reader (Thermo Labsystems, Franklin, MA).

Isolation of rabbit RPTC, culture conditions, and inhibitor treatment.

Rabbit RPTC were isolated using the iron oxide perfusion method and grown under improved conditions as previously described (30). Confluent monolayers were treated with R- or S-BEL or diluent control for 30 min before harvesting and isolation of mitochondria. RPTC mitochondria and cytosol were isolated by differential centrifugation as described previously (20).

Measurement of iPLA₂ activity.

PLA₂ activity was determined under linear reaction conditions in mitochondria as described previously (10, 11, 20). Activity was measured using synthetic (16:0, [³H]18:1) plasmenyl- or phosphatidylcholine (100 μM) in the presence (1 mM CaCl₂) and absence of Ca²⁺ (4 mM EGTA). For PLA₂ activity inhibition studies, mitochondrial samples were incubated with solvent control [DMSO < 0.1% (vol/vol)], racemic, R-, or S-BEL, MAFP, or AACOCF₃ for 10 min before the addition of the phospholipid substrate to start the reaction.

Protein determination.

Protein determination was performed using the bicinchoninic acid assay method as described by Sigma.

Statistical analysis.

Mitochondria, cytosol, or RPTC isolated from one rabbit or human kidney represent one experiment (n = 1). The appropriate ANOVA was performed for each data set using SigmaStat statistical software. Individual means were compared using Fisher’s protected least significant difference test with P ≤ 0.05 being considered indicative of a statistically significant difference between mean values.

RESULTS

Rabbit RCM were assayed for PLA₂ activity using synthesized plasmenylcholine and phosphatidylcholine phospholipids with palmitic acid (16:0) at the sn-1 position and [³H]oleic acid (18:1) at the sn-2 position. Percoll-purified RCM and RCM inner membrane possess PLA₂ activity that is not dependent on Ca²⁺ and has a preference for plasmenylcholine (Fig. 1A). The effect of pH on iPLA₂ activity in Percoll-purified RCM is presented in Fig. 1B. A significant increase in plasmenylcholine cleavage is observed as the pH decreases from 8 to 7.5, while a significant increase in phosphatidylcholine cleavage occurs as the pH decreases from 9 to 8. Cleavage of both substrates decreases to background levels at a pH of ∼6. Ca²⁺-independent activity was observed in the endoplasmic reticulum (ER) and inner mitochondrial membrane of kidney cortex tissue from cadaveric human kidneys (Fig. 1C). Similar to rabbit kidney (10) there was little basal cytosolic PLA₂ activity in human kidney cortex. In summary, RCM possess iPLA₂ activity that prefers plasmenylcholine over phosphatidylcholine, the activity is differentially affected by pH, and human kidney cortex mitochondria exhibit a similar level of iPLA₂ activity as the ER and mitochondrial inner membrane in rabbit kidney cortex.

Immunoblot analysis with antibodies for iPLA₂β and iPLA₂γ revealed immunoreactive proteins to both antibodies in rabbit RCM (Fig. 2). While the bands are similar in size, the antigenic peptides used for both antibodies are unique for their respective proteins. The protein recognized by the Cayman iPLA₂β antibody is similar in size to iPLA₂β reported in rabbit heart mitochondria (43) and rat liver mitochondria (6). iPLA₂γ in RPTC mitochondria is the long isoform of iPLA₂γ (∼88 kDa), based on the nucleotide sequence analysis of Mancuso et al. (24). Consistent with the very low cytosolic PLA₂ activity, no iPLA₂ expression was detected in RPTC cytosol.

