Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

N-acetylcysteine (NAC), a thiol reducing agent, is often used as an "antioxidant" in experimental models of acute and chronic lung injury, including those associated with endotoxin (1), hypoxia (2), bleomycin (3), and paraquat (4). In addition, because reactive oxygen radicals are considered to play an important role in the pathogenesis of various pulmonary disorders, such as idiopathic pulmonary fibrosis, acute respiratory distress syndrome, and cystic fibrosis, therapeutic enhancement of antioxidant defense mechanisms in these disorders have been attempted (5). Indeed, administration of NAC often causes a time-dependent increase in the cell content of glutathione (GSH), the most abundant thiol antioxidant. However, reagents that are traditionally considered to be antioxidants may undergo auto-oxidation and thus may also behave as an oxidant, depending upon the local milieu (6, 7). Thus, in conditions in which lung GSH is already depleted, NAC may auto-oxidize and behave as an oxidant. This phenomenon is one plausible reason why the potential beneficial effects of NAC have been controversial and at times contradictory (8, 9). In a recent review, Domenighetti and coworkers (9) astutely noted that "a direct in vivo antioxidative effect of NAC remains to be established in humans" and that "the mode of action of NAC may not be the same in different pathologies and clinical situations."

Reactive oxygen species (ROS), such as superoxide (O₂^.), hydrogen peroxide (H₂O₂), and hydroxyl radical (^.OH) may function as second messengers in signal transduction (10, 11). Although mitogen-activated protein kinases (MAPKs) may be directly activated by oxidants such as H₂O₂ (12), the role of ROS in the activation of the MAPKs by cytokines is largely inferred on the basis of the inhibition of MAPKs by NAC (13). Natoli and coworkers (11) previously showed that NAC was an inhibitor of c-Jun NH₂-terminal kinase (JNK) with tumor necrosis factor (TNF)- stimulation. Moreover, we reported that under serum-deprived conditions (0.1% fetal bovine serum [FBS]), NAC inhibited the activation of extracellular signal-regulated kinase (ERK), p38^mapk, and JNK by TNF- (14). In PC12 cells, serum deprivation was demonstrated to activate JNK, and this activation was abrogated by NAC (15). In contrast, other studies have found that NAC has no effect on MAPK activation (16). Further, Yan and Greene (17) showed that NAC promoted survival of neuronal cells by activation of ERK. Lipopolysaccharide (LPS) is a cell wall glycolipid from gram-negative bacteria that is known to induce a wide variety of inflammatory genes in sepsis and acute lung injury, many of which are mediated by the MAPKs. Because the level of serum in experimental cell cultures may have an effect on the redox state of the cellular milieu, we sought to determine in this in vitro study whether there is a differential effect of NAC on LPS activation of the MAPKs under serum-replete versus serum-deprived conditions.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

RAW 264.7 macrophages were used for all the studies. FBS, heat-inactivated and determined to be LPS-free, was purchased from Atlanta Biologicals (Atlanta, GA). LPS purified from Salmonella typhimurium was purchased from Sigma (St Louis, MO). GSH-sepharose beads were purchased from Pharmacia (Piscataway, NJ). Enhanced chemiluminescence assay kits were obtained from Amersham Life Sciences (Arlington Heights, IL). Recombinant c-Jun_1-79-GSH-S-transferase (GST) was kindly provided by Dr. Gary Johnson (Department of Pharmacology, University of Colorado Health Sciences Center, Denver, CO). Antiphosphospecific p38^mapk antibody was purchased from New England Biolabs (Beverly, MA). Rabbit polyclonal antiphosphospecific ERK antibody was purchased from Promega (Madison, WI). Rabbit polyclonal anti-p46 JNK, rabbit polyclonal anti-p42 ERK, and mouse monoclonal antiphosphospecific p46-p54 JNK antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). [³²P]--adenosine triphosphate (ATP) (> 3,000 Ci/mmol) was purchased from NEN Research Products DuPont (Wilmington, DE). All other reagents were of the highest purity.

