Brain trauma leads to enhanced lung inflammation ... - Semantic Scholar

Journal of Cerebral Blood Flow & Metabolism (2006), 1–12 & 2006 ISCBFM All rights reserved 0271-678X/06 $30.00 www.jcbfm.com

Brain trauma leads to enhanced lung inflammation and injury: evidence for role of P4504Fs in resolution Auinash Kalsotra1, Jing Zhao2, Sayeepriyadarshini Anakk1, Pramod K Dash2 and Henry W Strobel1 1

Department of Biochemistry and Molecular Biology, The Vivian L Smith Center for Neurologic Research, The University of Texas, Houston Medical School, Houston, Texas, USA; 2Department of Neurobiology and Anatomy, The Vivian L Smith Center for Neurologic Research, The University of Texas, Houston Medical School, Houston, Texas, USA

Traumatic brain injury is known to cause several secondary effects, which lead to multiple organ dysfunction syndrome. An acute systemic inflammatory response seems to play an integral role in the development of such complications providing the potential for massive secondary injury. We show that a contusion injury to the rat brain causes large migration of inflammatory cells (especially macrophages and neutrophils) in the major airways and alveolar spaces at 24 h post-injury, which is associated with enhanced pulmonary leukotriene B4 (LTB4) production within the lung. However, by 2 weeks after injury, a temporal switch occurs and the resolution of inflammation is underway. We provide evidence that 5-lipoxygenase and Cytochrome P450 4Fs (CYP4Fs), the respective enzymes responsible for LTB4 synthesis and breakdown, play crucial roles in setting the cellular concentration of LTB4. Activation of LTB4 breakdown via induction of CYP4Fs, predominantly in the lung tissue, serves as an endogenous signal to ameliorate further secondary damage. In addition, we show that CYP4Fs are localized primarily in the airways and pulmonary endothelium. Given the fact that adherence to the microvascular endothelium is an initial step in neutrophil diapedesis, the temporally regulated LTB4 clearance in the endothelium presents a novel focus for treatment of pulmonary inflammation after injury. Journal of Cerebral Blood Flow & Metabolism advance online publication, 20 September 2006; doi:10.1038/sj.jcbfm.9600396 Keywords: blood brain barrier; cytochrome P450 4Fs; cytokines; leukotriene B4 metabolism; pulmonary inflammation

Introduction Traumatic brain injury (TBI), especially closed head injury resulting from hemorrhage, blunt trauma, or traffic accidents, has become an increasing problem in medical care (Marion, 1999). Apart from mechanical disruption of brain tissue and vasculature, the extracellular and intracellular changes that follow have been termed ‘secondary injury’. Secondary Correspondence: Dr HW Strobel, Department of Biochemistry and Molecular Biology, The University of Texas-Houston Medical School, PO Box 20708, 6431 Fannin Street, Houston, TX 77225, USA. E-mail: [email protected] The work is supported by Grants R01 MH070054 and RO1 NS044174 from NIH and US Department of the Army Grant T5 0004268 for Texas Training and Technology against Trauma and Terrorism to Henry W Strobel, NS35457 to Pramod K Dash and President’s Research Scholarship awarded to Auinash Kalsotra. Received 13 April 2006; revised 27 July 2006; accepted 30 July 2006

injury is manifested by excitotoxicity, changes in concentrations of ions, specific peptides/proteins and neurotransmitters, etc. These secondary events occur within hours or days after primary injury, and can lead to further damage of the nervous system (Faden et al, 1989). Unfortunately, trauma to the brain, besides causing damage to the primary site of insult, leads to multiple organ dysfunction syndrome (Campbell et al, 2003; Gruber et al, 1999). This results in a myriad of pathophysiologic complications affecting neurologic, psychological, cardiovascular, pulmonary, vascular, and metabolic functions, which in turn impede the recovery process (Clifton et al, 1983; Fisher et al, 1999; Kalsotra et al, 2003b; McIntosh et al, 1999). Although many features of multiple organ dysfunction syndrome may be driven by low-grade systemic infection commonly associated with acute brain injury (Woiciechowsky et al, 1998, 2002), it is possible that injury to the brain per se drives this generalized response.

Role of CYP4F mediated LTB4 metabolism after brain trauma A Kalsotra et al 2

More recently, the association between brain injury and subsequent pulmonary dysfunction has been recognized clinically. Indeed, development of acute lung injury is associated with the worst neurologic outcome in patients with severe brain trauma or hemorrhage (Holland et al, 2003; Mascia and Andrews 1998; Paine et al, 1952). While the episodes of hypoxia after head trauma do not correlate with lung injury (Holland et al, 2003), several pieces of evidence suggest it may occur as a result of a systemic inflammatory response (De Simoni et al, 1990; Stubbe et al, 2004; Woiciechowsky et al, 1998, 2002). Fisher et al (1999) have reported enhanced pulmonary inflammation in organ donors after fatal brain injury. This severe inflammatory response stems from increased cytokine production and infiltration of activated neutrophils into the lung, which provides the potential for massive localized tissue injury. We hypothesize leukotriene B4 (LTB4) involvement in aggravating such inflammatory pathways because LTB4 functions as a potent neutrophil chemoattractant after injury or infection (Ford-Hutchinson et al, 1980). Leukotriene B4, originally identified as an activator of granulocytes (Samuelsson et al, 1987), initiates and amplifies chemotaxis, cell stimulation and release of granule products/superoxide anions (Ford-Hutchinson et al, 1980) by a cell-surface G-protein coupled receptor (Yokomizo et al, 1997). Although 5-lipoxygenase (5-LOX), an important enzyme for LTB4 biosynthesis, was purified as a cytosolic protein, the enzyme is known to translocate to the nuclear envelope (Brock et al, 1995). 5-Lipoxygenase catalyzes the conversion of arachidonic acid to leukotriene A4. Hydrolytic attack of leukotriene A4 by leukotriene A4 hydrolase in the cytoplasm, and potentially in the nucleus, yields LTB4 (Haeggstrom 2004). Leukotriene B4 inactivation is catalyzed by the Cytochrome P450 4F (CYP4F) subfamily implicating their immediate role in modulation of LTB4 levels and, therefore, function in resolving inflammation (Kikuta et al, 2002). In this report, we evaluated subsequent to contusion brain injury, LTB4 associated lung inflammatory effects in a rat TBI model. This model closely simulates contusion head injury in humans who have been victims of traffic accidents or subjected to blunt trauma (Kalsotra et al, 2003b). We also sought the mechanism that might resolve pulmonary inflammation after brain injury. Finally, we examined whether blood–brain barrier (BBB) compromise and systemic release of proinflammatory cytokines function as a link between brain and lung contributing to the pathogenesis of experimental brain injury.

