INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Injection of lipopolysaccharide (endotoxin, LPS) results in inflammatory mediator production in the lungs as well as other organs; however, local production of CXC chemokines in the lungs may be particularly important for determining neutrophil migration into the alveolar space by creating a gradient of neutrophil attractant chemokines from the blood toward the alveoli. To examine the concept of a chemokine gradient between the alveolar space and blood regulating neutrophil influx into the lung, we have studied two different routes of delivery of LPS: intraperitoneal (IP) and intratracheal (IT). In a rat model of IT LPS-induced neutrophilic alveolitis, neutrophils are detectable by bronchoalveolar lavage (BAL) by 2 h after treatment with LPS (1). In contrast, IP LPS injection does not produce significant neutrophilic alveolitis until 18- 24 h after LPS injection (2). In addition, the dose of LPS required to initiate neutrophilic alveolitis is significantly lower when given intratracheally rather than intraperitoneally (0.01 versus 6.0 mg/kg) (our unpublished observation, 1998).

We postulated that the early onset and increased intensity of neutrophilic alveolitis after IT LPS compared with IP LPS treatment were due to increased lung versus systemic chemokine production, leading to an increased alveolar space-blood gradient of neutrophil chemotactic chemokines after IT LPS treatment. To address this hypothesis, we treated rats with (1) IP LPS at 6.0 mg/kg, (2) IT LPS at 0.1 mg/kg, or (3) IP + IT LPS given simultaneously. We measured lung NF-B activation, BAL and peripheral blood cell counts and differentials, total lung myeloperoxidase (MPO) activity, and alveolar space-blood chemokine gradients (measured as BAL-serum chemokine gradients) for the rat CXC chemokines macrophage inflammatory protein 2 (MIP-2) and cytokine-induced neutrophil chemoattractant (CINC) 4 h after LPS treatment. In addition, we measured BAL cell counts and differentials and alveolar space-blood chemokine gradients 24 h after IP LPS injection.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Model

Male Sprague-Dawley rats weighing between 200 and 300 g were used in all experiments. Escherichia coli lipopolysaccharide (serotype 055; B5) (Sigma Chemical Co., St. Louis, MO) was given by IP or IT injection. For IT injections, rats were anesthetized with ketamine and xylazine, and LPS (diluted in sterile physiologic saline) was instilled directly into the trachea in a total volume of 200-300 µl. After treatment with IT LPS, rats were asphyxiated with carbon dioxide. The rat tracheas were cannulated after death, and the lungs were lavaged in situ with sterile pyrogen-free physiologic saline. Saline was instilled in two 5-ml aliquots and gently withdrawn with a 5-ml syringe. For other studies, lungs were removed, quickly frozen in liquid N₂, and stored at 70° C.

BAL and Peripheral Blood Cell Counts and Differentials

Lung lavage fluid was centrifuged at 500 g for 10 min to separate cells from supernatant. Supernatant was saved separately and frozen, and the pelleted cells were suspended in a small amount of serum-free RPMI culture medium. Total cell counts were determined on a grid hemocytometer. Differential cell counts were determined by staining cytocentrifuge slides with a modified Wright stain (Diff-Quik; Baxter, McGaw Park, IL) and counting 400-600 cells in cross-section. Blood was obtained from heavily anesthetized rats by direct puncture of the inferior vena cava or aorta followed by direct aspiration into a standard purple-top, EDTA-containing vacuum tube. A complete blood count (CBC) was determined in the clinical laboratory at the Veterans Affairs Medical Center (Nashville, TN), using an automated Advia-120 hematology analyzer (Bayer, Leverkusen, Germany). Differential cell counts were performed on blood smear slides that were stained with Wright stain by visually categorizing 100-200 random cells under a light microscope (×400).

Measurement of Immunoreactive Rat MIP-2, TNF-, and CINC

Rat MIP-2 and tumor necrosis factor  (TNF-) were measured in cell culture supernatant using commercially available enzyme-linked immunosorbent assays (ELISAs) according to the manufacturer instructions (rat MIP-2 [Biosource International, Camarillo, CA] and rat TNF- [Genzyme, Cambridge, MA]).

