Role of macrophage migration inhibitory factor (MIF) in allergic and endotoxin-induced airway inflammation in mice.

Macrophage migration inhibitory factor (MIF) has recently been forwarded as a critical regulator of inflammatory conditions, and it has been hypothesized that MIF may have a role in the pathogenesis of asthma and chronic obstructive pulmonary disease (COPD). Hence, we examined effects of MIF immunoneutralization on the development of allergen-induced eosinophilic inflammation as well as on lipopolysaccharide (LPS)-induced neutrophilic inflammation in lungs of mice. Anti-MIF serum validated with respect to MIF neutralizing capacity or normal rabbit serum (NRS) was administered i.p. repeatedly during allergen aerosol exposure of ovalbumin (OVA)-immunized mice in an established model of allergic asthma, or once before instillation of a minimal dose of LPS into the airways of mice, a tentative model of COPD. Anti-MIF treatment did not affect the induced lung tissue eosinophilia or the cellular composition of bronchoalveolar lavage fluid (BALF) in the asthma model. Likewise, anti-MIF treatment did not affect the LPS-induced neutrophilia in lung tissue, BALF, or blood, nor did it reduce BALF levels of tumor necrosis factor-alpha (TNF-alpha) and macrophage inflammatory protein-1alpha (MIP-1alpha). The present data suggest that MIF is not critically important for allergen-induced eosinophilic, and LPS-induced neutrophilic responses in lungs of mice. These findings do not support a role of MIF inhibition in the treatment of inflammatory respiratory diseases.


Introduction
Macrophage migration inhibitory fac tor (MIF) is considere d to be a c ritical re gulator of various inflammatory conditions. For ex ample, MIF is a pivotal me diator in the host re sponse to endotox ic shock, 1 and plays an important role in the de ve lopme nt of the delayed type hyp erse nsitivity reaction and c ollagen-induce d arthritis in mice . 2,3 MIF may further have a proinflammatory role in the deve lopme nt of human ac ute respiratory distress syndrome . 4 MIF w as described originally to be a T-ce ll produc t, w hich inhibited the random migration of mac rophage s. 5 ,6 Inte restingly, rec ent data indic ate that MIF is pre dominantly ex pre sse d by Th2-like T-c ells. 7 Howe ve r, MIF also ex is ts pre formed in monocyte s /mac rophage s, eosinophils, B-cells, airw ay epithelial cells, and c ortic otrophic c ells w ithin the anterior pituitary gland, and is release d in resp onse to various proinflammatory stimuli. 1 ,8 -1 2 MIF shares w ith other cytokines sensitivity to inhibition by the rapeutic conc entrations of gluc oc ortic oids. How eve r, at low physiological glucoc ortic oid conce ntrations MIF ex pre ssion is induc ed. 13 ,1 4 MIF has the prope rty of counte rac ting anti-inflammatory and immunosuppressive actions of glucocorticoids. 13 ,1 4 Anti-MIF therapeutic strategie s are thus under deve lopme nt w ith the aim to increase the immunosuppre ssive and antiinflammatory propertie s of endogenously release d gluc ocorticoids, thereby reducing the requireme nt for ste roid therapy in a variety of inflammatory conditions. It has furthe r bee n suggeste d that drugs inhibiting MIF w ould be e ffec tive also in inflammatory c onditions that ex hibit ste roid re sistanc e. 14 Airw ay mucosal inflammation in allergic asthma is charac te rize d by infiltration and /or activation of eosinophils, macrophages, T-lymphocyte s, and mast ce lls. 15 Sinc e all the se immune c ells c an produc e MIF in significant quantities, and since e levated le vels of MIF have bee n detected in BALF from asthmatic patients c ompare d w ith controls, it has be en forw arded that MIF may have a role in eosinophilic airw ay disease . 9 Re ce nt studies have show n that genetic ally MIFdefic ie nt mice are resistant to the lethal e ffec ts of a high syste mic dose of LPS, but susceptible to a combination of a low -dose of LPS and D-galac tosamine. 1 6 These mic e ex hibit no impairme nt of neutrophil migration to the p eritoneum elicited by thioglyc ollate; de spite diminished ne utrop hil acc umulation in BALF the y clear Ps e u d o m o n a s a e ru g ino s a instille d into the trachea better than w ild type mic e. 1 6 Makita e t a l. 11 re ce ntly demonstrate d that immunone utralization of MIF atte nuate s pulmonary ne utrophil influx and acute lung injury induc ed by intraperitone al (i.p.) administration of LPS in rats. These e ffec ts w e re assoc iated w ith re duce d BALF levels of mac rophage inflammatory protein -2, a pow e rful neutrophil chemokine. 1 1 Moreover, it has bee n de monstrated that MIF induce s alveolar c ells / macrophage s to se crete TNF-a and IL-8, 4,17 tw o cytokines w ide ly thought to be c ritically important for neutrophil infiltration in pulmonary inflammatory conditions such as chronic obstructive pulmonary disease (COPD). 18 , 1 9 In an atte mpt to further ex plore the hypothesis that MIF may have a role in bronchopulmonary eosinophilic and ne utrophilic inflammation, w e ex amine d effects of MIF inhibition both in an e stablishe d murine mode l of allergic asthma, and in a mode l involving LPS-induc ed ne utrop hilic inf lammation in the lungs of mic e. Spec ifically, anti-MIF serum w as administe re d repeate dly during allerge n aerosol ex p osure of immunized mic e, or give n once be fore instillation of a low dose of LPS into the low er airw ays of mic e.