Culture and Stimulation of RAW 264.7 Macrophages

RAW 264.7 cells were grown in 75-cm² polystyrene tissue culture flasks in medium. During log-phase growth, the cells were scraped off the flasks and the cell suspension was added to six-well polystyrene tissue culture plates at 200,000 cells/ml of culture medium (4 ml/well). After 24 h of growth at 37°C in 5% CO₂, the cells were preincubated with 30 mM NAC for 2 h in either 10% FBS or 0.1% FBS-containing media, followed by addition of LPS at a final concentration of 100 ng/ml for an additional hour. After stimulation, the cells were washed once with N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (Hepes)-saline buffer, followed by lysis of cells with 500 µl ice-cold lysis buffer (50 mM Tris/HCl, pH 8.0, containing 137 mM NaCl, 10% [vol/vol] glycerol, 1% [vol/ vol] Nonidet P-40, 1 mM NaF, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 2 mM Na₃VO₄, and 1 mM phenylmethylsulfonyl fluoride). After centrifugation at 20,000 × g for 10 min, the supernatants were used for in vitro kinase assay and Western blotting.

Determination of JNK Activity

For measurement of JNK activity, RAW 264.7 cell monolayers were lysed at 4°C with 500 µl of ice-cold lysis buffer. After the protein content was normalized between samples, JNK in each sample of lysate was bound to 15 µl of a 1:1 slurry of lysis buffer/ GST-c-Jun_1-79-sepharose beads and incubated at 4°C for 2 h. The beads were then washed twice with 500 µl lysis buffer and twice with 500 µl of JNK buffer (20 mM Hepes buffer, pH 7.2, containing 30 mM -glycerophosphate, 10 mM p-nitrophenylphosphate, 10 mM MgCl₂, 0.5 mM dithiothreitol [DTT], and 50 µM Na₃VO₄). The activity of JNK was detected by phosphorylation of c-Jun-GST in an in vitro kinase assay and was assessed by incorporation of [³²P]-ATP (10 µCi/sample) in JNK buffer incubated at 30°C for 30 min. The kinase reactions were then stopped with an equal volume of 2× Laemmli sample buffer containing 20 mM DTT and boiled for 3 min. The proteins present in the supernatants were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) through a 12% polyacrylamide gel and transferred onto nitrocellulose membranes. [³²P]-labeled c-Jun-GST was detected by autoradiography.

Western Blot Analysis

To determine the phosphorylation of ERK, p38^mapk, and JNK, whole-cell lysates were separated by SDS-PAGE and transferred onto nitrocellulose membranes. The blots were then washed in Tris-Tween buffered saline (TTBS) (20 mM Tris/HCl buffer, pH 7.6, containing 137 mM NaCl and 0.05% [vol/vol] Tween 20), blocked overnight with 5% (wt/vol) nonfat dry milk, and probed with phosphospecific antibodies to p42/p44 ERK, p38^mapk, and p46-p54 JNK in 5% (wt/vol) bovine serum albumin dissolved in TTBS. Bound antibodies were detected by enhanced chemiluminescence using horseradish peroxidase-conjugated secondary antirabbit antibody. To determine equal loading of proteins between samples, the corresponding membranes were probed with rabbit polyclonal p42/p44 ERK, p38^mapk, and p46 JNK antibodies.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

LPS Activates ERK, p38^mapk, and JNK

Activation of MAPKs is dependent upon the phosphorylation of specific threonine and tyrosine residues on TXY tripeptide motifs that are specific for each MAPK: TEY for ERK, TGY for p38^mapk, and TPY for JNK. Before determining the effects of NAC on LPS activation of the MAPKs, monolayers of RAW 264.7 cells were stimulated with LPS alone at concentrations from 0.1 to 500 ng/ml for a previously determined optimal time of 1 h. As shown in Figure 1, LPS induced the phosphorylation of ERK and p38^mapk in a concentration-dependent fashion beginning at ~ 50 ng/ml, and of JNK at much lower concentrations of LPS (0.1 ng/ml).

View larger version (63K): [in this window] [in a new window]   Figure 1.   LPS phosphorylates ERK, p38mapk, and JNK. RAW 264.7 cells were stimulated with LPS at the indicated concentrations for an optimal time of 1 h as previously determined. A total of 20 ng of protein in the nuclear-free lysate fraction for each condition was immunoblotted with specific antibodies that recognize Thr and Tyr phosphorylated forms of (A) ERK, (B) p38mapk, and (C) JNK. The same nitrocellulose membranes were then immunoblotted with antibodies for p42/p44 ERK, p38mapk, and p46-p54 JNK to show equal loading of protein (bottom panels of A-C). Data are representative of three independent experiments.