In an effort to determine which PLA₂ isoform is responsible for the iPLA₂ activity in RCM, the inhibitor sensitivity profile was determined. RCM iPLA₂ activity was inhibited by racemic BEL, but not MAFP or AACOCF₃ (Fig. 3A). This profile is consistent with iPLA₂γ, and not iPLA₂β or cPLA₂γ, both of which are inhibited by MAFP and AACOCF₃ at concentrations lower than 10 μM (2, 23, 38). As an additional approach to confirm that iPLA₂γ is responsible for the mitochondrial PLA₂ activity, the effect of the R- and S-enantiomers of BEL was investigated. R-BEL selectively inhibits iPLA₂γ at low micromolar concentrations while S-BEL selectively inhibits iPLA₂β (15). In Percoll-purified RCM, R-BEL was effective at inhibiting iPLA₂ activity while S-BEL had no effect (Fig. 3B). In mitochondria isolated from confluent RPTC, after a 30-min incubation with increasing concentrations of inhibitors or diluent control, R-BEL was significantly more effective at inhibiting iPLA₂ activity than S-BEL (Fig. 3C). These results demonstrate that iPLA₂γ is responsible for mitochondrial iPLA₂ activity in RCM and that treatment of RPTC with R-BEL selectively decreases mitochondrial iPLA₂γ activity.

The addition of ATP (1 mM) to RCM increased iPLA₂ activity ∼30% (Fig. 4). Previous investigators suggested that membrane-associated iPLA₂γ activity is regulated by PKCε-mediated phosphorylation (11, 28, 37). and the addition of the PKCε-specific inhibitor PKCε V1–2 blocked ATP-induced increases in iPLA₂ activity in RCM (Fig. 4A). To ensure that ATP was not activating latent iPLA₂β in the mitochondria, the effect of R-BEL and S-BEL (5 μM) on ATP-induced iPLA₂ activity was determined. S-BEL had no inhibitory effect on iPLA₂ activity in the presence or absence of ATP while R-BEL inhibited iPLA₂ activity under both conditions, suggesting that the increase in iPLA₂ activity in the presence of ATP is solely due to an increase in iPLA₂γ activity and not the activation of a latent iPLA₂β (Fig. 4B).

To investigate the role of iPLA₂γ in oxidant-mediated mitochondrial damage, RCM were treated with Fe²⁺ to initiate oxidant stress and mitochondrial swelling (Fig. 5) (14). Mitochondrial swelling induced by 10 μM Fe²⁺ was completely blocked by the antioxidants butylated hydroxyanisole (25 μM BHA; Fig. 5A) and N,N′-diphenyl-1,4-benzenediamine (DPPD; 5 μM, data not shown). Pretreatment of mitochondria with R-BEL (8 μM) accelerated the onset of Fe²⁺-induced mitochondrial swelling, whereas pretreatment with S-BEL (8 μM) had no effect (Fig. 5A). The time required for Fe²⁺ to induce maximal mitochondrial swelling in the presence of R- and S-BEL was compared with the time required for Fe²⁺ alone to induce maximal swelling. Pretreatment with R-BEL decreased the time required for Fe²⁺ to initiate mitochondrial swelling by ∼120 s, while S-BEL had no effect (Fig. 5B). Pretreatment with cyclosporine A (CsA; 1 μM) had no effect on Fe²⁺-induced mitochondrial swelling (data not shown), suggesting the Fe²⁺-induced swelling is not the result of mitochondrial permeability transition (MPT) pore opening. In summary, Fe²⁺ induced RCM swelling that is mediated by oxidative stress, but not the result of MPT, and specific iPLA₂γ inhibition accelerated the oxidant-induced swelling.

The fluorescent lipid cis-parinaric acid was used to measure the rate of lipid peroxidation under the same conditions as Fig. 5A. cis-Parinaric acid incorporates into membranes and loses fluorescence as it becomes oxidized (21, 29). Treatment with Fe²⁺ resulted in an increased rate of cis-parinaric acid oxidation compared with control mitochondria. The Fe²⁺-induced oxidation rate was further increased by R-BEL pretreatment; however, pretreatment with S-BEL had no effect (Fig. 6). BHA and DPPD significantly inhibited cis-parinaric acid oxidation in a concentration-dependent manner (Fig. 6 and data not shown, respectively). Similar to Fe²⁺-induced swelling, iPLA₂γ inhibition significantly accelerated the rate of oxidant-induced lipid peroxidation in RCM.