CINC was measured by sandwich ELISA. The ELISA was constructed in our laboratory using goat anti-CINC IgG (R&D Systems, Minneapolis, MN) and rabbit anti-CINC IgG. The polyclonal rabbit antiserum against CINC was raised using recombinant CINC. A glutathione S-transferase (GST)-CINC cDNA was a gift from J. DeLarco and A. Wittwer (Monsanto Research Institute, St. Louis, MO). This GST-CINC fusion protein construct in the plasmid pJZ240 was expressed in E. coli strain XL-1B Blue (Stratagene, La Jolla, CA) and expression of GST-CINC fusion protein was induced by adding 100 mM isopropyl--D-thiogalactopyranoside (IPTG). Rabbit antiserum against CINC was raised by injecting GST-CINC and adjuvant into naive rabbits. Rabbit IgG was purified by precipitation with saturated ammonium sulfate, and then washed and dialyzed against 50 mM TRIS (pH 7.5). Anti-CINC IgG was prepared by two passages of IgG over a cyanogen bromide-activated Sepharose column that was cross-linked with GST-CINC (ligand affinity chromatography). Anti-CINC IgG was eluted, concentrated with a Centricon filter (molecular weight cutoff 3,000; Amicon, Danvers, MA), dialyzed against 50 mM TRIS, and saved. Preliminary data showed that this rabbit anti-CINC IgG detects rGST-CINC or rCINC but does not react to albumin, GST, or rat rMIP-2 on Western immunoblots (not shown). The CINC ELISA is linear against rCINC between 30 and 2,000 pg/ml in biological fluids.

Nuclear Protein Extractions and Electrophoretic Mobility Shift Assays

Nuclear protein extraction from lung tissue and electrophoretic mobility shift assays (EMSAs) for NF-B were done as previously described (3). An oligonucleotide probe containing a consensus NF-B motif (Stratagene) was used in these studies.

Total Lung Myeloperoxidase Activity

Neutrophil infiltration into the lung was quantified by measuring myeloperoxidase activity in lung tissue. A portion of frozen lung (125- 175 mg) was weighed and homogenized in 2 ml of 50 mM potassium phosphate, pH 6.0, with 5 mM EDTA. The homogenate was sonicated and centrifuged at 3,000 rpm for 15 min. The supernatant was mixed with assay buffer (1:30, vol/vol) and placed in a spectrophotometer for reading at 460 nm. The assay buffer consisted of 100 mM potassium phosphate (pH 6.0), H₂O₂ (30% stock diluted 1:100; Sigma), and o-dianisidine hydrochloride (1 mg/ml) (Sigma). The activity of purified MPO (Sigma) was also determined, showing that the assay was linear over a range of at least 0.03-3 U/ml (data not shown). The results were measured as MPO activity per milligram of lung tissue and are reported as MPO units per lung (using the average rat lung weight in the experiment).

Statistical Analysis

For comparison among groups, a one-way analysis of variance (ANOVA) was used with the Tukey-Kramer multiple comparisons test (p values of  0.05 were considered significant).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Simultaneous Treatment with Intraperitoneal LPS Inhibits Neutrophilic Alveolitis Induced by Intratracheal LPS

Intratracheal (IT) administration of LPS in rats causes early production of polymorphonuclear leukocyte (PMN) chemotactic chemokines and PMN influx by 2 to 4 h after LPS injection (1, 4). Intraperitoneal (IP) LPS treatment also results in systemic inflammation and increased serum cytokine concentrations but no neutrophilic alveolitis in the first few hours after LPS injection (2). We postulated that a major reason for the differential PMN influx into the lung after IT or IP LPS is that the route of LPS administration affects the production of chemokine gradients, which influences PMN migration from blood into alveolar spaces. After IT LPS, increased local chemokine production in the lung compared with systemic chemokine production might produce a large gradient favoring PMN migration into the alveoli; however, IP LPS injection might produce more systemic than lung chemokine production, producing an unfavorable chemokine gradient for PMN influx into the alveolar space in the first few hours after LPS injection.

