INTRODUCTION

TOP ABSTRACT INTRODUCTION DISCUSSION REFERENCES

The initial step in leukotriene (LT) biosynthesis is the activation of phospholipase A₂ (PLA₂), which then hydrolyzes membrane phospholipids to release arachidonic acid (AA). In a calcium- and ATP-dependent reaction, AA is metabolized by 5-lipoxygenase (5-LO) in a two-step process to yield the epoxide intermediate, LTA₄ (1). This step is dependent upon the interaction of the 5-LO with a nuclear membrane protein, termed 5-lipoxygenase activating protein (FLAP) (2). LTA₄ is then converted into either LTB₄ by an enzymatic reaction mediated by LTA₄ hydrolase or conjugated with glutathione by LTC₄ synthase to form the cysteinyl LTs (cysLTs).

	HISTORICAL PERSPECTIVE

Until recently, we knew that specific stimuli triggered cells to release LTs extracellularly, but we knew little about the intracellular location of the LT enzymatic machinery. The cell could truly be described as a black box. In the last five years, however, three key proteins in the initial steps of LT biosynthesiscytosolic PLA₂, 5-LO, and FLAPwere found to be distributed into different cellular compartments. Following cell fractionation into soluble and particulate fractions, FLAP was localized in the particulate fraction from both resting and activated cells (3). In contrast, both 5-LO (4) and PLA₂ (5) were found in the soluble fraction of resting cells, but in response to cell stimulation and increased intracellular calcium levels, both enzymes left the soluble compartment and moved into the particulate fraction.

This seminal observation led to the hypothesis (circa 1993) that in response to increased intracellular calcium levels, 5-LO translocates to the plasma membrane, where it interacts with FLAP, which was believed to be a docking protein for 5-LO (3) (Figure 1A). This complex was then somehow able to use AA, presumably liberated from plasma membrane phospholipids by PLA₂, to form the LTs, which would then be secreted extracellularly. At that time, it made perfect sense to speculate that these reactions occurred at the plasma membrane, because LTs are largely secreted extracellularly, where they are believed to act on other target cells that have LT receptors on their cell surface. While this model made perfect sense at the time, it was incorrect.

View larger version (13K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Figure 1.   Intracellular site of leukotriene synthesisthe early model. (A) In response to increased intracellular calcium levels following activation, 5-LO was believed to translocate to the plasma membrane, where it could interact with FLAP. Nuclear envelope as the site of leukotriene synthesisthe current model. (B) Following cell activation, cPLA2 translocates from the cytoplasm to the nuclear envelope. Both nuclear and cytoplasmic 5-LO also translocate to the nuclear envelope. cPLA2 hydrolyzes nuclear membrane phospholipids to release arachidonate, which binds to FLAP for presentation to 5-LO for oxygenation.

	NUCLEAR ENVELOPE AS SITE FOR ARACHIDONIC ACID RELEASE AND 5-LO METABOLISM

Recently, work from a number of groups has led to the surprising conclusion that these enzymatic reactions occur instead at the nuclear membrane. First, it was observed that FLAP is located predominantly at the nuclear envelope and not at the plasma membrane as initially speculated (6, 7). Second, during activation, cytosolic PLA₂ (cPLA₂) translocates primarily to the nuclear envelope, and not to the plasma membrane (6, 8). Third, the cPLA₂ that translocates to the nuclear envelope selectively hydrolyzes phospholipids from the nuclear envelope, and not from other cellular phospholipids (10). This latter observation demonstrated the functional significance of the translocated cPLA₂. Finally, FLAP is no longer considered a docking protein for 5-LO (11); rather, it is an arachidonate transfer protein (12). Therefore, the current concept is that cPLA₂ hydrolyzes nuclear envelope phospholipids to release arachidonate (Figure 1B). The free arachidonate binds to FLAP and is somehow presented to the 5-LO, enabling the oxygenation reaction to occur. 5-LO also has been shown to move predominantly to the nuclear envelope (6, 7), where it is localized in close proximity to FLAP and cPLA₂. However, 5-LO does not appear to bind to FLAP during these reactions (11). Furthermore, LTC₄ synthase also has been found at the nuclear envelope, suggesting that LTC₄ also is synthesized from 5-LO-derived LTA₄ at this site (13).

The significance of this new model (Figure 1B) is that LT synthesis is not initiated at the plasma membrane as originally believed; rather, it begins at the nuclear membrane. This raises an important question: If LTs are destined to be exported from cells, why have the enzymes responsible for their synthesis clustered at the nuclear envelope? The answer to this question is unknown at present. However, this model demonstrates that LTs are not only exported but also are found deep within the cell, where they potentially enter the nucleus.