NAC Differentially Affects LPS-Induced ERK, p38^mapk, and JNK Activation in the Absence or Presence of Serum

To determine whether the presence of serum alters the effects of NAC on MAPK activation, we stimulated the macrophages with LPS plus NAC in media that contained either 10% or 0.1% serum. It is important to emphasize that the stock LPS used contained 10% FBS, and thus there was no functional deficiency of LPS-binding protein (LBP) in all experimental conditions. We confirmed this in a functional assay in which we quantified the amount of nitrite (NO₂) produced in cells cultured in media containing either 10% or 0.1% FBS and stimulated with interferon- plus stock LPS reconstituted in media containing 10% FBS. We found that there was no difference in the amount of NO₂ produced in 10% versus 0.1% FBS-containing media provided the stock LPS did not lack LBP (data not shown). In contrast, there was essentially no NO₂ produced when stock LPS was reconstituted in serum-free media and the culture media was serum-deprived (0.1% FBS) (data not shown).

Macrophages were pretreated with NAC at 30 mM for 2 h before costimulation of cells with NAC and 100 ng/ml of LPS for 1 h. After the cells were lysed and the lysates were normalized for protein content, the proteins were separated by SDS-PAGE. This was followed by immunoblot of the separated proteins with phosphospecific antibodies to ERK and p38^mapk. In identically treated samples, whole-cell lysates were subjected to a solid-phase in vitro kinase assay in the presence of [³²P]-ATP using c-Jun- GST beads as the substrate. As shown in Figures 2-4, there is induction of p42-p44 ERK and p38^mapk phosphorylation and JNK activation in the presence of LPS in media that contained either 10% or 0.1% FBS (compare lanes 1 and 4 versus lanes 2 and 5, respectively, in Figures 2-4). However, whereas NAC inhibited LPS-induced ERK phosphorylation in serum-deprived media, there was no such inhibitory effect in the presence of 10% serum (Figure 2, lane 6 versus lane 3). NAC also inhibited LPS-induced p38^mapk phosphorylation in serum-deprived media, but in the presence of 10% serum there was a modest increase in LPS-induced p38^mapk phosphorylation with NAC treatment (Figure 3, lane 6 versus lane 3). JNK activity was also inhibited by NAC in serum-deprived media but was in fact augmented by NAC in serum-replete media as shown in Figure 4, lane 6 versus lane 3. The bottom panels of Figures 2 and 3 are immunoblots with anti-ERK 1/2 and anti-p38^mapk, respectively, and show that there was equal loading of samples. The bottom panel of Figure 4 is an immunoblot of the in vitro JNK activity assay with p46 JNK1 antibody and shows that p46 JNK1 binding to c-Jun-GST correlated with JNK activity, as we and others have previously shown (18).

View larger version (45K): [in this window] [in a new window]   Figure 2.   Differential effects of NAC on LPS-induced phosphorylation of ERK in serum-replete (10% FBS) versus serum-deplete (0.1% FBS) conditions. Note that in serum-deplete conditions NAC completely inhibited LPS-induced ERK phosphorylation (compare lanes 5 and 6, top panel), whereas NAC had no effect on ERK phosphorylation in serum-replete conditions (compare lanes 2 and 3, top panel). Immunoblot with p42-p44 ERK antibody revealed equal loading of protein (bottom panel). Data are representative of three independent experiments.

View larger version (43K): [in this window] [in a new window]   Figure 3.   Differential effects of NAC on LPS-induced phosphorylation of p38mapk in serum replete (10% FBS) versus serum-deplete (0.1% FBS) conditions. In serum-deplete conditions, NAC inhibited LPS-induced p38mapk phosphorylation completely (compare lanes 5 and 6, top panel). In contrast, there was no inhibition and in fact a slight augmentation of LPS-induced p38mapk phosphorylation by NAC in serum-replete conditions (compare lanes 2 and 3, top panel). Immunoblot with p38mapk antibody revealed equal loading of protein (bottom panel). Data are representative of three independent experiments.

View larger version (43K): [in this window] [in a new window]   Figure 4.   Differential effects of NAC on LPS-induced activation of JNK in serum-replete (10% FBS) versus serum-deplete (0.1% FBS) conditions. In serum-deplete condition, NAC inhibited LPS-induced JNK activation by ~ 75% (compare lanes 5 and 6, top panel). Similar to that observed with p38mapk, there was no inhibition and in fact an augmentation of LPS-induced JNK activation by NAC in serum-replete conditions (compare lanes 2 and 3, top panel). Immunoblot with p46 JNK1 antibody revealed that, as expected, the amount of p46 JNK bound to the c-Jun substrate paralleled the activity of JNK (bottom panel). Data are representative of three independent experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

These results demonstrate that profoundly different conclusions can be reached with regard to signaling pathways according to how purported antioxidants are used. Activation of the JNK, ERK, and p38^mapk pathways were each potently affected by oxidation-reduction (redox) balance in the extracellular milieu. The oxidizing or reducing conditions prevailing in the extracellular environment can be drastically altered according to the presence or absence of serum. Cell culture media by themselves are oxidizing in that they promote thiol auto-oxidation and associated peroxide generation within cells (19). Addition of serum, which contains both GSH and albumin, or albumin alone change this to a reducing environment (19). Serum can also provide additional proteins which may bind redox-active iron and make it unavailable for various free-radical reactions.