We thought that simultaneous treatment with both IT and IP LPS could decrease early PMN influx into the alveolar spaces by decreasing alveolar space-blood chemokine gradients. Therefore, we treated rats with (1) IP LPS at 6.0 mg/kg, (2) IT LPS at 0.1 mg/kg, or (3) IP + IT LPS given simultaneously. We measured BAL cell counts and differentials 4 h after LPS treatment. Total PMN and macrophage counts in 10 ml of BAL are shown in Figure 1A, and the percentages of lavage cells, PMNs, and macrophages are shown in Figure 1B. The total number of cells obtained by lavage was significantly higher in the IT LPS group than in the other three groups (IT LPS = 109 × 10⁴, control = 24 × 10⁴, IP LPS = 27 × 10⁴, IP + IT LPS = 13 × 10⁴). As shown in Figure 1A, this difference in lavage cells was due entirely to the increased numbers of PMNs in the IT LPS group. There was a striking migration of PMNs into the alveolar space in the rats treated with IT LPS that was largely blocked by simultaneous treatment with IP LPS (IT LPS versus IP + IT LPS groups, Figure 1A). BAL PMN counts were 30-fold higher in the IT LPS group than in the IP + IT LPS group (IT LPS = 90 × 10⁴, IP + IT LPS = 3 × 10⁴, p < 0.001). The percentage of PMNs in BAL was also significantly different between these two groups, with PMNs representing 82% of BAL cells in the IT LPS group and 24% in the IP + IT LPS group (p < 0.001) (Figure 1B). As expected, no neutrophilic alveolitis was evident at 4 h in the control group or in the group receiving IP LPS alone. We have already reported that treatment with IP LPS does not produce neutrophilic alveolitis detectable by BAL until 18-24 h after injection (2).

View larger version (20K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   Figure 1.   Lung lavage cell counts in rats treated with intraperitoneal (IP) LPS, intratracheal (IT) LPS, or IP + IT LPS. (A) Total cells (macrophages and neutrophils) in lung lavage (means ± SEM) in untreated controls (n = 3), rats treated with IP LPS at 6.0 mg/kg (n = 3), rats treated with IT LPS at 0.1 mg/kg (n = 5), and rats treated simultaneously with IP and IT LPS (n = 5). In this experiment, lung lavage was done 4 h after LPS treatment with 10 ml of sterile saline. (*p < 0.001 for IT LPS compared with all other groups for total neutrophils.) (B) Percentage of macrophages and neutrophils in lung lavage from this experiment. (*p < 0.001 for IT LPS compared with IP + IT LPS.)

Total white blood cell, red blood cell, and platelet counts were performed using an automated cell counter (Figure 2). There was a 60% reduction in circulating leukocyte counts in the IP LPS and IP + IT LPS groups compared with the control group at 4 h after LPS treatment (Figure 2A). Thrombocytopenia was evident in the IP + IT LPS group but there were no differences in red blood cell counts in any of the groups (Figure 2A). Differential cell counts revealed a more than threefold increase in circulating neutrophils in the IT LPS group compared with the control group. There was a trend toward neutropenia in the IP + IT LPS group that did not reach significance (Figure 2B). Interestingly, both lymphopenia and monocytopenia were evident in the IP LPS and IP + IT LPS groups.

View larger version (33K): [in this window] [in a new window]   View larger version (31K): [in this window] [in a new window]   Figure 2.   Circulating blood cell counts in rats treated with intraperitoneal (IP) LPS, intratracheal (IT) LPS, or by combination treatment. (A) Total white blood cell (WBC) counts, red blood cell (RBC) counts, and platelet (PLT) counts (mean cell count/mm3 ± SEM) in untreated controls (open bars, n = 6), rats treated with IT LPS at 0.1 mg/kg (forward cross-hatched bars, n = 6), rats treated with IP LPS at 6.0 mg/kg (backward cross-hatched bars, n = 5), and rats treated simultaneously with IP and IT LPS (double cross-hatched bars, n = 5). (B) Differential leukocyte counts from the same experiment (mean cell count/mm3 ± SEM) for neutrophils (PMN), lymphocytes (LYMPH), and monocytes (MONO). In this experiment, blood was obtained 4 h after LPS treatment. (Single [*] and double [**] asterisks indicate counts that are significantly lower or higher than in the control group [p < 0.05], respectively.)

