Published online before print November 7, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
,,
,1
* Departments of Microbiology and Immunology and
Pathology and Laboratory Medicine, Weill Medical College of Cornell University, and Graduate Programs in
Immunology and Microbial Pathogenesis and
Molecular Biology, Weill Graduate School of Medical Sciences of Cornell University, New York, New York
1 Correspondence: Box 62, Weill Cornell Medical College, 1300 York Avenue, New York, NY 10021. E-mail: cnathan{at}med.cornell.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: phox • respiratory burst • tumor necrosis factor
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Inflammatory mediators can prime neutrophils in suspension to undergo an enhanced secretory response to other inflammatory mediators. However, soluble inflammatory mediators only induce a full-scale secretory response if the neutrophils are able to spread on biological surfaces while engaging their integrins [4 ]. "Full-scale" responses are those whose magnitude is similar to the maximal effect induced by phagocytic particles or phorbol esters. The full-scale response to soluble, physiologic mediators requires binary signals. Ligation of integrins during adherence to extracellular matrix (ECM) initiates one type of signal. The other set of signals comes from the ligation of receptors for inflammatory cytokines, chemokines, and oligopeptides, such as tumor necrosis factor (TNF), lymphotoxin, C5a, macrophage inflammatory protein-1, granulocyte macrophage-colony stimulating factor (GM-CSF), G-CSF, and formylated peptides [5 6 7 8 ]. The requirement for binary signaling may help prevent activation of circulating neutrophils in the blood.
TNF-activated neutrophils have been implicated in the pathology of numerous inflammatory syndromes and diseases, such as septic shock, adult respiratory distress syndrome, rheumatoid arthritis, ankylosing spondylitis, and Crohn’s disease. Activated neutrophils accumulate at the affected sites; TNF is elevated; and anti-TNF treatment reduces tissue damage in several of these settings [9 10 11 12 ]. However, despite intensive research on TNF signaling [6 , 13 14 15 16 17 18 19 20 21 22 23 ], it remains unclear how TNF activates neutrophils. Most investigations on TNF signaling have focused on pathways that lead to gene expression, cell proliferation, or cell death, usually in transformed cell lines. In contrast, activation of neutrophils by TNF is independent of transcription and translation, and none of the proteins that are known to be associated with the intracellular domains of TNF receptors have been implicated in the shape changes and secretory responses that characterize neutrophil activation.
In a recent study, we took a chemical-biologic approach to search for compounds that can selectively inhibit neutrophil activation by inflammatory mediators [24 ]. By means of high throughput screening of 15,000 "drug-like", small chemicals, we identified four compounds, 1–4, which inhibited the respiratory burst (RB) of human neutrophils in response to soluble inflammatory mediators, such as TNF and formylated methionyl-leucyl-phenylalanine (fMLF), but not in response to phorbol myristate acetate (PMA) and did not suppress the antibacterial activity of neutrophils. From the phenotypic pattern of their effects, the four compounds appeared to target different steps in the TNF signaling pathway. Compound 2 has been described [24 ]; based on its mechanism of action, 2 was termed neucalcin. Compounds 1, 3, and 4 appear to act at different sites in the signaling pathway of inflammatory mediators, as described here.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
H2O2 release and degranulation
As described [25
], the RB was measured in microtest plates coated with fetal bovine serum (FBS) by recording the horseradish peroxidase-dependent oxidation of fluorescent scopoletin by 1.5 x 104 neutrophils per well in response to buffer control, TNF (PreproTech, Rocky Hill, NJ) or PMA, each at 100 ng/ml. The supernate was used to measure degranulation as reported [24
]. Where indicated, neutrophils were incubated with dimethyl sulfoxide (DMSO) or test compounds for 30 min at 37°C and then exposed to Salmonella enterica var. Typhimurium (ATCC 14028s) or Listeria monocytogenes (ATCC 104035) opsonized with 10% autologous serum at a multiplicity of infection of 0.5 bacteria per neutrophil to measure H2O2 release.
Automated screening assay
The collection of 15,000 compounds and the screening were as described [24
]. In brief, using robots, human neutrophils and H2O2 assay mix were dispensed in microtest plates and preincubated with each compound individually for 30 min. Fluorescence was recorded, TNF added, and fluorescence recorded again 90 min later. Compounds affording >90% inhibition of the TNF-triggered RB were tested again using TNF and PMA as stimuli. Compounds showing inhibition of the RB when TNF was the trigger but not PMA were resupplied (ChemDiv, San Diego, CA) to confirm their identity and tested with cells from multiple normal donors in a concentration-response manner.