Materials and methods Materials One hundred and two Long Evans rats of either sex (280 to 330 g body wt) were purchased from Harlan Laboratories Journal of Cerebral Blood Flow & Metabolism (2006), 1–12

(Indianapolis, IN, USA). Purified LTB4, 20-OH LTB4, and LTB4 EIA kit were purchased from Cayman Chemical Company (Ann Arbor, MI, USA). 17-Octadecynoic acid was obtained from Tocrist (Ellisville, MO, USA). Cytochrome P450 4F standard amplicons, primers, and probes were custom synthesized by IDT DNA Technology Inc. (Coralville, IA, USA). 5-Lipoxygenase antibody was purchased from Research Diagnostics Inc. (Flanders, NJ, USA). Neutrophil elastase antibody was obtained from EMD Biosciences (San Diego, CA, USA). Cytochrome P450 4F2, CYP4F5, cyclooxygenase-2 (COX-2) and von Willebrand factor (VWF) antibodies were respective gifts from Dr Yasushi Kikuta, Department of Allied Biological Science Fukuyama University Hiroshima Japan; Dr Hidenori Kawashima, Osaka City University Graduate School of Medicine Japan. All reagents utilized in the experiments reported here where not specifically defined, were of reagent grade quality or higher.

Experimental Brain Injury Model A controlled cortical impact injury model as previously described (Long et al, 1996) was used to cause TBI in the animals. All protocols were conducted in compliance with guidelines set forth in the NIH Guide for the Care and use of Laboratory Animals and approved by The University of Texas Health Science Center’s Institutional Animal Care and Use Committee. Briefly, rats were initially anesthetized with 5% isoflurane with a 1:1 O2:N2O mixture. Animals were mounted on a stereotaxic frame and were secured by two ear bars and an incisor bar. Anesthesia was maintained with 2.5% isoflurane with a 1:1 O2:N2O mixture. Bilateral craniotomies were performed midway between the bregma and the lambda with the edges of craniotomies 1 mm lateral to the midline, one at the site of impact, the other contralateral to the impact site. Rats received a single, unilateral impact at 6 m/secs, 2.0 mm deformation. Previous studies have shown that 2.0 mm deformation does not cause widespread hippocampal cell loss, a defining feature of moderate brain injury (Long et al, 1996). Two controls were employed. One is the contralateral hemisphere. The other is the brain of sham-operated but uninjured rats. Core body temperature was monitored continuously by a rectal thermistor probe and maintained at 371C to 381C. After injury, the scalp was sutured and closed and the animal extubated to minimize anesthetic effects and to allow rapid recovery. The impact apparatus for controlled cortical impact injury consists of a small (1.975 cm)-bore, double-acting strokeconstrained, pneumatic cylinder with a 5.0 cm stroke. The cylinder is rigidly mounted perpendicular to the cortex on a cross bar. The lower rod end has an impactor tip attached (i.e., that part of the shaft that actually contacts the exposed dura mater). The upper rod end is attached to the transducer core of a linear variable differential transformer. The velocity of the impacter shaft is controlled by gas pressure. Impact velocity is directly measured by the linear variable differential transformer (Shaevitz model 500 HR), which produces an analog signal that is recorded by a PC based data acquisition (RC


Electronics) for analysis of time and displacement parameters of the impactor. Six animals in each group were killed at different time points postinjury. The molecular, histopathologic and immunohistochemical analyses were performed in different sets of animals.

a chromagenic solution consisting of diaminobenzidine supplemented with 1% nickel chloride for signal enhancement (Vector labs), after which the sections were counterstained with nuclear red solution and mounted.

Membrane and Soluble Protein Preparations Bronchial Alveolar Lavages At specific time points, postsurgery (3 h, 24 h, 3days, or 2 weeks) rats were anaesthetized with sodium pentobarbitone orally (Sleepaway, Fort. Dodge Laboratories Inc., Fort Dodge, IA, USA). A midline thorax to neck incision was made, the trachea was exposed, and a blunt end 18-gauge needle was inserted and tied into the trachea. Lungs were lavaged three times with 2 mL phosphate-buffered saline (PBS), and 4.5 to 5 mL of pooled lavage fluid was recovered. Samples were centrifuged at 2500 r.p.m. for 5 mins to recover cells. Both cells and supernatant fractions from these spins were collected and stored at 801C for the LTB4 analysis. Bronchio-alveolar lavage fluid (BALF) cells were resuspended in 200 mL PBS. Total cell counts were determined from an aliquot counted using a hemocytometer, while another aliquot, cytospun onto microscope slides, was stained with Diff-Quik (Dade Behring, Newark, DE, USA) for cellular differentials. Two hundred cells per sample were identified and counted under oil immersion.

Protein Concentration in the Bronchio-Alveolar Lavage Fluid Lung permeability was determined by measuring the total protein concentration in the cell-free BALF with bichinchoninic acid procedure using albumin as a standard.

Histologic Preparation After the BALF was collected, the animal’s chest was opened and the ribs were removed to allow uncompromised lung inflation. Rats were intracardially perfused with 50 mL PBS followed by 100 mL freshly prepared 4% paraformaldehyde to drain all blood. The trachea was cannulated and the lungs were inflated with 4% paraformaldehyde (8 mL) by means of a 10 cm3 syringe. The inflated lungs were tied and then fixed in 4% paraformaldehyde for 24 h. At the conclusion of fixation, lung samples were rinsed in PBS, dehydrated, and embedded in paraffin. After embedding, 5 mm sections were cut, transferred onto microscope slides and stained with hematoxylin and eosin (Shandon-Lipshaw Inc., Pittsburgh, PA, USA), according to the manufacturer’s instructions. Neutrophil recruitment into lungs was assessed by immunohistochemistry, by overnight incubation with antineutrophil elastase antibody (1:100 dilutions) followed by three washes with PBS. Sections were then incubated at room temperature for 2 h with a biotinylated secondary antibody (1:500 dilutions). Endogenous peroxidase activities were quenched with a 0.6% hydrogen peroxide solution. Sites of antigen expression were visualized by incubating tissue sections for 5 mins with

