| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* The Sackler Institute of Pulmonary Pharmacology, King’s College London, Guy’s Campus, London, UK; and
The School of Pharmacy, University of London, London, UK
1Correspondence: The Sackler Institute of Pulmonary Pharmacology, King’s College London, 5th Floor Hodgkin Bldg., Guy’s Campus, London SE1 1UL. E-mail: james.moffatt{at}kcl.ac.uk
ABSTRACT
Urocortins are members of the corticotropin-releasing factor (CRF) family of peptides that bind to two receptors, CRF1 and CRF2. While CRF1 is a high-affinity receptor for CRF, urocortin III binds with much greater affinity to CRF2. In the present study we investigated the effect of CRF2 receptor activation with urocortin III on airway smooth muscle tone in vitro and in an acute model of airway inflammation in mice. Urocortin III caused relaxation of methacholine-precontracted mouse tracheal segments. CRF caused similar relaxation, but with reduced potency compared to urocortin III, consistent with the CRF2 receptor subtype. Relaxation induced by urocortin III was concentration-dependently inhibited by the CRF2 antagonist, astressin 2B, with an IC50 in the nanomolar range. These relaxations were potentiated by inhibition of phosphodiesterases but unaffected by inhibition of cyclooxygenase and NO or by removal of the epithelium. Finally, the number of neutrophils retrieved by bronchoalveolar lavage after administration of bacterial LPS (LPS) was reduced by prior intraperitoneal (i.p.) injection of urocortin III. This effect was also suppressed by astressin 2B, implicating CRF2 receptors. Therefore, CRF2 agonists appear to have both bronchorelaxant and anti-inflammatory activities and might represent an interesting therapeutic approach to the treatment of inflammatory lung diseases.—Moffatt, J. D., Lever, R., Page, C. P. Activation of corticotropin-releasing factor receptor-2 (CRF2) causes bronchorelaxation and inhibits pulmonary inflammation in mice.
Key Words: urocortin • airway • neutrophil • smooth muscle • epithelium
CORTICOTROPIN-RELEASING FACTOR (CRF) is an important neuropeptide of the central nervous system (CNS) that regulates mood, gastrointestinal motility, and the immune response to stress (1
2
3)
. Released by hypothalamic neurons, CRF subsequently initiates release of ACTH into the circulation, the principal effect of which is to release anti-inflammatory glucocorticocoids from the adrenal gland. In addition to this role in the hypothalamic-pituitary-adrenal axis, centrally acting CRF also acts to produce anxiety and depression-like behaviors. Since the initial discovery of CRF, three peptides (the urocortins) with close homology have been cloned using molecular biology approaches (2)
. The physiological roles of these novel peptides are under investigation. In the CNS, urocortins and CRF have different anatomical distributions, suggesting potentially different physiological roles.
CRF and urocortins act at two G-protein-coupled receptors, CRF1 and CRF2. CRF binds with high affinity to CRF1 whereas urocortin I binds to both receptors. However, urocortin II and urocortin III are high-affinity agonists for CRF2. Similar to the different localization of CRF and urocortins, CRF1 and CRF2 are found in different regions of the CNS, again suggesting different physiological roles for these peptides. These structural observations are matched by behavioral studies that indicate that while CRF1 receptor activation exerts central and peripheral effects concordant with stress, CRF2 activation has opposite effect such as decreased feeding, lowered anxiety/depression, vasodilatation, and lowered blood pressure (1)
.
While the functions of CRF and urocortins in the CNS are becoming clear, much less is known about the roles of these peptides in the periphery. Urocortins are widely expressed in peripheral organs such as the spleen, thymus, skin, kidney, and heart, as well as in inflammatory cells (1)
, and therefore may be important paracrine or autocrine mediators. Indeed, peripheral actions of CRF and urocortins have been reported. For example, urocortins potently relax vascular smooth muscle in vitro (4)
and modulate gastrointestinal function in isolated preparations (5)
. As in the CNS, activation of CRF1 and CRF2 appears to have opposite effects. Thus, in isolated preparations of rat duodenum, CRF1 activation increases smooth muscle activity, whereas in segments of ileum CRF2 activation is inhibitory (5)
. To date, no study has examined the possible bronchomotor effects of these emerging mediators.