Bacterial killing
Neutrophils and bacteria were brought together for tests of the bacteria-triggered RB. At the indicated intervals, neutrophils were lysed with detergent, and surviving bacterial colonies were counted after overnight growth on agar plates [24
].
Additional studies
Methods for morphologic evaluation, transendothelial migration, immunoprecipitation, Western blot, and kinase assays were as reported [24
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
90% inhibition of H2O2 release triggered by TNF. Among the 460, 190 were selected for individual testing against TNF- and PMA-triggered H2O2 release. Potent inhibitors of PMA-triggered H2O2 release were discarded as being likely to intoxicate the cells, inhibit protein kinase C or phagocyte oxidase (phox), or interfere with the assay. After we screened the same compounds on neutrophils from multiple donors, four compounds emerged as reproducibly effective at inhibiting TNF-triggered H2O2 release at much lower concentrations than they inhibited the PMA-induced RB. The chemical structures of 1, 3, and 4 are illustrated in Figure 1A
; neucalcin (Compound 2) [24
] is included for comparison. Compounds 1 {2-[(2-methyl-5,6,7,8-tetrahydro-benzo[4
, 5
]thieno[2,3-d]pyrimidin-4-yl)-hydrazonomethyl]-phenol} and 3 {N-pyridin-3-ylmethylene-N'-(5,6,7,8-tetrahydro-benzo[4
, 5
]thieno[2,3-d]pyrimidin-4-yl)-hydrazine} are derived from the same core structure. Compound 5 {(3-morpholin-4-yl-propyl)-(5,6,7,8-tetrahydro-benzo[4
, 5
]thieno[2,3-d]pyrimidin-4-yl)-amine}, which had no effect on H2O2 release, shares the same core structure with 1 and 3 and was used as an inactive control in subsequent experiments. Dose-response curves (Fig. 1A)
yielded mean 50% inhibitory concentrations (IC50±SEM) as follows: 1, 0.18 ± 0.04 µM; 3, 2.5 ± 1.5 µM. However, we were not able to assign a consistent IC50 for 4 {(benzothiazol-2-ylsulfanyl)-acetic acid (2-hydroxy-5-nitro-benzylidene)-hydrazide}. In some experiments, 4 was exceedingly potent, with an IC50 of 3 nM. In others, the IC50 was 2 µM. The variation was donor-independent (data not shown) and remains unexplained. Concentrations that consistently yielded 95% inhibition of the TNF-triggered RB were used for all the subsequent experiments as follows: 2 µM for 1, 5 µM for 3, and 10 µM for 4. The control compound 5 was used at 5 µM. The results for 5, which have been reported previously [24
], were obtained in the same experiments in which 1, 3, and 4 were tested. Thus, results for 5 are reproduced here as the appropriate control group and as an aid in comparison. Inhibition by 1, 3, and 4 persisted after neutrophils were washed to remove these compounds from the medium, indicating slowly reversible or irreversible binding to their targets (Fig. 1B)
 View larger version (22K): [in a new window] |
Figure 1. Identification of compounds that inhibit the neutrophil RB in response to TNF. (A) Chemical structures and concentration-response curves. The indicated concentrations of each compound were added at 37°C for 30 min before stimulation with TNF (100 ng/ml, red squares) or PMA (100 ng/ml,  ). H2O2 release, measured at 90 min, is displayed as percent H2O2 release seen with TNF or PMA in the presence of the vehicle DMSO. Negligible amounts of H2O2 were released if no stimulus (NS) was added. Results are from one experiment representative of three and display mean ± SEM for triplicates. Many of the error bars fall within the symbols. (B) Reversibility. Neutrophils incubated with DMSO (D) or 1, 3, 4, or 5 were left unmanipulated (solid bars) or washed (open bars) with buffer before both sets were plated and stimulated with TNF. Other neutrophils from which the compounds had been washed off were re-exposed to the compounds (red hatched bars). Results are from one experiment representative of two and display mean ± SEM for triplicates.