Lungs collected from shams, 3 h, 24 h, 3 days, and 2 weeks postinjured rats were individually homogenized (20 strokes) in six volumes of potassium phosphate buffer (pH 7.4) containing 20 mmol/L Kpi, 0.25 mol/L sucrose, 1 mmol/L EDTA, and a cocktail of protease inhibitors (1 mmol/L phenylmethylsulfonylfluoride, 1 mg/mL leupeptin and 0.7 mg/mL of pepstatin). Each homogenate was centrifuged at 10,000 g for 20 mins and the supernatant fraction was collected (a portion of it was kept for cytokine and soluble 5-LOX determinations). The supernatant fraction was ultracentrifuged at 100,000 g for 45 mins, the pellet washed in fresh buffer, and again centrifuged at 100,000 g for 45 mins. The washed pellets were resuspended and stored in 801C until analyzed for CYP4F, COX-2 or membrane translocated 5-LOX. The protein concentrations were determined using the bicinchoninic acid procedure.

Immunoblotting Protein samples were boiled in Laemmli buffer and resolved on 4 to 15% gradient Tris-glycine sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels. Proteins were transferred onto nitrocellulose membranes using a semidry transfer apparatus. Membranes were blocked overnight and incubated with CYP4F5 (1:250 dilution), 5-LOX (1:500 dilution), and COX-2 (1:1000 dilution) antibodies. Membranes were then washed and incubated at room temperature with horse radish peroxidase-conjugated secondary antibody (1:1000 dilution) for 1 h. Immunoreactivity was detected using a horse radish peroxidase chemiluminescense system (Pierce Rockford, IL, USA).

Measurement of Leukotriene B4 by Enzyme Immunoassay Leukotriene B4 levels were determined in duplicate using a commercially available ACEt competitive enzyme immunoassay kit (Cayman Chemicals, Ann Arbor, MI, USA) as per the manufacturer’s recommended protocols. Assays were performed using 100 mL of recovered BALF or serum. The absorbance was read at recommended wavelength on Dynex ELISA reader within 30 mins of stopping the reactions.

Leukotriene B4 Metabolism Metabolism of LTB4 by lung microsomes was determined using a high-performance liquid chromatography method as described previously (Kalsotra et al, 2003a). Briefly, the reaction mixture containing 500 mg lung microsomal protein, 20 mmol/L HEPES buffer (pH 7.5) including 340 mmol/L sucrose, 1 mmol/L EDTA, and 1 mmol/L dithiothreitol (DTT), 1 mmol/L NADPH, and 60 mmol/L Journal of Cerebral Blood Flow & Metabolism (2006), 1–12


LTB4, in a total volume of 0.1 mL, was incubated at 371C for 20 mins. The reaction products were extracted with 3 mL of ethyl acetate, dried gently under N2 gas and resolublized in 100 mL of mobile phase. The products were passed over a C18 column and measured with an UV detector set at 270 nm. Prostaglandin B1 served as an extraction control. The quantities of newly formed metabolite were determined from the peak area read from the standard curve prepared for 20-OH LTB4 (0.3 to 5 mmol/L). Cytochrome P450 4F activity inhibition was tested by preincubating the microsomes with 10 mmol/L 17-octadecynoic acid for 10 mins at 371C.

RNA Preparations and Quantitative Real-Time Polymerase Chain Reaction A portion of BALF cells and the lungs from which BALF was collected were processed to obtain total RNA using RNA-STAT. All samples were DNAse treated using RQ1 DNAse (Promega, Madison, WI, USA). The quality of the isolated RNA was assessed by electrophoresis on 1% agarose gels based on the integrity of 28S and 18S bands after ethidium bromide staining. Cytochrome P450 4F gene expression analysis was performed using quantitative real-time polymerase chain reaction assays as previously described (Kalsotra et al, 2003a).

Immunohistochemistry Animals were intracardially perfused with ice-cold PBS as described earlier. The lungs were inflated with 8 mL TissueTek OCT embedding compound (Miles Inc., Elkhart, IN, USA) and slowly frozen in precooled isopentane. The fresh frozen tissues were sectioned into 10-mm-thick slices using a cryostat and carefully mounted on 2% gelatin subbed microscopic slides. The sections were dried at room temperature for 1 h and fixed with 100% methanol for 20 mins at 201C. Sections were rinsed in 1 PBS three times for 10 mins each and then blocked with 5% goat serum in PBS with 0.25% Triton X-100 for 1 h. For doublelabeling experiments, sections were incubated with primary antibodies CYP4F2 (cross reacts with CYP4F4, and 4F6) (1:500), and vWF (1:100 dilution) for 48 h at 41C. Cytochrome P450 4F and vWF primary antibodies were detected using species-specific secondary antibodies conjugated to Alexafluors. Sections were cover-slipped using Fluoromount-G (Fisher Scientific) and the immunofluorescence was visualized using either a UV microscope (Axiophot, Zeiss) with the appropriate filter sets or a BioRad confocal microscope. To assess nonspecific fluorescence, tissue sections were incubated as described above in the blocking buffer without primary antibodies and tested for immunoreactivity using the secondary antibodies. No signal was detected for any of the detection methodologies utilized in this study in the absence of primary antibodies.

argon laser for excitation at 488 and 568 nm. The confocal was attached to an Olympus BX-50WI upright microscope. Images were acquired using LaserSharp 2000 software, analyzed using MetaMorph 6.1 and adjusted for size using Illustrator 9.0 (Adobe).

Evans Blue Extravasations Blood–brain barrier permeability was assessed using extravasation of Evans blue dye. At the assigned time points, animals were anesthetized with 7% chloral hydrate (0.7 mL/100 g). Evans blue (3%) prepared in saline (4 mL/kg) was injected through the Jugular vein and allowed to circulate for 1.5 h. Animals were then transcardially perfused with PBS followed by 4% paraformaldehyde (400 mL). The brain was removed and the hemispheres separated and weighed. Brain sections (2 mm) were made from each hemisphere and incubated in 5 mL of formamide (for each hemisphere) in a 601C water bath for 24 h. The resultant solution was centrifuged at 14,000 g for 20 mins. The supernatant solution was collected and the optical density at 610 nm was measured to determine the relative amount of Evans blue dye in each sample.