Previous studies in RPTC demonstrated that oxidants directly inactivate ER-iPLA₂γ (12). To determine the effect of oxidative stress on mitochondrial iPLA₂γ activity, RCM were incubated with diluent control or 10 μM Fe²⁺ and assayed for iPLA₂ activity over time. Similar to ER-iPLA₂γ, oxidant treatment significantly inhibited iPLA₂γ activity in RCM (Fig. 7).

DISCUSSION

In this study, we demonstrate the novel finding that human and rabbit RCM possess iPLA₂ activity. The preference of RCM iPLA₂ for plasmenylcholine phospholipid substrates over phosphatidylcholine substrates is similar to iPLA₂γ in microsomes of RPTC and rabbit ventricular myocytes (10, 20). The localization of kidney mitochondrial iPLA₂ activity to the inner mitochondrial membrane is in agreement with a previous report in rabbit heart mitochondria (43). While immunoblot analysis results suggested that iPLA₂γ and iPLA₂β are localized to RPTC mitochondria, inhibitor sensitivity analysis revealed that in rabbit RCM and RPTC mitochondria iPLA₂ activity is due solely to the iPLA₂γ isoform and not iPLA₂β or cPLA₂γ.

The iPLA₂ isoform in rabbit heart mitochondria and rat liver mitochondria was hypothesized to be iPLA₂β based on iPLA₂β immunoblot results and sensitivity of the activities to racemic BEL (6, 43). The iPLA₂β antibody also detected an immunoreactive protein of ∼85 kDa in the microsomal fraction (10) and in mitochondria of RPTC. This size protein is similar to those reported by others in the mitochondria (6, 43). However, the presence of iPLA₂β in RPTC is not supported by RT-PCR (11) or inhibitor sensitivity analysis of the iPLA₂ activity in RPTC microsomes (10, 20) and mitochondria. In contrast, iPLA₂γ expression has been confirmed at the mRNA (11) and protein level (20) (Fig. 2) and the inhibitor sensitivity profile, using racemic, R- and S-BEL, MAFP and AACOCF₃, of microsomes (10, 20) and mitochondria (Figs. 3 and 4) from RPTC confirms iPLA₂γ. The concentrations of MAFP and AACOCF₃ used (10 μM each) are well above the IC₅₀ values reported for iPLA₂β, cPLA₂γ, and PAF-AH (2, 15, 18, 23). In summary, immunoblot analysis results prohibit ruling out iPLA₂β expression in RPTC, but RT-PCR and inhibitor sensitivity of the iPLA₂ activity in RPTC raise questions about the identity of the protein recognized by the iPLA₂β antibody in RPTC.

Regulation of iPLA₂γ activity has not been extensively studied. The pH of the environment surrounding iPLA₂γ is a potential regulator of its activity. Our experiments revealed that mitochondrial iPLA₂ activity increases toward plasmenylcholine substrates but is unchanged toward phosphatidylcholine substrates as pH decreases from 8 to 7.5. This is relevant because a significant portion of rabbit RPTC phospholipids is made up of plasmalogens (plasmenylcholine and plasmenylethanolamine) and these phospholipids are enriched with arachidonic acid (20:4) at the sn-2 position (32). Mitochondrial matrix pH is ∼8 and decreases toward 7.5 as the inner mitochondrial membrane potential is lost and by agents that increase cytosolic Ca²⁺ (1). During mitochondrial stress (i.e., ischemia-reperfusion, oxidative stress, Ca²⁺ overload), the decrease in matrix pH could increase iPLA₂γ activity to cleave arachidonic acid containing plasmalogens, which are targets of ROS in cellular membranes (19). Changes in pH may result in conformational changes of iPLA₂γ, phospholipids, or both leading to the observed changes in activity.

Previous observations led to the hypothesis that microsomal iPLA₂ activity in kidney, heart, and coronary artery endothelial cells is regulated by PKC-mediated phosphorylation (11, 28, 37). Our recent studies demonstrated that the microsomal iPLA₂ in kidney and heart is iPLA₂γ (20). In support of this hypothesis, sequence analysis of iPLA₂γ revealed the presence of multiple potential serine/threonine phosphorylation sites (5). The addition of ATP to isolated RCM increased iPLA₂γ activity, and the increase in activity was blocked by pretreatment with a PKCε-specific inhibitor, suggesting that PKCε regulates mitochondrial iPLA₂γ activity. Our findings in conjunction with the observation by Nowak et al. (31) that PKCε is present in RPTC mitochondria and that PKCε translocation to mitochondria is increased after oxidant treatment support the hypothesis that oxidant-induced PKCε translocation may result in upregulation of mitochondrial iPLA₂γ activity in response to oxidant stress in renal cells.