Until recently, roles for CRF and related peptides in inflammatory processes had been obscure because it was not realized that two receptors, with possibly opposite roles, are present (1)
. Apart from its central role in indirectly initiating release of anti-inflammatory glucocorticosteroids, CRF may have a direct peripheral proinflammatory role (1)
. For example, local knockdown of CRF (CRF1-preferring ligand) with RNA interference in rats inhibits C. difficile toxin A-induced intestinal inflammation (6)
. In the same study, a similar reduction of urocortin II (CRF2-preferring ligand) had no effect, suggesting no role for CRF2 in intestinal inflammation. However, urocortin II administration converts a sublethal model of bacterial infection in mice into a lethal model by inducing the production of anti-inflammatory cytokines and thus lowering host resistance (7)
. However, whether selective activation of CRF1 or CRF2 modulates inflammatory processes in the lung has not been explored.
In the present study, we have investigated for the first time the ability of a CRF2-selective agonist, urocortin III, to alter smooth muscle tone and to modulate pulmonary neutrophilia in response to intranasally administered LPS.
MATERIALS AND METHODS
Animals
All experiments were carried out in accordance with the UK Home Office Animals (Scientific Procedures) Act, 1986. Female BALB/c mice (6–8 wk, Charles River Laboratories, Margate, UK) were studied. For in vitro studies, mice were killed with an i.p. injection of urethane (20 g/kg) and the full length of the trachea was removed and placed in a physiological salt solution (composition (mM):NaCl 118; KCl 5.4; MgSO4 0.57; glucose (Glc) 11; KH2PO4 1.2, NaHCO3, 25; CaCl2 2.5). After clearing the trachea of surrounding connective tissue the preparation was either cut into two rings (for mechanical recording) or cut along the ventral surface to produce a single flat sheet (for electrical recordings of ion transport). In some experiments, we attempted to remove the epithelium from the trachea by lavaging a mild detergent solution (0.1% Triton X-100) through the trachea in situ, prior to excision, as described (8
, 9)
. Full removal of the epithelium was confirmed by lack of responsiveness to ATP, a known epithelium-dependent relaxant in this preparation (10)
.
Isometric recording of airway smooth muscle
Preparations of trachea were mounted in 10 ml organ baths on two L-shaped stainless steel supports as described previously (9)
. After 15 min equilibration at 37°C the preparations were contracted with a supramaximal (100 µM) concentration of the cholinergic agonist methacholine to determine the maximal contractile force (Fmax) capacity of each preparation. After washout of methacholine, the preparations were allowed to re-equilibrate for 30 min, during which time inhibitors were added to the baths. Each preparation was carefully contracted to 
50% Fmax with titrated concentrations of methacholine, then exposed to cumulative concentrations of urocortin III or CRF. The full capacity of the tissues to relax was determined at the end of the experiment by the addition of a supramaximal concentration of the nonselective ß-adrenoreceptor agonist isoprenaline (10 µM). In experiments using inhibitors, one ring from each animal served as a control preparation.
Recording of epithelial ion transport
Sheets of trachea, prepared as described above, were mounted in an Ussing chamber (World Precision Instruments, Stevenage, UK) as described previously (11)
. Voltage and current electrodes on either side of the preparation were connected to an amplifier and the tissue was voltage clamped at 0 mV for the duration of the experiment. Changes in required short-circuit current (ISC) were used as an indication of changes in ion transport.
LPS-induced airway inflammation
Mice were administered 50 µl of a 0.05 mg/ml solution of LPS in saline via the intranasal route while lightly anesthetized with 5% isofluorane. To assess the effect of urocortin III on the influx of neutrophils that occurs after LPS administration, urocortin III (or saline) was given at a dose of 150 µg/kg (i.p.) 1 h prior to LPS. To verify that the effect of urocortin III was mediated via CRF2 receptors, another group of mice was given the CRF2 antagonist astressin 2B (12)
at a dose of 500 µg/kg (or saline) 1 h prior to urocortin III administration. All mice were killed (as described above) 3 h after administration of LPS and bronchoalveolar lavage (BAL) was performed as described previously (13)
. Total cell counts and differential counts of stained cytospin preparations were performed to determine the absolute numbers of neutrophils recovered in BAL fluids. BAL TNF-
concentrations were determined using capture and biotinylated detection antibodies (R&D Systems, Abingdon, Oxon, UK) following the manufacturer’s instructions.