Cytokine Enzyme-Linked Immunosorbent Assays Using rat-specific ELISA kits for interleukin (IL)-1b and IL-6 cytokine levels in sham and injured animals were determined as per the manufacturer’s instructions (Biotrak Amersham Biosciences Piscataway, NJ, USA). Two hundred microgram aliquots of soluble protein prepared from whole lungs were utilized in duplicate in each assay. For the serum analysis, trunk blood was collected in BD Vacutainer Tubes (BD Vacutainer, Preanalytical Solutions, Franklin Lakes, NJ, USA) and the serum was separated by centrifugation at 3000 r.p.m. for 10 mins at 41C. One hundred microliter of serum from each sample was assayed in duplicate. The absorbance readings were read at recommended wavelengths on Dynex ELISA reader within 30 mins of stopping the reactions. Each experiment was repeated at least two additional times.

Statistical Analysis Each experiment represents at least six independent determinations. Results are expressed as mean7s.d. To determine statistical differences, multiple groups were comparisons using one-way analysis of variance with a post hoc Tukey’s multiple range test. Values of P < 0.05 were considered statistically different.

Results

Confocal Microscopy

Initiation and resolution of pulmonary inflammation after TBI, evidence for a time-based switch of inflammatory response in lungs after TBI

Confocal images were captured using a Bio-Rad MRC 1024 2000 confocal microscope using a mixed gas krypton/

Histopathologic examination of hematoxylin and eosin-stained lung tissue sections of brain-injured

Journal of Cerebral Blood Flow & Metabolism (2006), 1–12


Temporal Synthesis and Degradation of Leukotriene B4 after Traumatic Brain Injury Coordinates Inflammation in Lungs

Since 5-LOX translocation from cytosol to the nuclear membrane is critical for LTB4 biosynthesis (Brock et al, 1995), the extent of transition in these two compartments was determined by immunoblotting. We found a substantial amount of 5-LOX translocation from cytosol to the nucleus at 24 h, indicating increased production of LTB4. Similarly, COX-2 expression increased to a maximum at 24 h with a subsequent decline by 2 weeks (Figure 3A). When immuno-reactive LTB4 levels were measured in the BALF, a striking increase (up to seven-fold higher than sham) was seen at 24 h postinjury

A

B

0.25

C Total Protein (mg/mL)

rats showed massive accumulation of acute inflammatory cells. We observed thickening of basement membrane with massive inflammatory cell infiltration around bronchioles, pulmonary blood vessels, and alveoli at 24 h postinjury (Figure 1A). Immunohistochemical staining identified appearance of significant number of neutrophils within the alveolar walls at 24 h after injury (Figure 1B). These cells started to disappear spontaneously by 3 days and the lung histology of 2 week injured rats was essentially similar to the shams. To evaluate the time and severity of lung injury caused by focal brain trauma, the lung permeability was assessed by determining the total protein content in the BALF. The results obtained from our study reveal nearly 1.8-fold higher protein concentration in the BALF at 24 h postinjury in comparison to shams (Figure 1C). The injury-induced protein increases subsided substantially on day 3 and the levels were completely normal by 2 weeks. Bronchio-alveolar lavage fluid cells were examined to determine the cellular differentials as a function of time after TBI. While macrophages were the predominant cell type in all groups, epithelial cells and eosinophils constituted less than 1% of the total number of cells. Small numbers of neutrophils (B 1%) were also present in the BALF of naive rats that were not altered by the sham surgeries. The number of neutrophils was significantly increased at 24 h postinjury (five-fold higher compared with shams), which returned to normal levels after 2 weeks (Figure 2A). Similarly, a significant increase (54% higher than shams) in total number of macrophages was also seen at 24 h after injury. Moreover, at 24 h postinjury, we observed considerable activation (measured as number of activated/ total number) of macrophages (characterized by increased clumping, size and granule appearance) in the BALF (Figures 2B and 2C). These results show that brain-injured rats develop pronounced pulmonary inflammation at 24 h characterized by the increase in neutrophils and accumulation of activated alveolar macrophages.

a

b

c

d

a

b

c

d

*

0.2 #

0.15 0.1 0.05 0

Sham

Inj 24h

Inj 3d

Inj 2Wk

Figure 1 (A) Histopathologic analysis of lungs after brain injury. Hematoxylin and eosin staining of paraffin sections of lung tissues showed acute inflammatory reactions in the lungs at 24 h after TBI. (a) Sham, (b) postinjury 24 h, (c) postinjury 3 day, and (d) postinjury 2 week. Red arrows indicate thickening of basement membrane and black arrows indicate the presence of inflammatory cells. By day 3, inflammatory infiltrates started to disappear and by 2 weeks the lungs of brain injured rats were completely normal compared with shams. Bronchioles (BR) and blood vessels (BV). Bars, 100 mm for 10 magnification and 50 mm for the 40 magnification (in sets). (B) The neutrophils (arrows) were detected by immunohistochemistry using antineutrophil elastase antibody. (a) Sham, (b) postinjury 24 h, (c) postinjury 3 day, and (d) postinjury 2 week. Significant numbers of neutrophils were present only in the lungs at 24 h after brain injury. (C) Total protein concentration in BALF displaying increased lung permeability at 24 h after brain injury. Each data point represents n = 6. * indicates P < 0.05 compared with the sham and # indicates P < 0.05 compared with post-injury 24 h.



A

*

BAL Cells × 105

2

Sham Injury 24h

*#

1.5

Injury 2 Wk 1 0.5

*

#

0 Macrophages

Lymphocytes

Eosinophils

Neutrophils

B a

C

b

% Activated macrophages

80

c

*

60

*#

40

20

0 Sham

Injury 24h

Injury 2 Wk

Figure 2 Analysis of the inflammatory cells in BALF after brain injury. (A) Quantitation of BALF cellularity. Total cells retrieved by BALF were increased 24 h after TBI. Significant increase in neutrophils at postinjury 24 h was evident whereas alveolar macrophages account for the overall increase of total cells. Cell counts and differentials were returning to normal by postinjury 2 week. (B) Hematoxylin and eosin stained BALF cells. Bronchio-alveolar lavage fluid cells were collected, cytospun and stained as described in Materials and methods section. (a) Sham, (b) postinjury 24 h, (c) postinjury 2 week. Red arrows point to activation of alveolar macrophages. Bars, 20 mm. (C) Quantitation of macrophage activation after injury. Each data point represents n = 6. * indicates P < 0.05 compared with sham and # indicates P < 0.05 compared with postinjury 24 h.