NAC has often been mentioned as an "antioxidant" in the recent literature regarding signal transduction (11, 15, 20). However, an understanding of thiol chemistry is required to use it as such. Thiols have the capacity to behave either as anti- or pro-oxidants depending on the local redox environment. The fact that thiols have the capacity to "auto-oxidize" has been known for some time (6). The presence of metals such as iron and the presence of ROS such as O₂^. or H₂O₂ potentiate this process (see the following reaction pathway) (6, 19). Having undergone auto-oxidation, thiols no longer act as "antioxidants." In addition, once initiated, these reactions can produce additional ROS including O₂^. , H₂O₂, and ^.OH. Here, the pathway for auto-oxidation of NAC is shown as adapted from Misra (6), where Me denotes a transition metal:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

The fact that these reactions can profoundly influence signaling pathways such as nuclear factor (NF)-B and activator protein (AP)-1 has been demonstrated previously by us (19). The current finding that these same thiol-dependent reactions can also cause important changes in signaling pathways in which phosphorylation reactions are required for activity highlights the need for continued attention to the local redox environment of cultured cells when designing experiments in this field. It could be argued that the presence of LPS in these experiments could have exaggerated the findings because some investigators have shown that LPS itself can promote oxidative stress (24). However, a similar divergence in findings in the absence and presence of serum has also been observed with other signaling pathways (AP-1, NF-B) and with other inducers of signaling, such as interleukin-1, which have not been shown to reproducibly cause oxidative stress (7).

Previously, we demonstrated that addition of 10% serum markedly decreases background levels of peroxides within cells as demonstrated by 2',7'-dichlorofluorescein fluorescence (19). The presence of this level of serum in experiments evaluating NF-B activation, AP-1 activation, and manganese superoxide dismutase (SOD) messenger RNA (mRNA) induction caused findings opposite from the results noted in the absence of serum. Thereby, agents such as NAC, DTT, and 2-mercaptoethanol were able to act as reductants in the presence of serum, but instead as oxidants in its absence. Antioxidants such as SOD and catalase, or chelators such as ethylenediaminetetraacetic acid, also decreased auto-oxidation and allowed thiols to act as reductants or "antioxidants" in these signaling experiments (19). The present experiments demonstrate that the presence of serum causes sufficient changes in the oxidation-reduction status of media to result in diametrically opposite findings regarding the effect of NAC on JNK, ERK, and p38^mapk signaling. Among the antioxidants in serum, albumin is the most prevalent, with each molecule bearing a single reactive sulfhydryl group. The presence of albumin has been known for many years to greatly prolong the useful life of isolated-perfused organs (25), and it is also capable of markedly decreasing H₂O₂-mediated damage to such organs (26). However, most available preparations of albumin do not provide a single purified antioxidant molecule. Indeed, some contaminants of purified albumin can also contribute to its antioxidant properties (27, 28). In cell signaling studies, omission of albumin, serum, or other potential redox buffers would likely lead to a pro-oxidant effect of NAC and potential alteration of the signaling pathway being studied. In some circumstances this could then cause misinterpretation of the findings, with attribution of the signaling event inhibited as being due to "oxidants." In addition, serum withdrawal itself can be a stress in many cultured cell types, and the oxidative stress induced by introduction of high concentrations of NAC under such conditions would be expected to further exaggerate such stress. We speculate that where the signal is mediated by oxidative stress, such as the activation of ERK 1/2 or other MAPKs by H₂O₂, such an enhancement of oxidative stress also might push concentrations of oxidants beyond optima for induction of signaling. In light of this, Abe and coworkers (29) found that JNK activation in bovine tracheal smooth-muscle cells was optimally increased at a concentration of 200 µM H₂O₂, with decreased activation at 500 µM H₂O₂.