Immunohistochemistry
Four mice, killed as described above, were perfused transcardially with saline for 2 min before similar perfusion with a solution of paraformadehyde (4% in PBS) at 4°C. Tracheal tissues were dissected from the animal and fixed in the same solution for 1 h at 4°C before several 30 min washes in PBS. Tissues were cryopreserved in 10% sucrose in PBS overnight at 4°C, then frozen in OCT compound. Frozen sections (10–14 µm) were allowed to air dry for 30 min, then washed twice in PBS. The sections were then incubated overnight in a humid chamber with rabbit antisera raised against CRF1 or CRF2 (Santa Cruz Biotechnology, Calne, Wiltshire, UK) diluted 1:50 in 0.1% BSA and 0.1% Triton X-100 in PBS. To clearly delineate the smooth muscle layer, sections were simultaneously incubated with an anti-smooth muscle actin antibody (Ab) (Sigma, 1:500). CRF1 or CRF2 were visualized with a CY3-conjugated goat anti-rabbit secondary Ab (Jackson Immunoresearch, Soham, Cambridgeshire, UK), while smooth muscle actin was detected using an FITC-conjugated donkey anti-mouse Ab (Jackson Immunoresearch). The slides were examined with a Zeiss Axioskop microscope, to which a CCD camera was attached to capture digital images.
Sources of materials
Isoprenaline, IBMX, methacholine, indomethacin, LPS, L-NAME, astressin 2B, LPS (E. coli 0127:B8), and urethane were purchased from Sigma (Poole, Dorset, UK). Urocortin III and CRF were purchased from Bachem (St. Helens, UK). Other reagents were purchased as indicated above.
RESULTS
Localization of CRF1 and CRF2
Using commercial antibodies raised against CRF1 and CRF2, we employed fluorescence immunohistochemistry to probe the location of these receptors in the mouse trachea. Both antibodies revealed intense staining of the airway epithelium (Fig. 1
A, B) with little apparent immunoreactivity in other cells. Colocalization of CRF1 with smooth muscle actin was not observed in sections from any of four animals. However, faint CRF2-immunoreactivity was consistently observed in the smooth muscle layer. Adventitial fibroblast-like cells displayed both CRF1 and CRF2 immunoreactivity, as did chondrocytes near the point of insertion of smooth muscle (not shown). Only very faint or no staining of any these structures was observed when the Ab was preadsorbed with the immunizing peptide antigen (not shown).
|
Effects of urocortin III and CRF in vitro
The CRF2-preferring ligand urocortin III (1
2
3)
caused slowly developing relaxation of smooth muscle (Fig. 1C
), which represented an approximate 30% reversal of methacholine-induced tone. Isoprenaline caused prompt and full relaxation of the tissues, demonstrating the full reversibility of mechacholine-induced relaxations (Fig. 1C
). Similar to urocortin III, CRF caused relaxation of the mouse trachea (Fig. 2
A), although it was significantly less potent than urocortin III (pEC50 6.79±0.08 vs. 7.153±0.09; P=0.02).
|
As CRF1 and CRF2 staining were most pronounced on epithelial cells, we also examined the effects of CRF1-preferring (CRF) and CRF2-preferring (urocortin III) peptides on ion transport across the airway epithelium in voltage clamped preparations. Surprisingly, neither CRF (not shown) nor urocortin III (Fig. 1D
) caused significant changes in ion transport, although the effect of elevating intracellular cyclic nucleotide levels with the phosphodiesterase inhibitor isobutylmethylxanthine (IBMX; 100 µM) produced large responses (Fig. 1D
).
The highly selective CRF2 receptor antagonist (>100 fold vs. CRF1) astressin 2B (12)
potently inhibited relaxations induced by urocortin III (Fig. 2B
). Although a full pharmacological analysis (to determine a pA2 value) was not possible owing to availability and solubility of urocortin III, the IC50 for astressin 2B appeared to be 3 nM (Fig. 2B
), consistent with the nanomolar affinity of astressin 2B for CRF2.