We measured total myeloperoxidase activity (MPO) in lungs from six individual rats in each group at 4 h after LPS treatment (Figure 3). Compared with the control group, there was a 79% increase in MPO activity in the IP LPS group, a 155% increase in MPO activity in the IT LPS group, and a 134% increase in MPO activity in the IP + IT group (p < 0.05). These data indicate that the total lung PMN burden, as reflected by MPO activity, is increased in the IP LPS, IT LPS, and IP + IT groups. Because similar numbers of PMNs are present in the lungs of rats treated with IT LPS or IP + IT LPS, combined treatment apparently decreases the fraction of PMNs that migrate across the alveolar-capillary barrier in response to IT LPS.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Myeloperoxidase activity per average lung weight. MPO is reported as MPO activity per lung (mean ± SEM) for controls (n = 6), rats treated with IP LPS at 6.0 mg/kg (n = 6), rats treated with IT LPS at 0.1 mg/kg (n = 6), and rats treated simultaneously with IP LPS and IT LPS (IP + IT LPS) (n = 6). Total lung weights were not different among the four groups and the results were calculated using an average total lung weight of 1,860 mg. (* Significantly more activity than in the control group [p < 0.05].)

Simultaneous Treatment with Intraperitoneal LPS Alters BAL-Serum Cytokine Gradients Induced by Intratracheal LPS Treatment

Because the route of LPS administration (intratracheal or intraperitoneal) determines the timing and intensity of neutrophilic alveolitis, we sought to determine whether this differential effect on PMN influx could be explained by differences in the gradients of neutrophil chemotactic chemokines favoring emigration of PMNs from the circulation into the alveoli. Therefore, we measured the levels of two important CXC chemokines in the rat, CINC and MIP-2, in both serum and BAL, and calculated BAL-serum gradients (Figures 4 and 5). For MIP-2, high concentrations were detected in BAL from both IT LPS (10.6 ± 0.2 ng/ml) and IP + IT LPS (12.0 ± 0.4 ng/ml) groups, and significantly higher concentrations were found in the sera of rats treated with IP + IT LPS (9.9 ± 1.6 ng/ml) than with IT LPS alone (0.4 ± 0.1 ng/ml, p < 0.001) (Figure 4A). No MIP-2 was detected in the BAL or serum of untreated controls, and modest increases in both BAL and serum MIP-2 were found in the IP LPS group. The calculated chemokine BAL-serum gradient was significantly higher for MIP-2 in the IT LPS group than in the IP + IT LPS or IP LPS group (Figure 4B).

View larger version (29K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 4.   MIP-2 levels in serum and lung lavage 4 h after treatment with LPS. (A) Concentration of MIP-2 as measured by ELISA (mean ± SEM) in serum and lung lavage of untreated controls (n = 3), rats treated with IP LPS at 6.0 mg/kg (n = 3), rats treated with IT LPS at 0.1 mg/kg (n = 5), and rats treated simultaneously with IT LPS and IP LPS (IP + IT LPS) (n = 5). (B) BAL/serum ratios of MIP-2 were calculated for each group from the ELISA measurements presented in Figure 4A. Values are means ± SEM. *p < 0.01 for IT LPS group compared with IP LPS and IP + IT LPS groups.

View larger version (32K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Figure 5.   CINC levels in serum and lung lavage 4 h after treatment with LPS. (A) Concentration of CINC as measured by ELISA (mean ± SEM) in serum and lung lavage of untreated controls (n = 3), rats treated with IP LPS at 6.0 mg/kg (n = 2), rats treated with IT LPS at 0.1 mg/kg (n = 5), and rats treated simultaneously with IT LPS and IP LPS (IP + IT LPS) (n = 5). (B) BAL/serum ratios of CINC were calculated for each group from the ELISA measurements presented in Figure 5A. Values are means ± SEM. *p < 0.01 for IT LPS group compared with IP + IT LPS group.