(Figure 3B) explaining the recruitment of neutrophils and inflammatory cells in the lung at this time point. Immuno-reactive LTB4 levels were significantly lower at 2 weeks compared with 24 h but were still elevated in comparison to shams. No significant changes of LTB4 levels in serum were observed after TBI, which rules out the possibility of its humoral spillover into the lung (Figure 3C). The CYP4F expression levels increased significantly with time after injury with maximum expression being detected at 2 weeks (Figure 4A). To quantitate CYP4F specific activity in lungs after injury, high-performance liquid chromatography Journal of Cerebral Blood Flow & Metabolism (2006), 1–12

analysis of LTB4 metabolism was performed and the quantities of newly formed 20-OH metabolite determined (Figure 4B). The results show a gradual increase in CYP4F activity with respect to time. Two weeks after injury there was a nearly two-fold increase in LTB4 o-hydroxylation compared with the shams (Figure 4C). 17-Octadecynoic acid, a known CYP4F suicide inhibitor, when tested for inhibition, entirely abolished the 20-OH LTB4 formation by the lung microsomes at 2 week postinjury. Increase in CYP4F expression and the associated LTB4 o-hydroxylase activity suggest the effulgence of the resolution phase for inflammation.

Role of CYP4F mediated LTB4 metabolism after brain trauma A Kalsotra et al

2W k In j

h In j3 d

24 In j

A

In j3 h

Sh am

7

Anti-5-LOX (Soluble) Loading Control Anti-5-LOX (Membrane) Anti-COX-2 (Membrane) Loading Control

B

*

BALF LTB4 (pg/mL)

160

Compared with the lung tissue, very little CYP4F expression in the BALF cells was observed at all of the time-points analyzed. To localize CYP4Fs more precisely, frozen sections of rat lung were probed with anti-CYP4F antibody (Figure 4E). Intense staining within the bronchial airways was apparent (see panels a and b). In addition, some areas of staining within the lung vasculature were in the smooth muscles of medium- and large-size vessels. Inspection with double-label immunofluoroscence at a higher magnification (see panels d to f) revealed the presence of specific staining in endothelial cells. Thus CYP4Fs are expressed in both vascular smooth muscle and endothelial cells of rat pulmonary vessels. Negative controls without the primary antibodies are shown in panel c. The overall background intensity was low, and the images had minimal staining, in contrast to the specific signal with the Anti-CYP4F antibody.

120

*# 80 40 0 Sham

Inj 24h

Inj 2Wk

Sham

Inj 24h

Inj 2Wk

C Serum LTB4 (pg/mL)

100 80 60 40 20 0

Figure 3 5-Lipoxygenase, COX-2 protein estimation and LTB4 quantitation in lungs after brain injury. (A) COX-2 induction and 5-LOX translocation from cytosol to nuclear membrane after brain injury. (B and C) EIA based pulmonary and serum LTB4 estimation after injury. Each data point represents n = 6. * indicates P < 0.05 compared with sham and # indicates P < 0.05 compared with postinjury 2 week.

mRNA and Protein Localization of Cytochrome P450 4Fs in Rat Lung

To uncover the source of CYP4F increase after brain injury, CYP4F4 mRNA expression in BALF cells or the lung tissue was examined in parallel. The results show a clear increase in CYP4F4 transcripts in the lung tissue, whereas no change is evident in the BALF compartment. An increase of four-, five- and six-fold from sham was observed at postinjury 24 h, 3 days, and 2 weeks, respectively (Figure 4D).

Blood–Brain Barrier is Compromised after Brain Injury but Brain-Derived Proinflammatory Cytokines do not Link Lung Inflammation and Traumatic Brain Injury

After TBI, the permeability of BBB to Evans blue bound albumin is dramatically increased on the ipsilateral hemisphere as represented by the OD values. The picture represents an injured brain removed at 24 h after TBI. Evans Blue extravasation can be clearly seen within the ipsilateral hemisphere (Figure 5A). Evans Blue extravasation was maximal between 24 h to 3 day after TBI, which subsided gradually thereafter. Two weeks after injury, the dye extraction values were significantly lower than 24 h (Figure 5B). A modest increase in permeability was observed on the contralateral hemisphere (24 h and 3 day time points), although this change was statistically insignificant. Next, we evaluated serum levels of the IL-1b and IL-6 after TBI to investigate whether the communication from the brain to the lung is mediated through these cytokines released in the blood after the transient leakage in the BBB (Figure 5C). Although IL-1b was modestly higher in serum at postinjury 2-week time point, IL-6 was undetectable in serum by enzyme-linked immunosorbent assay at the time points examined. On the contrary, IL-1b and IL-6 levels were significantly upregulated in the lungs at postinjury 24 h that returned to baseline by 2 weeks. (Figure 5C).

Discussion Brain injury is often associated with secondary injury, which turns out to be a major contributor to the extensive pathophysiology of TBI (Faden et al, 1989). Although there is some clinical data linking brain injury with lung inflammation (Fisher et al, 1999; Holland et al, 2003), to our knowledge, this is Journal of Cerebral Blood Flow & Metabolism (2006), 1–12


A

B

C

0.0 0.04 0.03 0.02 0.01 0.0

Retention Time (min)

D

*#

In j2 W k 17 -O DY A

0.03 0.02 0.01

* *

h In j3 d

0.0 0.04

9 8 7 6 5 4 3 2 1 0 24

0.01

In j

0.02

In j3 h

20-OH LTB4 (pmoles/min/mg)

Loading Control

0.03

Sh am

Anti-CYP4F (Microsomal)

Absorbance Units (270 nm)

In j2 W k

In j2 4h In j3 d

In j3 h

Sh am

0.04

E CYP4F4 mRNA (% Sham)