	5-LIPOXYGENASE

The nuclear envelope appears to be the universal destination for 5-LO translocation in activated cells. However, the site where 5-LO resides in resting cells has been the object of intense investigation and is turning out to be a very complex story. In addition to the cytosolic pool of 5-LO, recent results indicate that there is actually a pool of 5-LO within the nucleus of resting cells (14).

Localization of 5-Lipoxygenase in Resting Cells

When human peripheral blood neutrophils are stained with an anti-5-LO antibody and then a fluorescent secondary antibody, the fluorescence is found diffusely throughout the cytoplasm. This suggests that 5-LO is localized to the cytoplasm in these resting peripheral blood neutrophils. Similarly, neutrophils can be gently disrupted in order to spare the nucleus, and a nuclear fraction isolated by low-speed centrifugation. The supernatant of this preparation can additionally be fractionated into cytosolic and crude non-nuclear membrane fractions. Analysis of these three fractions again demonstrates that 5-LO is predominantly cytosolic in these resting cells (14). A similar localization, predominantly in the cytosol, is found in other cell types at rest, including peripheral blood monocytes (11) and peritoneal macrophages (6).

However, this is not the story in all cells. Using rat basophilic leukemia (RBL) cells, a model for mast cells, immunofluorescence microscopy demonstrated faint 5-LO staining in the cytoplasm, with more intense staining in the nucleus (14). The staining pattern for 5-LO in RBL cells is similar to that of the histones, which are intranuclear proteins. Interestingly, during cell division, the nuclear membrane breaks down and the 5-LO leaks into the cytoplasm (14). A similar, nuclear-predominant localization for this enzyme has also been reported for murine bone marrow-derived mast cells (15).

Also of relevance to asthma and other lung diseases is the alveolar macrophage. Its pattern of 5-LO distribution is similar to that of RBL cellsfaint cytoplasmic staining, with intense nuclear staining (11) (Figure 2A). Following fractionation of resting alveolar macrophages, 5-LO can be localized in both a cytoplasmic pool and a soluble pool from the nucleus (Figure 2B). The intranuclear localization of 5-LO was confirmed using immunogold electron microscopy. Interestingly, 5-LO was not randomly distributed throughout the nucleus (Figure 2C). Rather, it was localized to the euchromatin region (11), the intranuclear site where active gene transcription occurs.

View larger version (128K): [in this window] [in a new window]   View larger version (72K): [in this window] [in a new window]   Figure 2.   Localization of 5-lipoxygenase in resting human alveolar macrophages. Immunofluoresence staining for 5-LO demonstrates faint cytoplasmic staining and intense nuclear staining (A). Immunoblot analysis reveals a band of 5-LO protein in both cytosolic (lane C ) and soluble nuclear (lane Ns) fractions, but not in membrane (lane M) or particulate nuclear (lane Np) fractions (B). Using immunogold electron microscopy, nuclear 5-LO was localized to the euchromatin region of the nucleus (C ). Abbreviations: nuc = nucleus; ne = nuclear envelope; er = endoplasmic reticulum; pm = plasma membrane.

Translocation of Nuclear 5-Lipoxygenase to the Nuclear Envelope

The activity of nuclear 5-LO and its role in LT synthesis was next investigated by first isolating nuclei, then performing a cell-free 5-LO assay. The results demonstrated that nuclear 5-LO indeed had enzymatic activity (16). The second step was to ascertain whether nuclear 5-LO translocated to the nuclear envelope following cell activation. When RBL cells were activated with calcium ionophore, immunofluorescence microscopy showed that 5-LO was depleted from its diffuse localization within the nucleus, and that the fluorescence now localized as a rim around the nuclear envelope (Figure 3). Furthermore, cytoplasmic 5-LO staining likewise decreased. Such results suggested that both nuclear and cytoplasmic 5-LO translocated to the nuclear envelope. The translocation of nuclear 5-LO to the nuclear envelope has been subsequently confirmed using two cell types important to the lungalveolar macrophages (11, 16) and cultured primary mast cells (15).

View larger version (91K): [in this window] [in a new window]   Figure 3.   Translocation of nuclear 5-lipoxygenase to nuclear envelope. Resting RBL cells (A) exhibit diffuse cytoplasmic and intranuclear staining. In cells that were activated with calcium ionophore (B), cytoplasmic and intranuclear staining were both reduced relative to non-activated cells, and 5-LO staining was localized as a rim around the nuclear envelope.

Localization of 5-Lipoxygenase Using In Situ Preparations

The results described thus far involve cell isolation and processing, and it was desirable to verify them using in situ preparations not subjected to such perturbations. Alveolar macrophages were therefore evaluated in situ by immunohistochemistry in normal-appearing regions of lung tissue obtained from patients undergoing lung resection for bronchogenic carcinoma (17). The in situ pattern was identical to the intranuclear staining pattern in alveolar macrophages removed from the lung and evaluated ex vivo. Again, there was some staining in the cytoplasm but intense staining within the nucleus.