For CINC, low concentrations were detected in the BAL from untreated controls, but none was detected in the sera of these rats. Markedly increased levels of CINC were found in the BAL from rats treated with IT LPS (1.7 ± 0.1 ng/ml) and IP + IT LPS (1.4 ± 0.1 ng/ml); however, as with MIP-2, a significant difference was detected in serum CINC concentration between these two groups (IT LPS = 1.2 ± 0.1 ng/ ml, IP + IT LPS = 2.0 ± 0.1 ng/ml, p < 0.001) (Figure 5A). The calculated CINC BAL-serum gradient was significantly higher in the IT LPS group than in the IP + IT LPS or IP LPS group (Figure 5B).

In addition to measuring the levels of CXC chemokines, we also measured levels of TNF- in BAL and serum to determine whether the detected differences in BAL-serum gradients were specific for chemokines or generalizable to other cytokines (Figure 6). TNF- was not detectable in the BAL or sera from control rats or in the BAL from rats treated with IP LPS. Low TNF- concentrations were found in the sera of rats treated with IP LPS (31 ± 15 pg/ml). As with the chemokines, high levels of TNF- were detectable in the BAL of both the IT LPS group (935 ± 306 pg/ml) and the IP + IT LPS group (943 ± 243 pg/ml), but the concentration of TNF- in the serum was much higher in the IP + IT LPS group (IP + IT LPS = 197 ± 72 pg/ml, IT LPS = 10 ± 6 pg/ml). For TNF- and the chemokines MIP-2 and CINC, IT LPS produced high concentrations in the BAL at 4 h whether given alone or in combination with IP LPS; however, the serum concentration of these mediators was much greater when IP LPS was given in combination with IT LPS.

View larger version (24K): [in this window] [in a new window]   Figure 6.   TNF- levels in serum and lung lavage 4 h after treatment with LPS. The concentrations of TNF- were measured by ELISA (means ± SEM) in serum and lung lavage of untreated controls (n = 3), rats treated with IP LPS at 6.0 mg/kg (n = 3), rats treated with IT LPS at 0.1 mg/kg (n = 5), and rats treated simultaneously with IT LPS and IP LPS (IP + IT LPS) (n = 5).

Activation of NF-B in the Lungs Is Increased in Rats Treated with Intraperitoneal Plus Intratracheal LPS Compared with Rats Treated with Intratracheal LPS Alone

To address the question of whether treatment with IP LPS in combination with IT LPS might affect local inflammatory mediator production in lungs induced by IT LPS, we measured activation of the transcription factor NF-B in lung tissue by electrophoretic mobility shift assay (EMSA). NF-B activation is required for the gene expression of both MIP-2 and CINC and is involved in the upregulation of TNF- gene expression (4). NF-B activation was detected in lung tissue nuclear protein extracts 4 h after IT LPS treatment (Figure 7A, lanes 1-4) and after treatment with IP + IT LPS (Figure 7A, lanes 5-8), and in both groups was increased when compared with controls (not shown). In this EMSA, band A contains authentic NF-B (intact p65/p50 heterodimers) and band B contains p50 homodimers. By densitometry of band A (Figure 7B), which contains the trans-activating form of NF-B, activation of NF-B in the lung was increased in the IP + IT LPS group when compared with the IT LPS group, implying that local activation of NF-B is not sufficient to direct PMN migration into the alveoli after an inflammatory stimulus. In this experiment, lung activation of NF-B and total chemokine production (serum and BAL) did not predict neutrophil influx into the alveolar space.

View larger version (15K): [in this window] [in a new window]   View larger version (70K): [in this window] [in a new window]   Figure 7.   Electrophoretic mobility shift assay (EMSA) for NF-B-binding activity in nuclear extracts from lung tissue 4 h after treatment with LPS. (A) Lanes 1-4 contain samples from lungs of rats treated with IT LPS (0.1 mg/kg), and lanes 5-8 contain samples from lungs of rats treated with IT LPS (0.1 mg/kg) and IP LPS (6.0 mg/kg). Fifty nanograms of cold (C) unlabeled NF-B probe or an unrelated DNA sequence (NS, nonspecific) was added to the nuclear protein sample used in lane 8 to assess specificity of protein binding for the NF-B motif (lanes 9 and 10, respectively). Cold NF-B oligonucleotide successfully competed for NF-B binding and eliminated bands A and B (lane 9), but addition of an oligonucleotide without an NF-B-binding motif did not affect either band (lane 10), indicating that both bands are specific for NF-B. Antibody to p65 was added to the lane 8 sample and resulted in a supershifted band with diminution of band A (lane 11). Antibody to p50 supershifts both bands (A and B) (not shown), indicating that band A contains p65/p50 heterodimers and band B contains p50 homodimers. (B) The mean ± SEM of the laser densitometry readings of band A in lanes 1-4 (IT LPS) compared with lanes 5-8 (IP + IT LPS). The asterisk (*) indicates that these groups are significantly different, p < 0.05.