800 700

Lung tissue

600

BALF

*

a

b

c

d

e

f

* *

500 400 300 200 100 0 Sham

Inj 3h

Inj 24h Inj 3d

Inj 2Wk

Figure 4 Immunoblot and specific activity analyses of cytochrome P450 4Fs in lungs after brain injury. (A) Delayed CYP4F induction in lungs after TBI. (B) High-performance liquid chromatography chromatograms showing LTB4 o-hydroxylation. Leukotriene B4 metabolism was performed as described in the Materials and methods section. (C) Quantitation of 20-OH LTB4 production, specific activity of CYP4Fs, from lung microsomes. 17-Octadecynoic acid significantly inhibited the increased CYP4F activity at 2 week. (D) Cytochrome P4504 mRNA localization. Total RNA from BALF cells or lung tissue was harvested at 3 h, 24 h, 3 day or 2 weeks postinjury and expression of CYP4F4 quantitated by quantitative real-time polymerase chain reaction. Cytochrome P4504 expression was induced only in the lung tissue. (E) CYP4F protein localization. Representative confocal images of double-labeled lung sections using CYP4F2 (green) and vWF (red) antibodies. The immunoreactivity for CYP4F colocalizes with the endothelial marker, vWF. (a) Sham, (b) postinjury 24 h, (c) no primary antibody (negative control). High magnification ( 60) pictures showing cross-section of a pulmonary blood vessel (d) vWF, (e) CYP4F, and (f) colocalization. Bars, 100 mm for 20 (a–c) and 50 mm for 60 magnification (d–f). Each data point represents n = 6. * indicates P < 0.05 compared with sham and # indicates P < 0.05 compared with postinjury 2 week.

the first experimental report describing that how does central nervous system trauma initiate acute injury to the lung. This study demonstrates increased proteinaceous fluid in the air spaces (characterized through BALF total protein estimation) at 24 h after TBI. Total protein content in the BALF is a direct indicator of lung permeability. The increase in lung permeability can occur as a consequence of vascular breakdown, or via neutrophil (PMNLs) and other immune cell degranulation (Martin et al, 1989). Leukocyte trafficking in pulmonary tissue and air spaces is critical in the host defense response; however, migration and activation of neutrophils into lungs can contribute to inflammatory injury (Weiss, 1989). Activated PMNLs contribute to tissue damage by release of oxygen radicals, release of proteolytic enzymes such as elastase or metalloproJournal of Cerebral Blood Flow & Metabolism (2006), 1–12

teases, and stimulation and/or release of proinflammatory cytokines (Weiss, 1989). Moreover, influx of neutrophils into the air spaces is thought to cause flooding of alveoli by plasma liquid and proteins (Kaslovsky et al, 1993). This increase in protein permeability across the endothelial and epithelial barriers of the lung is an early characteristic of lung injury. Our results provide evidence that there is significant recruitment of cells into the lung after brain trauma, which acts to amplify the inflammatory response and modulate secondary tissue damage. We show that brain injury directs pulmonary neutrophil mobilization during the early 24 h phase. Whether a large increase in neutrophils and macrophages observed at this stage contributes to increased lung permeability remains unclear; nevertheless, pronounced neutrophilia as well as


A

B 0.8

*

a

O.D. value/g tissue

*

b

0.6

0.4

0.2

# 0.0 Sham

Injury 24h Injury 3d Injury 2Wk

C *

160 120 80

#

40 0 Sham

Injury 24h Injury 2Wk

c 150

*

120 90 60 30 0 Sham


Lung IL-6 (pg/mg protein)

b 200

Serum IL-1β (pg/mg/mL)

Lung IL-1β (pg/mg protein)

a

*

6 5 4

#

3 2 1 0 Sham


Figure 5 Evaluation of BBB compromise and major pro-inflammatory cytokines after brain injury. (A) In a separate set of animals, the Evans blue dye was injected through the jugular vein for 1.5 h of circulatory time, which was estimated in the brain tissue by visual analysis. Increased dye extravasation was observed at 24 h after TBI. (B) Quantification of extracted dye was performed using spectrophotometry at 610 nm. The dye extracted at 24 h and 3 days was up to 30-fold higher than shams. (C) Proinflammatory cytokines expression after brain injury. Interleukin-1b and IL-6 levels were estimated in the lungs or serum using commercially available ELISA kits. (a) Interleukin-1b in lungs, (b) IL-1b in serum, and (c) IL-6 in lungs. No IL-6 was detected in the serum at the time points examined. Each data point represents n = 6. * indicates P < 0.05 compared with sham and # indicates P < 0.05 compared with postinjury 24 h.

alveolar macrophage activation are clear signs of pulmonary inflammation. The data from our study are consistent with high levels of inflammation demonstrated in lungs of individuals who die from fatal brain injury (Fisher et al, 1999). Similar to our observations, a high neutrophil content in the BALF of the traffic accident victims was found (Fisher et al, 1999). These authors claimed further that the preinflamed lungs of such victims should not be used for transplantation as they may contribute to early graft dysfunction. Interestingly, we see a spontaneous recovery phase starting around 3 days. By 2 weeks after initial injury, inflammation resolves and the lungs have essentially recovered. Because LTB4 recruits neutrophils into the lungs (Martin et al, 1989), its role in PMNL mobilization after brain injury was considered. Previous studies have demonstrated increased LTB4 in rat brains after central nervous system injury (Lindgren et al, 1984) accompanied by accumulation of PMNLs and cerebral edema (Denzlinger et al, 1985; Schoettle et al, 1990; Tarlowe et al, 2003; Xu et al, 1990). In this study, we found significantly high levels of immuno-reactive LTB4 in

the lungs at 24 h postinjury, which is related to enhanced 5-LOX translocation from cytosol to the nuclear membrane, which in turn mobilizes and activates inflammatory cells. Similar to 5-LOX increase, we observed time-dependent and injurydriven induction of pulmonary COX-2, an enzyme responsible for prostaglandin synthesis during pathologic conditions (Funk, 2001). The precise role and nature of rise in COX-2 in this case is not clear at present; nevertheless, ample evidence exists for its role in exacerbating inflammatory responses both in the brain as well as in the periphery (Bazan, 2001; Ek et al, 2001). Since 5-LOX and immunoreactive LTB4 levels at 2 weeks after injury were still high compared with control, COX-2 derived prostanoids at this time may be playing pivotal roles in resolving inflammation. Similarly, the production of counterregulatory mediators like lipoxins at 2 weeks that may inhibit LTB4 signaling is an alternate possibility (Serhan and Savill, 2005). Cytochrome P450 4Fs catalyze LTB4 degradation, which manifests an important role for them in curtailing inflammation. We have previously reported that CYP4F expression is altered in a Journal of Cerebral Blood Flow & Metabolism (2006), 1–12