Previous studies from our laboratory revealed that inhibition of iPLA₂ activity with BEL in rabbit RPTC-potentiated oxidant-induced lipid peroxidation and necrotic cell death (10) and inhibition of ventricular myocyte iPLA₂ activity with BEL potentiated doxorubicin- and TBHP-induced cell death (27, 39). These findings led to the hypothesis that iPLA₂γ prevents or repairs lipid peroxidation. In this study, we used the iPLA₂γ inhibitor, R-BEL, to show that specific inhibition of iPLA₂γ in isolated RCM accelerated oxidant-induced lipid peroxidation and mitochondrial swelling. These observations support our hypothesis and demonstrate that mitochondrial iPLA₂γ activity protects mitochondria against oxidative stress. Our previous demonstration that iPLA₂γ is expressed in ER of rabbit RPTC and ventricular myocytes (20) is consistent with the idea that iPLA₂γ localized to these two organelles, which are routinely exposed to oxidants, serves as a defense mechanism against lipid peroxidation-induced organelle dysfunction.

A protective role for iPLA₂β in mitochondria has recently been proposed. Seleznev et al. (34) reported that overexpression of iPLA₂β in insulinoma cells and Chinese hamster ovary cells results in mitochondrial localization of the GFP-linked iPLA₂β and protects against staurosporine-induced apoptosis. Staurosporine-induced apoptosis is at least partially mediated by mitochondrial ROS production and lipid peroxidation (34). The expression and localization of different iPLA₂ isoforms vary greatly among different cell types and the role of lipid peroxidation repair may be determined by the available isoforms in a specific cell or organelle.

The ER iPLA₂γ in RPTC is directly inactivated by diverse oxidants in a dithiothreitol-sensitive manner, implicating reduced thiols are required for iPLA₂γ activity and are a target of oxidants (12). Our current results demonstrate that mitochondrial iPLA₂γ also is inactivated by oxidants. Recently, iPLA₂β also was shown to be inactivated by oxidants and the specific oxidant-induced damage was elucidated (36). Oxidant-induced inhibition of iPLA₂ may be a form of signaling (negative regulation) to prevent excessive activity. Conversely, it may represent an additional mechanism of oxidant-induced toxicity (i.e., oxidants cause lipid peroxidation and inactivate the lipid peroxidation repair enzyme). Prevention of oxidant inactivation of iPLA₂γ may provide protection to cells and organelles during oxidative stress.

In conclusion, we demonstrated that iPLA₂γ is expressed and active in RPTC mitochondria, and inhibition of iPLA₂γ accelerates lipid peroxidation and swelling in isolated RCM. Human kidney cortex displays significant iPLA₂ activity in the inner mitochondrial membrane and ER similar to rabbit kidney cortex. Our data suggest that mitochondrial iPLA₂γ activity is regulated by pH and phosphorylation by PKCε. Finally, similar to iPLA₂γ in RPTC ER, mitochondrial iPLA₂γ is inactivated by oxidant stress. Efforts to preserve and enhance mitochondrial iPLA₂γ activity may be useful to prevent oxidative stress-induced cell death in the proximal tubular cells of the kidney and in other tissues, including the heart.

GRANTS

This research was supported by National Institutes of Health Grant DK-62028 to R. G. Schnellmann and G. R. Kinsey was supported by a training grant from the National Institute of Environmental Health Sciences (NIEHS), NIH (T32 ES-012878) and MUSC animal facilities were funded by NIH Grant no. C06 RR015455. Its contents are solely the responsibility of the authors and do not represent the official views of the NIEHS, NIH.

FOOTNOTES

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.