Chemokine Gradients Predict PMN Migration into the Lungs 24 h after Intraperitoneal Injection of LPS

After determining that BAL-serum gradients for MIP-2 and CINC correlated well with neutrophilic alveolitis 4 h after IT and IP LPS treatment, we performed further experiments to determine whether chemokine gradients could explain the delayed PMN alveolitis that occurs 24 h after intraperitoneal injection of LPS. Rats were injected intraperitoneally with LPS at 6.0 mg/kg, and BAL was done 24 h later for evaluation of neutrophilic alveolitis and calculation of BAL-serum chemokine gradients. Lavage cell counts and differentials are shown in Figure 8. A significant PMN influx into the alveolar spaces occurred after IP LPS treatment. Total BAL cell counts were higher in the rats treated with IP LPS compared with controls [controls, 32 (± 10) × 10⁴; IP LPS, 516 (± 100) × 10⁴; p < 0.01]. As shown, there was a marked increase in PMNs but no significant change in alveolar macrophages in the lavagable air space 24 h after treatment with IP LPS. PMNs accounted for 76% of BAL cells (on average) 24 h after IP LPS injection (compared with < 1% in controls).

View larger version (22K): [in this window] [in a new window]   Figure 8.   Lung lavage cell counts in rats 24 h after treatment with intraperitoneal (IP) LPS. Counts represent total cells (macrophages and neutrophils) in lung lavage (mean ± SEM) in untreated controls (n = 6) and rats treated with IP LPS at 6.0 mg/kg (n = 6). In this experiment, lung lavage was done 24 h after LPS treatment with 10 ml of sterile saline. *p < 0.001 for IP LPS compared with the control group for total neutrophils.

Low levels of MIP-2 were detected in the sera of untreated controls and LPS-treated rats in this experiment (controls, 74 ± 50 pg/ml; LPS treated, 40 ± 10 pg/ml) (Figure 9). In BAL, substantial MIP-2 levels were found in BAL from LPS-treated rats (859 ± 144 pg/ml) but only low concentrations were found in controls (96 ± 5.0 pg/ml). The calculated BAL- serum gradient for MIP-2 in the LPS group was significantly higher than in the control group (p < 0.001) and compared favorably with the gradients that were measured 4 h after IT LPS injection. In this experiment, CINC was not detectable in sera or BAL from controls or 24 h after intraperitoneal injection of LPS.

View larger version (24K): [in this window] [in a new window]   Figure 9.   MIP-2 level in serum and lung lavage 24 h after treatment with IP LPS. Concentration of MIP-2 was measured by ELISA (mean ± SEM) in serum and lung lavage of untreated controls (n = 6) and rats 24 h after treatment with IP LPS (6.0 mg/kg) (n = 6). *p < 0.001 for IP LPS compared with the control group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PMN immigration into the alveolar space is a complex process involving attachment to the endothelium and migration across the endothelial barrier. This process is regulated by adhesion molecules that control attachment to the endothelium and by chemotactic factors that influence transendothelial migration to sites of inflammation. Several chemotactic factors for PMNs have been identified, including complement component C5a, platelet-activating factor, leukotriene B₄, formylmethionyl bacterial peptides, and CXC chemokines containing the ELR motif (5, 6). CXC chemokines include interleukin-8 (IL-8); Gro-, -, and -; MIP-2; CINC; and epithelial neutrophil-activating protein (ENA-78). These mediators are upregulated by a variety of inflammatory stimuli and are critical for directing neutrophil accumulation in a variety of animal models of inflammation (7).