The translocation of 5-LO to the nuclear envelope was studied in situ using lung tissue from patients with idiopathic pulmonary fibrosis (IPF), a disease in which there is overproduction of LTs (17). In IPF, inflammation and fibrosis occur in peripheral lung parenchyma, in the air spaces, and in alveolar structures, but not in the larger airways. Some alveolar macrophages exhibited an immunohistochemical staining pattern consistent with the resting state, in which 5-LO was distributed principally throughout the nucleus (17) (Figure 4). Other cells, however, exhibited a concentration of staining around the rim of the nucleus, a pattern that was confirmed by confocal microscopy, and which suggested cell activation and 5-LO translocation to the nuclear envelope in situ.

View larger version (128K): [in this window] [in a new window]   Figure 4.   Localization of 5-lipoxygenase in alveolar macrophages in situ by immunohistochemical staining. In histologically normal lung tissue excised from patients undergoing resection for lung cancer (left panel ), 5-LO was localized throughout the nucleus of alveolar macrophages (open arrow). By contrast, in lung tissue from patients with idiopathic pulmonary fibrosis (IPF) (right panel ), 5-LO was localized to the nuclear membrane of some alveolar macrophages (solid arrow), whereas in other cells a "resting" pattern was observed. Adapted from Reference 17 by permission.

Potential Significance of Nuclear 5-Lipoxygenase Localization

In rat blood neutrophils, like their human counterparts, most 5-LO staining is found in the cytoplasm. However, it is neutrophils recruited to sites of inflammation that are of greater interest. Therefore, we examined neutrophils recruited to the peritoneal cavity following glycogen instillation (18). In contrast to blood neutrophils, these elicited cells have both a cytosolic and a soluble nuclear pool of 5-LO, suggesting movement of 5-LO into the nucleus (18) (Figure 5). This nuclear import occurs rapidly and may be due to endothelial adherence that is an early event in neutrophil extravasation from the bloodstream (18). Of note, nuclear import is not associated with LT synthesis. However, when peritoneal neutrophils were stimulated with calcium ionophore, they exhibited a loss of intranuclear 5-LO and the appearance of 5-LO staining on the nuclear envelope (18).

View larger version (95K): [in this window] [in a new window]   Figure 5.   Nuclear import of 5-lipoxygenase in elicited rat peritoneal neutrophils. Neutrophils isolated from rat blood (A) exhibited 5-LO in the cytosol, as demonstrated by immunofluorescence microscopy and cell fractionation/immunoblot analysis. By contrast, neutrophils isolated from the peritoneal cavity of rats injected with glycogen (B) exhibited 5-LO in both cytosolic and soluble nuclear pools. Adapted from Reference 18 by permission.

In order to assess the possible metabolic consequences of this nuclear import of 5-LO, blood and peritoneal neutrophils were stimulated with a range of concentrations of calcium ionophore, and LTs in the medium were quantitated. First, it was observed that elicited neutrophils with intranuclear 5-LO were less sensitive to calcium ionophore stimulation and required a higher dose to trigger LT synthesis (18). It is tempting to suggest that sequestration of 5-LO in the nucleus may be one way for cells to protect themselves from indiscriminate stimulation. Second, although the elicited neutrophils had a higher threshold to stimulation, they also had a 5-fold higher maximum capacity for LT synthesis. This LT synthetic profile would therefore be consistent with what might be expected for cells involved in an inflammatory response. It is not possible to be certain that these changes in LT synthesis are the specific consequences of a shift of 5-LO from the cytosol to the nucleus. However, similar differences in sensitivity and maximal capacity for LT synthesis are observed when two other cell types whose 5-LO localization differs are examined, i.e., alveolar and peritoneal macrophages from rats (19). The fact that LT synthesis is initiated at the nuclear envelope almost certainly means that significant levels of LTs are found in the nucleus or deep within the cell. Here, they could have significant autocrine implications, regulating processes such as gene transcription or signal transduction. It is unlikely that such effects, if they occur, result from activation of "traditional" LT receptors on the plasma membrane. Recently, in fact, an intranuclear receptor for LTB₄ has been identified, and it is a transcription factor of the steroid hormone superfamily (20). It is possible, therefore, that plasma membrane receptors are primarily responsible for mediating paracrine effects of LTs, while intranuclear receptors mediate some of the autocrine effects of these substances.