Like ß-adrenoceptors, CRF2 has repeatedly been reported to couple to Gs and hence elevate cAMP in cells (1
2
3)
. If similar coupling occurs in the mouse trachea, responses should be amplified by agents that inhibit phosphodiesterases. In agreement with this, IBMX (30 µM), a nonspecific phosphodiesterase inhibitor, significantly potentiated relaxation responses to urocortin III (Fig. 3
A).
|
The strong CRF1 and CRF2 immunoreactivity in airway epithelial cells also prompted us to consider that the response to urocortin III might depend on an epithelium-derived relaxing factor such as NO or prostaglandin E2, as has been demonstrated for other agonists (8
, 14)
. However, the combination of inhibitors of cyclooxygenase and NOS had no effect on the relaxation caused by urocortin III (Fig. 3B
), suggesting that the effect of this peptide is directly on smooth muscle cells in the trachea. Furthermore, in three cases in which removal of the epithelium was confirmed by a lack of relaxant effect of ATP (10)
, urocortin III produced similar maximal responses (Fig. 3B
).
Effect of urocortin III on LPS-induced pulmonary neutrophilia in vivo
Finally, we tested the potential anti-inflammatory activity of urocortin III in a standard acute model of lung inflammation induced by LPS, which elicits neutrophil influx into the airways of mice. Few to no neutrophils were found in BAL fluids from naive BALB/c mice in our hands, whereas LPS elicits neutrophil influx within 3 h. Pretreatment of mice with as little as 150 µg/kg i.p. of urocortin III reduced the number of neutrophils found in BAL 3 h after administration of LPS by more than half compared to saline (Fig. 4
A). This anti-inflammatory effect was suppressed when astressin 2B was administered (500 µg/kg i.p.) 1 h prior to urocortin III (Fig. 4)
, confirming the involvement of CRF2 receptors. It is known that the recruitment of neutrophils to the lungs of mice after LPS administration is dependent on production of TNF- in the first 3 h within the airways (15)
. However, TNF- levels, which were undetectable in naive animals (not shown), as determined in BAL samples, were not different between the experimental groups (Fig. 4B
). In a final set of experiments, we examined whether the anti-inflammatory effect of acute CRF2 activation was profound enough to inhibit airway neutrophilia for up to 24 h after LPS administration. The number of neutrophils recovered 24 h post-LPS was reduced, but not significantly (137±22x104 vs. 210±46x104 in saline controls, n=5), unless a second dose of urocortin III was administered after 6 h (183±11x104 vs. 140±14x104, n=5).
|
DISCUSSION
Activation of smooth muscle CRF2 receptors has been shown to cause relaxation in vascular preparations (4)
, and CRF2 has been shown to inhibit ileal motility (5)
. Since CRF1 and CRF2 are linked to cAMP production via coupling to Gs (1
2
3)
, we hypothesized that, like the similarly coupled ß-adrenoceptors, CRF2 activation would have a bronchorelaxant effect. Indeed, our data fully support a direct role of airway smooth muscle CRF2 receptors in causing airway relaxation. Thus, urocortin III was more potent than CRF, and the inhibitory responses to urocortin III were dose-dependently inhibited by astressin 2B in the nanomolar range. Both of these observations implicate CRF2 rather than CRF1 receptors or an unidentified receptor. Furthermore, our immunohistochemical studies found no evidence for CRF1 receptors on smooth muscle, whereas CRF2-immunoreactivity, albeit weak, was clear in all animals studied.
As both CRF1 and CRF2 were detected at very high levels on the epithelium and the epithelium is known to release inhibitory substances, we considered the possibility that the relaxant responses to urocortin III were mediated via epithelium-derived products. In mice, the predominant epithelium-derived inhibitory factor is prostaglandin E2 (8
, 16)
. However, inhibition of cyclooxygenase with indomethacin did not blunt the response to urocortin III (data not shown). In guinea pigs, both prostaglandin E2 and NO appear to be released by the epithelium (14)
. Therefore, we combined inhibitors of cyclooxygenase and NOS (L-NAME) in order to rule out a role for either of these epithelium-derived factors indirectly mediating the response to urocortin III. The lack of effect of this combination of inhibitors, combined with the lack of effect of removing the epithelium itself, unequivocally indicate an action of urocortin III directly on airway smooth muscle cells.