In our model of neutrophilic alveolitis induced by injection of LPS, the timing and intensity of PMN influx into the lavagable air space are related to the route of LPS delivery, implying compartmentalization of the inflammatory response. We have identified in vivo CXC chemokine gradients across the alveolar capillary barrier that may explain these findings. Intratracheal injection of LPS resulted in intense neutrophilic alveolitis by 4 h in association with BAL-serum chemokine gradients favoring immigration of PMNs into the alveolar space. In contrast, IP LPS injection resulted in elevated BAL and serum chemokine levels at 4 h, without a substantial gradient favoring neutrophilic alveolitis. However, the BAL-serum MIP-2 gradient measured 24 h after IP LPS was found to be similar to that measured 4 h after IT LPS. BAL MIP-2 levels were near 3 ng/ml at 4 h after IP LPS and had declined to near 1 ng/ml at 24 h after IP LPS. During the same interval, serum levels of MIP-2 had fallen from 3 ng/ml to less that 0.05 ng/ml. Thus, while the BAL-serum ratio was near 1 at 4 h, it increased to more than 20 by 24 h, apparently providing a sufficient chemotactic gradient to direct PMN migration into the alveolar spaces.

Four hours after LPS treatment, the average BAL-serum gradient for MIP-2 was 45-fold greater in rats treated with IT LPS compared with those treated with IP LPS and 28-fold greater in rats treated with IT LPS compared with rats undergoing combined treatment. A similar but less pronounced pattern was seen with CINC. The average BAL-serum gradient for CINC at 4 h was 10-fold greater in rats that were treated with IT LPS compared with those treated with IP LPS and 2-fold greater after IT LPS compared with treatment with IP and IT LPS. This pattern seemed to be relatively selective for CXC chemokines because the BAL-serum ratios for TNF- levels were high in both the IT and IP + IT LPS groups. Thus, for both MIP-2 and CINC, simultaneous treatment with IP LPS and IT LPS altered the BAL-serum chemokine gradients by increasing the serum level, as compared with levels in BAL fluid. These findings, taken together, provide strong evidence that in vivo chemotactic gradients direct PMN migration into the alveolus after a pulmonary inflammatory stimulus. The gradients of chemokines across the alveolar capillary barrier, rather than the absolute concentration of chemokines, determine the timing and intensity of neutrophilic alveolitis. Although others have shown that decreasing chemokine levels in the lung blunt the development of neutrophilic alveolitis in animal models, we now show that increasing systemic levels of chemokines have the same effect.

Because chemokines are basic proteins, they bind to negatively charged heparin and heparin sulfate (12, 13). These bound chemokines may not be identified by lung lavage, but they may augment chemokine gradients favoring immigration of PMNs into the lungs. In addition, binding of chemokines to the extracellular matrix may be a mechanism for the prolonged retention of chemokines in the lung compared with the circulation after IP LPS injection. We have observed that CINC, as detected by immunohistochemistry, lines the alveolar epithelial cell surfaces from 2 to 48 h after IP LPS injection (T. S. Blackwell and J. W. Christman, unpublished observation, 1998). In addition to heparin binding, chemokines also bind to the human Duffy antigen receptor for chemokines (DARC) on erythrocytes (14). DARC binds chemokines but does not signal, and may function as a neutralizing sink for chemokines, increasing clearance from the circulation (15). A mouse homolog of the human Duffy gene has been described; its product also binds chemokines and may function like human DARC (16). The combination of prolonged chemokine binding to the extracellular matrix in the lung and clearance of chemokines from the blood via DARC may establish the chemokine gradient necessary for PMN influx into the lungs 24 h after IP LPS injection.

The idea that chemokine gradients determine neutrophil influx to sites of inflammation is consistent with the observation that injection of chemokines into specific sites causes PMN recruitment, but systemic injection of chemokines does not produce tissue neutrophilic inflammation or injury. For example, when injected intradermally into humans, IL-8 induces a time-dependent perivascular neutrophil influx (17); however, intravenous injection of IL-8 into baboons causes no hemodynamic abnormalities, no detectable production of TNF-, IL-1, or IL-6, and no tissue PMN influx (18). In addition, this concept may also help explain previously published findings that intravenous injection of IL-8 blocks PMN accumulation induced by intradermal injection of a variety of inflammatory mediators and inhibits PMN influx into ischemic myocardium (19, 20).