It is clear that cysLT receptor antagonists and 5-LO inhibitors have clinical efficacy. Zileuton, for example, is equieffective at inhibiting LT synthesis in vitro in alveolar macrophages and blood neutrophils, even though these cells have different 5-LO distributions. Therefore, the distribution of 5-LO does not offer obvious directions for designing more specific 5-LO inhibitors, such as agents that may selectively inhibit 5-LO at key subcellular locations. However, it does emphasize the need for a better understanding of cell topographyhow the various lipids formed during LT synthesis traffic within the cell, how they exit cells, how different lipid pools are topographically defined, how effector molecules interact with these lipids, and finally, how the drugs of interest actually partition within cells. Because shifts in compartmentalization of 5-LO do influence cellular metabolic capacity, modulation of 5-LO localization in resting or activated cells represents an additional strategy by which pharmacologic agents could interfere with the 5-LO pathway.

	DISCUSSION

TOP ABSTRACT INTRODUCTION DISCUSSION REFERENCES

Busse: Do you have any data to indicate that there may be priming effects from some of these leukotrienes on either neutrophil or macrophage function? If there are, is tyrosine being phosphorylated, or are there some transcription factors that might be generated? The leukotrienes clearly have inflammatory effects rather than just agonist effects on airway smooth muscle. I think these would be more intriguing and would support a more central role for LTs in the pathogenesis of asthma.

Peters-Golden: There is a large body of work that has already been done in that area. For example, the leukotrienes directly activate a number of kinases and transcription factors.

Aharony: What about the proto-oncogenes c-fos and c-jun?

Peters-Golden: Yes, they activate c-fos and c-jun. The leukotrienes cause transcription of oncogenes; they also cause transcription of various cytokines and growth factors. Thus, the actions of the LTs are clearly involved in the fundamental cellular processes of differentiation and proliferation, going well beyond mere smooth muscle contraction. Moreover, there is also some evidence that, in addition to LTs produced by 5-LO, oxygen radicals generated during oxygenation can activate transcription factors such as nuclear factor kappa B.

Bernstein: How much is known about the way leukotrienes induce these effects?

Peters-Golden: Well, it's interesting. Most of the experiments were performed by adding leukotrienes to intact cells. Therefore, it isn't clear whether they are working through a plasma membrane receptor or via some other mechanism. The recent finding of an intranuclear LT receptor offers exciting new possibilities.

Bernstein: Since LTC₄ is actively exported from several cell types and the conversion of LTD₄ and LTE₄ occurs extracellularly, is there an impact on your hypotheses about the potential autocrine and paracrine roles of the leukotrienes and the intracellular nature of their synthesis?

Peters-Golden: Despite the fact that LTC₄ is exported, I think that LTC₄, LTB₄, or their precursors could still act intracellularly. If all you ever measure is the amount of leukotriene secreted extracellularly, you get spoiled. You get used to very high levels of leukotriene production. There is also some work to suggest that leukotrienes may have important effects in cells that produce much lower levels, such as parenchymal cells. For example, epithelial cells produce very low levels of cysLTs, levels that in the past, most of us would have ignored.

These very low concentrations of cysLTs participate in epithelial cell proliferation and work by affecting intracellular calcium regulation. Again, the absolute concentrations are very low. If you give either 5-LO inhibitors or cysLT receptor antagonists to these epithelial cells, you can inhibit proliferation and calcium flux. In the future, I think that you are going to see people measure local eicosanoid concentrations within the nucleus.

Busse: Is there any evidence in isolated nuclear preparations of a direct nuclear regulatory effect of the leukotrienes?

Peters-Golden: Yes, with the recent recognition of an intranuclear LT receptor, LTB₄ caused transcription in isolated nuclei.

Cohn: Is it possible that there is an inducible and a constitutive form of 5-LO?

Peters-Golden: Well, 5-lipoxygenase is itself inducible, and it can be regulated. There are a number of factorscytokines, colony stimulating factorsthat can upregulate 5-LO production. FLAP is also regulated under certain conditions. But as far as we know, there is only a single 5-LO gene, with no isoforms.

Drazen: I think it is important to understand that there has been a good deal of controversy about the role of cytosolic versus secretory PLA₂ as an arachidonate source. I think that both forms generate arachidonate. The secretory form probably acts on the external plasma membrane, and may be equivalent to adding exogenous arachidonic acid that diffuses into the cell through the membrane. I think that there is reasonable evidence to implicate both the cytosolic and secretory enzymes, although one may have more of a role in inducing inflammation and the other may have more of a cell-regulatory role.

Peters-Golden: The issue of different phospholipases adds another layer of complexity. I think the fact remains that eicosanoid-forming enzymes remain near the nucleus. What does that imply for a phospholipase that works at the plasma membrane? One possibility that intrigues me is that phospholipases acting at the plasma membrane may have other functionsgenerating arachidonate, not for eicosanoid synthesis, but for other functions.