Recent findings suggest that the effects of the presence of serum are due to intracellular events. Cantin and coworkers (30) very recently reported that incubation of pulmonary epithelial-like A549 cells, fibroblasts, or lymphocytes with human serum albumin (HSA) caused a 3-fold elevation of cellular GSH content. It was found that this elevation of cellular GSH was sufficient to protect cells against H₂O₂ toxicity and TNF-mediated biologic effects. Serum that was depleted of HSA had no effect on cellular GSH. Oxidation of the single free thiol (cysteine 34) of HSA did not impair its ability to augment cell GSH. On the other hand, reducing the disulfide bonds in HSA, such as through the addition of DTT, did inhibit the HSA-mediated increases in cell GSH. Incorporation of -glutamylcysteine groups, of which there are six in albumin, could be one mechanism whereby HSA causes an increase in cell GSH. However, the effect may be more complex than this. For example, there was induction of -glutamylcysteine synthase mRNA, which encodes the rate-limiting enzyme in GSH synthesis, but not that for glyceraldehyde-3-phosphate dehydrogenase, in cells treated with HSA.

These effects of HSA were specific. Other abundant serum proteins, such as -1 antitrypsin, had no effect. Overall, these findings confirmed that the intracellular oxidant- antioxidant balance is drastically altered by serum, and that the effect of serum is due to the presence of albumin. In addition, they showed that elevation of cell GSH may be the key event causing this effect (30). Whether the effects of NAC observed in the present study were due to intra- or extracellular events, or both, is not entirely clear. Given our previous findings that NAC can dramatically increase peroxides within cells in the absence of serum relative to those levels present in the presence of NAC and serum, coupled with the findings described earlier whereby serum can augment cell GSH via an albumin-mediated effect, we believe that the divergent effects of NAC observed in the presence and absence of serum are most likely due to intracellular effects on oxidant-antioxidant balance. Precisely how such a change might affect MAPK-dependent signaling is not known. ERK is activated by H₂O₂ in a variety of cell types, including macrophages (31, 32). In our studies, cells were preincubated with NAC, both in the presence and absence of serum, for 2 h before LPS exposure. Most investigators using NAC preincubate cells with the compound for 30 to 60 min or longer. This is more than sufficient time for considerable auto-oxidation of NAC in the absence of serum and secondary H₂O₂ production (28). We speculate that such prior exposure to H₂O₂ might attenuate ERK activation upon subsequent exposure to LPS in a manner similar to that caused by repeated exposure to endotoxin or phorbol ester (33, 34). These studies emphasize the need for caution in designing and interpreting signal transduction studies using thiol-containing antioxidants such as NAC. In addition, interpretation of others' results requires that experimental conditions, including the presence or absence of serum, be precisely defined.

NAC is used in a variety of clinical settings. In acetaminophen poisoning, a common disorder in which the pathogenesis is considered to be due, in part, to free-radical organ damage, it is a well accepted antidote. In that setting, the drug would encounter macrophage-like Kupffer cells in a serum-containing milieu. NAC also has been administered by aerosol to treat various pulmonary diseases such as chronic obstructive pulmonary disease and other obstructive airway disorders. In this situation, airway cells and alveolar macrophages would be exposed to NAC in a non-serum containing milieu. Although epithelial lining fluid normally contains some low molecular weight antioxidants, some of these, such as GSH, may be diminished in these disorders. This could result in a net pro-oxidant effect. In other lung diseases, such as adult respiratory distress syndrome, in which endotoxin, cytokines, and oxidative stress have been implicated in pathophysiology, no therapeutic benefit of systemic NAC therapy was found (35). In that situation, we speculate that a milieu of elevated iron and/or oxidant levels might have favored oxidation of the drug and caused there to be no net benefit. Recently, in a prospective, randomized, placebo-controlled study of critically ill patients with multiorgan failure, NAC showed no benefit and in fact was noted to be harmful if given 24 h after admission to the intensive care unit (8).

The physiologic significance of these findings is a reasonable area for speculation. In health, inflammation is normally initiated in a reducing environment and, therein, positive signaling through the MAPKs would be favored. Because various oxidants such as O₂^. , H₂O₂, nitric oxide, hypochlorous acid, ^.OH, peroxynitrite, and others are generated during ongoing inflammation, activation of MAPK-dependent signaling followed by inhibition of these responses due to the persistent oxidizing environment might be one means of first initiating inflammation, and then reversing these responses and downregulating inflammation. On the other hand, prior exposure to oxidative stress could impair necessary adaptive signaling responses to endotoxin, and potentially cause adverse clinical results.