biphasic fashion in hippocampus after brain injury (Cui et al, 2003). Results from this study showed that, 2 weeks after the initial brain trauma, increased CYP4F expression and associated LTB4 metabolism in the lung were temporally coincident with decreases in neutrophil number. Loss of pulmonary PMNL recruitment later on in the inflammatory response shows a functional link between the cessation of neutrophil infiltration and LTB4 breakdown. Together, these findings suggest that the onset of LTB4 clearance is a regulator of PMNL trafficking during the progression of an acute inflammatory response. Our next goal was to localize the cellular source of CYP4F induction. The mRNA expression analysis revealed that these changes are predominant in the lung parenchyma rather than in BALF, a medium that is rich in immune cells. Immunofluoroscence studies provided evidence that CYP4F proteins are expressed in bronchial airways, vascular smooth muscle and endothelial cells of rat pulmonary vessels. Perhaps the most interesting aspect of pulmonary CYP4F expression is localization to endothelial cells. Adherence of circulating neutrophils to the microvascular endothelium is the initial step in diapedesis, the process by which leukocytes migrate through blood vessels to accumulate at sites of infection or injury (Springer, 1994). Leukotriene B4 clearance by CYP4Fs in endothelium may be an ideal site to resolve pulmonary inflammation initiated after brain injury. Next, we show that blunt trauma to the brain leads to increased BBB permeability up to 3 days postinjury but the BBB is sealed back by 2 weeks as measured using Evans blue extravasations. The results of BBB compromise led us to hypothesize that systemic release of certain mediators after brain injury might drive the inflammatory response seen in the lung. In patients with acute brain injury, systemic elevation of cytokines is thought to play a role in the pathology of injury that often occurs in several organs distant from the brain and may be a factor in determining clinical outcome (Fisher et al, 1999; Gruber et al, 1999; Woiciechowsky et al, 1998, 2002). While a time-based induction in IL-1b and IL-6 was seen in the lungs, only a modest increase in the levels of IL-1b in the blood was evident at 2 weeks after TBI. Therefore, the possibility of humoral spillover of these cytokines into the lung remains scant. Alternatively, neural mechanisms may be operating via vagal efferents to indirectly control peripheral cytokine and chemokine synthesis. Elevated intracranial pressure is known to stimulate neuroimmune pathways such as the hypothalamic– pituitary–adrenal axis and the sympathetic nervous system, which can aggravate the systemic immune changes resulting from the injury-triggered inflammatory response in the brain. Moreover, parasympathetic pathways like vagal release of acetylcholine can modulate the release of cytokines in the liver Journal of Cerebral Blood Flow & Metabolism (2006), 1–12

(Borovikova et al, 2000). Similar neural amplification pathways may account for the pulmonary cytokine response after TBI. In a separate study, we have shown that proinflammatory cytokines—IL-1b and IL-6—can regulate CYP4F expression in rat hepatocytes (Kalsotra et al, 2003a). Based on that evidence, we speculate that the local production of cytokines after brain trauma might play the role of a molecular switch to produce time-dependent, highly regulated CYP4F increase in the lung. Such dynamic aspects of inflammation have been previously studied (Serhan et al, 1996; Serhan and Savill 2005). Levy et al (2001) have shown a time-based switching from proinflammatory to anti-inflammatory eicosanoids after tumor necrosis factor-a mediated inflammation in the murine dorsal air pouch. We believe that this model of mediator class switching might apply to TBI as well. In summary, the results from this study reveal that trauma to the brain can produce acute changes in organs distant from the injury site, which contribute to multiple organ dysfunction syndrome. We conclude that in the proinflammatory phase, increased local production of LTB4 leads to mobilization of inflammatory cells such as neutrophils and macrophages into the lung. In the resolution phase, CYP4F expression and associated activity increases with time. The temporal association between CYP4F expression and changes in LTB4 levels are consistent with the hypothesis that the decline in LTB4 concentration is caused by induction of CYP4Fs. Further studies such as pharmacological treatment of animals with a 5-LOX inhibitor or an LTB4 receptor antagonist would elucidate the cause and effect relationship between LTB4 signaling and lung injury initiated after brain trauma. A parallel study performed on the mouse in a 5-LOX or CYP4F knockout background would also establish the relative importance of LTB4 catabolic pathway for resolving lung inflammation in brain injury victims.

Acknowledgements The authors are thankful to Dr Cheri M Turman and Dr Eric T Williams for their technical assistance. Our sincere thanks go to Dr Yasushi Kikuta and Dr Michael R Blackburn for their collegial assistance.

References Bazan NG (2001) COX-2 as a multifunctional neuronal modulator. Nat Med 7:414–5 Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI, Watkins LR, Wang H, Abumrad N, Eaton JW, Tracey KJ (2000) Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature 405: 458–462