Because the level of expression of CRF1 and CRF2 was so high in epithelial cells compared with CRF2 expression on smooth muscle cells, we hypothesized that epithelial cell receptors might mediate robust biological responses. One property of the epithelium that is modulated by G-protein-coupled receptors is the rate of transport of ions across this layer, a process critical to maintaining the sol layer in the airways for the efficient transport of mucus. Elevating intracellular cAMP in airway epithelial cells causes the opening of cAMP-sensitive channels that increase the net flux of ions across the epithelium (17)
, causing a change in the ISC required to voltage clamp an isolated preparation. Since CRF1 and CRF2 are coupled to Gs, we expected activation of these receptors to cause such a change. We were surprised to find that this was not the case, even though elevation of cyclic nucleotide levels with IBMX produced a robust response. Therefore, levels of protein expression of these receptors do not appear to predict a biological response. It is noteworthy that we have also found very high levels of CRF1 and CRF2 expression on platelets, yet neither CRF nor urocortin III inhibit platelet aggregation (data not shown), as would be predicted for Gs-coupled receptors (18)
. We are currently exploring whether such CRF receptor pools become more effectively coupled following inflammatory or other stimuli.
Roles for CRF and urocortins in inflammatory pathways have been difficult to dissect without selective inhibitors of the two known receptors (1)
. Thus, until more recently, reports of the effects of these peptides in inflammatory conditions have been somewhat contradictory. This is probably due in part to the fact that CRF1 and CRF2 receptors often have opposite effects and peptides like urocortin have little selectivity. By selectively lowering levels of CRF1 and CRF2 mRNA, la Fleur et al. (6)
recently showed that CRF1 is important in a rat intestinal inflammation model, whereas lowered expression of CRF2 had little effect. On the other hand, administration of the CRF2-preferring ligand urocortin II to mice infected with bacteria lowered host resistance by inducing the secretion of the anti-inflammatory cytokine interleukin (IL)-10 (7)
. This effect was inhibited by the CRF2-selective antagonist astressin 2B. Thus, the emerging pattern for CRF receptors in inflammation is that CRF1 appears to be important to the development of an inflammatory response, whereas CRF2 has no, or an anti-inflammatory, role. In our model of LPS-induced airway inflammation, urocortin III potently inhibited neutrophil influx into the lungs without altering the levels of the initial inflammatory signal (i.e., TNF-) in the lungs. It is known that TNF- production in the lung is a crucial early event in response to LPS (15)
, and so our findings suggest that urocortin III somehow exerts a systemic anti-inflammatory effect that causes the immune system to ignore the inflammatory signal from the lung. This effect probably does not involve remote (e.g., splenic) secretion of IL-10 (7)
, as in a very similar model of LPS-induced lung inflammation, systemically delivered IL-10 was ineffective in reducing lung inflammation in response to inhaled LPS, even when produced continuously after administration of an IL-10-encoding plasmid (19)
. Similarly, although urocortin has been found to inhibit systemic TNF- production in vivo via an uncharacterized CRF receptor (20)
, levels of this cytokine in BAL fluids were identical in all experimental groups in our hands, and TNF- production is limited to the airway compartment when LPS is given by inhalation (21
, 22)
. The anti-inflammatory effect of urocortin III in our model was more apparent at the acute 3 h time point compared to the smaller effect observed at 24 h. Based on this observation and the lack of effect on inflammatory cytokine release, we consider it likely that urocortin has short-term effects on endothelial cell or leukocyte function. We are currently investigating these possibilities.
Another unresolved issue is whether peripheral CRF2 activation occurs in a normal inflammatory response or whether systemic administration of urocortin III is completely nonphysiological. Urocortin III is not found in naive rodent lungs (23)
, whereas only a low level of expression has been reported in human lung tissue (24)
. Therefore, it is possible that our data do not reflect any endogenous protective role for urocortin III in the lungs. However, urocortin II is present at high levels in peripheral blood leukocytes (24)
, which invade the airways during an inflammatory response. In our hands, using the LPS-induced inflammation model reported, we find no difference between the numbers of neutrophils found in BAL after pretreatment of mice with astressin 2B in the absence of exogenous urocortin III (data not shown), suggesting that endogenous CRF2 activation is not protective in this model. Further experiments are required to determine whether endogenous urocortin III is protective in other pulmonary disease models. Irrespective of whether our observations are relevant in terms of disease initiation and development, our findings point toward a novel strategy for the development of anti-inflammatory drugs having moderate bronchodilator activity.
In summary, we have demonstrated for the first time that activation of the urocortin II- and III-preferring CRF2 receptor subtype causes bronchorelaxation in vitro and inhibits inflammatory processes in the airway in vivo. Therefore, we suggest that CRF2 receptors might be an interesting therapeutic target for future treatment of inflammatory airway diseases.
ACKNOWLEDGMENTS
We thank Janice Cheung for her technical assistance.
Received for publication December 1, 2005. Accepted for publication April 21, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