Because peripheral blood PMN counts were markedly increased after treatment with IT LPS but blood PMN counts were not significantly different among the other groups, we thought that the availability of PMNs could be a factor in determining the intensity of neutrophilic alveolitis. After IP LPS, PMNs could be recruited to the peritoneum and other sites, and become unavailable for sequestration in the pulmonary vasculature and emigration to the alveolar space. This issue was addressed by measuring lung MPO activity as a reflection of the total lung PMN burden. Interestingly, no significant differences were detected among the three treatment groups, suggesting that similar numbers of neutrophils were sequestered in the lung vasculature and/or interstitium in the IP LPS and IP + IT LPS groups. These data, together, support the conclusion that a chemokine gradient rather than availability of neutrophils is the major determinant of the intensity of PMN influx into alveolar spaces.

Frevert and coworkers (21) pretreated rats with intravenous (IV) LPS 2 h before IT LPS injection and found that PMN influx into the alveolar spaces 4 h later was markedly reduced. They measured chemotactic activity in the BAL and plasma by in vitro assays and found no difference in the chemotactic gradient when measured in this way. The only difference they found in rats pretreated with IV LPS followed by IT LPS and rats treated with IT LPS only was that incubation of normal PMNs with plasma from rats treated with IV LPS resulted in loss of L-selectin and increased MAC-1. They speculated that this loss of L-selectin occurred in vivo and resulted in decreased neutrophil adhesion in the pulmonary vasculature after IV LPS. This explanation seems unlikely to account fully for the dramatic difference in neutrophilic alveolitis in these two groups, because on morphological examination equal numbers of PMNs were sequestered in the pulmonary vasculature in controls and rats pretreated with IV LPS.

Because NF-B is thought to be important in the gene expression of all neutrophil chemotactic chemokines and the neutrophil-endothelial cell adhesion molecules ICAM-1 and E-selectin (22), we measured NF-B activation in lung tissue to determine whether simultaneous treatment with IP LPS affected lung NF-B activation induced by IT LPS. We found that combined IP and IT LPS treatment resulted in increased NF-B activation in lung tissue compared with IT LPS treatment, implying that IP LPS treatment does not diminish the ability of the lung to respond to IT LPS. After treatment with IT LPS and with combined IT and IP LPS treatment, there was similar production of MIP-2, CINC, and TNF- in the lung; only the circulating concentrations were different in our various treatment groups. If NF-B-mediated chemokine production in the lungs compared with other organs determines the alveolar space-blood chemokine gradient and is critical for development of neutrophilic alveolitis, inhibiting lung NF-B activation could substantially alter the alveolar space- blood chemotactic gradient and significantly diminish the resultant neutrophilic influx. Attenuating the systemic release of chemokines without decreasing concentrations in the lung could, paradoxically, increase neutrophilic alveolitis. These data seem to indicate that specifically targeting NF-B in the lungs for therapeutic intervention may be a particularly effective strategy for modulating neutrophilic alveolitis.

In summary, these data show that local chemokine gradients across the alveolar capillary barrier predict the intensity of neutrophilic alveolitis after treatment with LPS. A high BAL/serum chemokine ratio is associated with alveolar neutrophilia. Interventions to limit neutrophilic alveolitis could either be targeted to block local lung chemokine production or to increase systemic production of chemokines. In our studies, circulating chemokine concentrations are a critical determinant of the intensity of alveolitis. This observation has an important implication for the treatment of lung inflammation if our data can be extrapolated to human disease. It is possible that systemic antiinflammatory strategies, including IL-1 receptor antagonists (IL-1ra), anti-TNF- antibodies, and soluble TNF receptor (sTNF-R) chimeric molecules, might augment PMN influx into the alveolar spaces. If therapeutic concentrations of these agents are not achieved in lung they could, paradoxically, augment neutrophilic alveolitis by decreasing systemic chemokine concentrations without affecting chemokine production in the lung. Future clinical trials of systemic inflammatory agents should use sensitive markers to monitor systemic and lung inflammation independently. Our data predict that local abrogation of chemokine production in the lung would be a more effective way to decrease inflammation than systemic antiinflammatory agents.