Role of CYP4F mediated LTB4 metabolism after brain trauma A Kalsotra et al

Brock TG, McNish RW, Peters-Golden M (1995) Translocation and leukotriene synthetic capacity of nuclear 5-lipoxygenase in rat basophilic leukemia cells and alveolar macrophages. J Biol Chem 270:21652–8 Campbell SJ, Hughes PM, Iredale JP, Wilcockson DC, Waters S, Docagne F, Perry VH, Anthony DC (2003) CINC-1 is an acute-phase protein induced by focal brain injury causing leukocyte mobilization and liver injury. FASEB J 17:1168–70 Clifton GL, Robertson CS, Kyper K, Taylor AA, Dhekne RD, Grossman RG (1983) Cardiovascular response to severe head injury. J Neurosurg 59:447–54 Cui X, Kalsotra A, Robida AM, Matzilevich D, Moore AN, Boehme CL, Morgan ET, Dash PK, Strobel HW (2003) Expression of cytochromes P450 4F4 and 4F5 in infection and injury models of inflammation. Biochim Biophys Acta 1619:325–31 De Simoni MG, Sironi M, De Luigi A, Manfridi A, Mantovani A, Ghezzi P (1990) Intracerebroventricular injection of interleukin 1 induces high circulating levels of interleukin 6. J Exp Med 171:1773–8 Denzlinger C, Rapp S, Hagmann W, Keppler D (1985) Leukotrienes as mediators in tissue trauma. Science 230:330–2 Ek M, Engblom D, Saha S, Blomqvist A, Jakobsson PJ, Ericsson-Dahlstrand A (2001) Inflammatory response: pathway across the blood–brain barrier. Nature 410: 430–1 Faden AI, Demediuk P, Panter SS, Vink R (1989) The role of excitatory amino acids and NMDA receptors in traumatic brain injury. Science 244:798–800 Fisher AJ, Donnelly SC, Hirani N, Burdick MD, Strieter RM, Dark JH, Corris PA (1999) Enhanced pulmonary inflammation in organ donors following fatal nontraumatic brain injury. Lancet 353:1412–3 Ford-Hutchinson AW, Bray MA, Doig MV, Shipley ME, Smith MJ (1980) Leukotriene B, a potent chemokinetic and aggregating substance released from polymorphonuclear leukocytes. Nature 286:264–5 Funk CD (2001) Prostaglandins and leukotrienes: advances in eicosanoid biology. Science 294:1871–5 Gruber A, Reinprecht A, Illievich UM, Fitzgerald R, Dietrich W, Czech T, Richling B (1999) Extracerebral organ dysfunction and neurologic outcome after aneurysmal subarachnoid hemorrhage. Crit Care Med 27:505–14 Haeggstrom JZ (2004) Leukotriene A4 hydrolase/aminopeptidase, the gatekeeper of chemotactic leukotriene B4 biosynthesis. J Biol Chem 279:50639–42 Holland MC, Mackersie RC, Morabito D, Campbell AR, Kivett VA, Patel R, Erickson VR, Pittet JF (2003) The development of acute lung injury is associated with worse neurologic outcome in patients with severe traumatic brain injury. J Trauma 55:106–11 Kalsotra A, Cui X, Antonovic L, Robida AM, Morgan ET, Strobel HW (2003a) Inflammatory prompts produce isoform-specific changes in the expression of leukotriene B(4) omega-hydroxylases in rat liver and kidney. FEBS Lett 555:236–42 Kalsotra A, Turman CM, Dash PK, Strobel HW (2003b) Differential effects of traumatic brain injury on the cytochrome p450 system: a perspective into hepatic and renal drug metabolism. J Neurotrauma 20: 1339–1350 Kaslovsky RA, Lai L, Parker K, Malik AB (1993) Mediation of endothelial injury following neutrophil adherence to extracellular matrix. Am J Physiol 264:L401–5

Kikuta Y, Kusunose E, Kusunose M (2002) Prostaglandin and leukotriene omega-hydroxylases. Prostaglandins Other Lipid Mediators 68–69:345–62 Levy BD, Clish CB, Schmidt B, Gronert K, Serhan CN (2001) Lipid mediator class switching during acute inflammation: signals in resolution. Nat Immunol 2:612–9 Lindgren JA, Hokfelt T, Dahlen SE, Patrono C, Samuelsson B (1984) Leukotrienes in the rat central nervous system. Proc Natl Acad Sci USA 81:6212–6 Long DA, Ghosh K, Moore AN, Dixon CE, Dash PK (1996) Deferoxamine improves spatial memory performance following experimental brain injury in rats. Brain Res 717:109–17 Marion DW (1999) Management of traumatic brain injury: past, present, and future. Clin Neurosurg 45: 184–91 Martin TR, Pistorese BP, Chi EY, Goodman RB, Matthay MA (1989) Effects of leukotriene B4 in the human lung. Recruitment of neutrophils into the alveolar spaces without a change in protein permeability. J Clin Invest 84:1609–19 Mascia L, Andrews PJ (1998) Acute lung injury in head trauma patients. Intensive Care Med 24:1115–6 McIntosh TK, Juhlen M, Raghupathi R, Saatman KE, Smith DH (1999) Secondary brain injury: neurochemical and cellular mediators in traumatic brain injury. In: Traumatic brain injury (Marion DW ed) New York: Thieme Medical publishers, Inc., 39–54 Paine R, Smith JR, Howard FA (1952) Pulmonary edema in patients dying with disease of the central nervous system. J Am Med Assoc 149:643–6 Samuelsson B, Dahlen SE, Lindgren JA, Rouzer CA, Serhan CN (1987) Leukotrienes and lipoxins: structures, biosynthesis, and biological effects. Science 237:1171–6 Schoettle RJ, Kochanek PM, Magargee MJ, Uhl MW, Nemoto EM (1990) Early polymorphonuclear leukocyte accumulation correlates with the development of posttraumatic cerebral edema in rats. J Neurotrauma 7:207–17 Serhan CN, Haeggstrom JZ, Leslie CC (1996) Lipid mediator networks in cell signaling: update and impact of cytokines. FASEB J 10:1147–58 Serhan CN, Savill J (2005) Resolution of inflammation: the beginning programs the end. Nat Immunol 6:1191–7 Springer TA (1994) Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76:301–14 Stubbe HD, Greiner C, Van Aken H, Rickert CH, Westphal M, Wassmann H, Akcocuk A, Daudel F, Erren M, Hinder F (2004) Cerebral vascular and metabolic response to sustained systemic inflammation in ovine traumatic brain injury. J Cereb Blood Flow Metab 24:1400–8 Tarlowe MH, Kannan KB, Itagaki K, Adams JM, Livingston DH, Hauser CJ (2003) Inflammatory chemoreceptor cross-talk suppresses leukotriene B4 receptor 1mediated neutrophil calcium mobilization and chemotaxis after trauma. J Immunol 171:2066–73 Weiss SJ (1989) Tissue destruction by neutrophils. N Engl J Med 320:365–76 Woiciechowsky C, Asadullah K, Nestler D, Eberhardt B, Platzer C, Schoning B, Glockner F, Lanksch WR, Volk HD, Docke WD (1998) Sympathetic activation triggers systemic interleukin-10 release in immunodepression induced by brain injury. Nat Med 4:808–13

11



Woiciechowsky C, Schoning B, Cobanov J, Lanksch WR, Volk HD, Docke WD (2002) Early IL-6 plasma concentrations correlate with severity of brain injury and pneumonia in brain-injured patients. J Trauma 52:339–45 Xu JA, Hsu CY, Liu TH, Hogan EL, Perot PL, Jr, Tai HH (1990) Leukotriene B4 release and polymorphonuclear


cell infiltration in spinal cord injury. J Neurochem 55:907–12 Yokomizo T, Izumi T, Chang K, Takuwa Y, Shimizu T (1997) A G-protein-coupled receptor for leukotriene B4 that mediates chemotaxis. Nature 387: 620–4