|
 View larger version (34K): [in a new window] |
Figure 2. Impact on the interaction between neutrophils and bacteria. Neutrophils were incubated with DMSO (D), 1, 3, 4, or 5 for 30 min, stimulated with (A) fMLF (100 nM), (B) L. monocytogenes or S. enterica (bacteria:neutrophil ratio of 5:1) and H2O2 release, depicted at the time it reached plateau (75–130 min). (C) Effect on bacteria killing ability of neutrophils. Survival of bacteria in culture only (open bars) or in culture with neutrophils (solid bars). Killing of bacteria in the presence of DMSO or each compound is graphed after 30 min (for Salmonella) or 60 min (Listeria) of incubation with neutrophils. CFU, Colony-forming units. (A–C) Results are means ± SEM for triplicates in experiments representative of three, two, and three experiments, respectively. (D) Impact on degranulation. Neutrophils were incubated for 90 min as indicated, and their supernates were assayed for lactoferrin (LTF; LF), myeloperoxidase (MPO), and lactate dehydrogenase (LDH) as a percent of that released when the cells were stimulated with TNF. Each dot is the mean of duplicates in one experiment. Data are pooled from four experiments for LTF and two for MPO, each with different donors. Horizontal bars are group means. Effects of 3 and 4 on LTF release and of 1, 3, and 4 on MPO release were statistically significant (P<0.05 and P<0.01, respectively, by one-way ANOVA with Dunnett’s test).
|
Impact on neutrophil spreading and transmigration
The large-scale RB triggered by TNF or fMLF requires that neutrophils undergo cytoskeletal rearrangement and spread on ECM. Therefore, we asked whether the inhibitors interfered with these processes. All three compounds allowed neutrophils to spread normally on serum-coated glass in response to PMA. In contrast, each compound interfered with spreading triggered by TNF (Fig. 3A
). That spreading proceeded partially in the presence of each inhibitor suggested that the inhibitors acted after binding of TNF to its receptors and transmission of some signals to the cytoskeleton. These compounds seem to block different stages of cell spreading. Neutrophils treated with 1 showed a heterogeneous morphology upon TNF stimulation, ranging from appearing almost nonstimulated to spreading extensively, yet not as fully as in the absence of the compound. Cells treated with 3 were arrested after extending a few pseudopodia, and 4-treated cells after flattening and extending several filopodia (Fig. 3A)
. Compound 3 most consistently blocked cell spreading at the earliest stage. In contrast, 5 had no effect on cell spreading triggered by TNF.
 View larger version (77K): [in a new window] |
Figure 3. Impact of inhibitors on neutrophil spreading and transmigration. (A) Neutrophils were plated on FBS-coated glass coverslips, incubated or not with each compound at 37°C for 30 min before stimulation with TNF (100 ng/ml), PMA (100 ng/ml), or an equal volume of buffer (NS). After 30 min, the cells were fixed and photographed with phase-contrast microscopy (100x original magnification). Results are from one experiment representative of five. (B) Interaction of neutrophils with TNF-activated HUVEC (ECs) monolayers. Neutrophils were incubated with or without each compound at room temperature for 30 min, layered on collagen-supported HUVECs, which had not been stimulated (shaded bars) or had been activated for 12–18 h with TNF (50 pg/ml; solid bars), and allowed to transmigrate at 37°C for 30 min. Unbound cells were washed off with phosphate-buffered saline, and the cultures were fixed and stained. Neutrophils above and below the HUVEC were counted microscopically. The total number of neutrophils per field was scored as adherent (left panel). The proportion of the total that was beneath the HUVEC was scored as having undergone transendothelial migration (TEM; right panel). Means ± SEM for six replicates in one of three similar experiments, each with a different donor.
|
Each active compound blocked TNF-elicited spreading of neutrophils yet did not interfere with the cytoskeletal rearrangements necessary for adhesion and transmigration through HUVEC. This suggested that the compounds were not targeting components directly involved in cytoskeletal rearrangement. Rather, they might be interfering with upstream events that can be bypassed when signals for spreading are delivered through receptors other than those for TNF.
Influence on several tyrosine kinases involved in TNF-induced activation of adherent neutrophils
To explore the biochemical basis for the inhibition of TNF-induced responses, we investigated the effect of the compounds on three tyrosine kinases known to be critical in activation of neutrophils by TNF: Src [23
, 28
], Syk [29
, 30
], and Pyk2 [27
].
None of the compounds inhibited recombinant Src or Syk in vitro (Fig. 4A
and 4B
). However, the tyrosine phosphorylation of Syk, normally triggered within adherent neutrophils by TNF, was inhibited completely by 3 and 4 and 
50% by 1 (Fig. 4B)
. Thus, these compounds appear to block signaling events upstream of Syk activation rather than Syk itself. Phosphorylation of Pyk2 on tyrosine 402, which is critical for the kinase activity of Pyk2 [31
], was not blocked by any of the compounds to a greater extent than seen with the functionally inactive control Compound 5 (Fig. 4C)
. Thus, the degree of inhibition observed appears to be functionally irrelevant. In sum, the compounds do not appear to inhibit TNF-triggered H2O2 release by targeting Src, Syk, or Pyk2 directly, but they do appear to act upstream of Syk.
 View larger version (28K): [in a new window] |
Figure 4. Impact of inhibitors on biochemical events involved in TNF-induced activation of adherent neutrophils. (A) Src activity. Recombinant Src (40 units, UBI) was incubated for 30 min with DMSO (D), 1, 3, 4, 5, or a known Src kinase inhibitor, pp2 (10 µM, Calbiochem, San Diego, CA) and assayed for its ability over 10 min at 30°C to incorporate 32P from adenosine 5'-triphosphate into a synthetic substrate peptide from p34cdc2 purchased from UBI (Saranac Lake, NY). Background was measured by omitting Src (sub only). CPM, Counts per minute. (B) Syk activity. (Left) Syk was immunoprecipitated (IP) from adherent neutrophils treated as in A, and Western blot (WB) was carried out with antiphosphotyrosine antibody (PY; upper panel) or anti-Syk antibody (lower panel). (Right) Activity of recombinant Syk (60 ng, UBI) was measured as for Src. Migration of markers is indicated on the left by their molecular weight in kD. (C) TNF-induced phosphorylation of endogenous Pyk2. Total cell lysates (TCL) from adherent neutrophils, treated as indicated, were Western blotted with antibody specific for Pyk2, phosphorylated on tyrosine 402 (PY402; upper panel) and with antibody for total Pyk2 (lower panel). (A–C) Results are from experiments representative of two, three, and three experiments, respectively.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Production of reactive oxygen intermediates (ROI) by the RB plays critical roles in bacterial killing [32 , 33 ]. However, ROI can also promote tissue damage during inflammation by inactivation of antiproteases in the tissue [3 ]; desorbing proteases from their proteoglycan bed by triggering K+ influx [34 ]; activation of matrix metalloproteinases [35 ]; and increasing susceptibility of proteins to proteases by oxidation [1 ]. Thus, targeting ROI production in response to inflammatory mediators is expected to reduce proteolytic tissue damage through multiple mechanisms. Blocking ROI production in response to soluble mediators and leaving ROI production intact or only partially diminished in response to bacteria appear to leave neutrophils competent to kill the same bacteria.
The chemical-biologic approach used in this study generated reagents that may be helpful to dissect the complex cellular response of neutrophils during activation by inflammatory mediators. As the phenotypic patterns of inhibition by 1, 3, and 4 are different from one another and from that reported for neucalcin (Compound 2) [24 ], it is likely that the target of each compound is different (Table 1 ). Phox cannot be a direct target for any of the compounds, as they were selected on the basis that they do not inhibit phox, as long as phox is triggered by PMA. Possible sites of action include a flavoprotein, distinct from phox, whose early, low-level production of H2O2 may help induce activation of phox [36 ]; phosphatidylinositol 3-kinase, which plays a role early and late in TNF-dependent activation of phox in adherent human neutrophils [28 ]; Src homology 2 domain-containing leukocyte phosphoprotein of 76 kD, an adaptor protein whose deficiency creates a defect similar to that of Syk deficiency [30 ] with respect to TNF-triggered neutrophil activation in the mouse [37 ]; and Vav1, a guanine nucleotide exchange factor for Rac2, which also regulates the TNF- and adherence-dependent responses [38 ]. Further work to identify the molecular locus of action of each compound could provide a better understanding of the transcription- and translation-independent TNF signaling pathway in neutrophils, as well as the basis of response to fMLF. As fMLF signals through G protein-coupled receptors, and TNF does not, it will be of interest to identify signaling intermediates that these pathways acivate in common, besides Src, Syk, and Pyk2.
|
View this table: [in a new window] |
Table 1. Summary of Effects of Compounds on Neutrophil Functions
|
Received June 8, 2005; revised August 2, 2005; accepted September 10, 2005.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
/cachectin activates the O2(–)-generating system of human neutrophils independently of the hydrolysis of phosphoinositides and the release of arachidonic acid Biochem. Biophys. Res. Commun. 166,308-315[CrossRef][Medline]; priming effect of solid-phase fibronectin and intervention of CD11b–CD18 integrins Clin. Exp. Immunol. 94,533-538[Medline]
receptor and integrin signaling in neutrophils Immunity 19,761-769[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  F. J. Roca, I. Mulero, A. Lopez-Munoz, M. P. Sepulcre, S. A. Renshaw, J. Meseguer, and V. Mulero Evolution of the Inflammatory Response in Vertebrates: Fish TNF-{alpha} Is a Powerful Activator of Endothelial Cells but Hardly Activates Phagocytes J. Immunol., October 1, 2008; 181(7): 5071 - 5081. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