Animals
Male C57BL/6 mice (n = 184, 6-8 w e eks of age ), w ere purchased from Bomholtgaard, De nmark. All mic e w ere kept in w ell-controlled animal housing fac ilitie s and had fre e acc ess to tap w ater and pe lle te d food throughout the ex pe rimental p eriod. All animals w ere used unde r protocols approve d by the Ethics Committe e of the Faculty of Medicine at the University of Lund.

Induction of allergic eosinophil-rich airway inflammation
We have use d a protocol slightly modified from that developed by Brusse lle and colleagues. 2 0 On the first day of the ex perime nt (Day 0 ), all mic e w ere ac tively immunize d by i.p. injec tion of 10 m g chic ken OVA (Grade III, Sigma, St Louis, MO, USA), adsorbed to 1 mg of alum adjuvant. Starting 14-16 days afte r immunization the mice w ere ex pose d once daily during 7 days to aerosolized saline (SAL) or OVA over a 30-min period by plac ing groups of 5 -10 aw ake mice in an ex position chamber. The ae rosol w as gene rate d into the chamber using a nebulize r (Bird 500 ml Inline Mic ronebulizer driven at 4 bar, Bird Co., Palm Springs, CA ). The conc entration of OVA in the ne bulizer w as 1% w /v. Animals w e re sac rifice d by i.p inje ction of pentobarbital 8 h after the last ae rosol ex p osure.

Induction of neutrophil-rich airway inflammation
Groups of mice rece ive d one intratracheal instillation of a low dose of Es ch e rich ia c o li LPS (Difc o Lab., MI, USA, 4 m g /kg, i.e. , 0.08 m g /animal, diluted in , 20 m l saline ) or SAL. The present LPS dose is comparable w ith occupational le ve ls. 21 For ex amp le , it has be en estimated that c otton mill w orkers are ex posed to 60 m g endotox in pe r day. 2 2 In pre liminary dosere sponse ex perime nts our sele cte d dose of LPS w as show n to induce submax imal re sponse s regarding TNF-a leve ls and total ce ll numbers in BALF. For the instillation proce dure, animals w e re anaesthe tize d w ith enflurane , and a blunted c annula w as introduce d perorally into the trache a. Animals w ere sacrific ed by i.p injec tion of p entobarbital 4 or 24 h after LPS or SAL administration.

In vivo neutralization of MIF
Starting the day before the first allergen aerosol challe nge, mice (n = 20 ) w ere injec te d i.p. w ith 200 m l of rabbit anti-murine MIF se rum; this treatme nt w as then repe ated e very 3 days until te rmination of the ex pe riment. Control mice (n = 20 ) w ere injec te d w ith a similar volume (200 m l) of NRS. Other groups rec eived no tre atment at all. Groups of mice subje cte d to intratra cheal LPS instillation rec eived one i.p. injection of anti-MIF serum (200 m l), NRS, or SAL 12 h before LPS challe nge (n = 8 in e ach group ). This anti-MIF treatme nt has previously been show n to neutralize MIF in v ivo using the same dose leve l and administration route in mic e as in this study. 1,2 3 Components in se rum may have some c apacity to dow nregulate inflammatory response s, unde rscoring the need for proper control groups in the evaluation of ex perime nts using anti-se rum. Thus, in this study groups of mic e treate d w ith anti-MIF serum w ere compare d w ith corresp onding groups of mice treate d w ith NRS.
In additional ex pe rime nts w e ex plore d w he ther anti-MIF tre atment might be effe ctive via the local route. Groups of mic e (n = 4 in e ach group ) re ce ive d 15 m l anti-MIF serum or NRS administe re d intratracheally, alone or toge the r w ith LPS solution (4 m g /kg ) in a total volume of , 20 m l.
The pre sent batch of anti-MIF serum w as che cke d for bioactivity at our laboratory using a protoc ol, w hich has pre viously been employe d to analyse the role of MIF in endotox ae mia. 16 Groups of mice (n = 5-6 in each group ) re ce ive d one i.p . inje ction of anti-MIF se rum (200 m l), NRS, or SAL 2 h before i.p. injection w ith a high dose of LPS (25 mg /kg ). Ninety minute s after LPS challe nge blood w as collected by cardiac p uncture and plac ed into EDTA tubes. As an indicator of LPS-induce d p lasma ex travasation, the haematoc rit w as de te rmined by an automate d haematology analyser (Sysmex K-4500, TOA Medical Elec tronics Co., Kobe, Japan ).

Histological analysis of allergic airway inflammation
Lung tissue spec imens obtained 8 h afte r the last OVA or SAL ex p osure w e re imme rsed ove rnight in Stefanini's fix ative (2% p araformaldehyde and 0.2% picric ac id in 0.1 M phosphate buffer, pH 7.2 ), rinse d repeate dly in buffer (Tyrode's buffe r supp le mented w ith 10% sucrose ), froze n in mounting me dium (Tis sue-Te k, Mile s Inc , Elkhart, IN, USA), and stored at -80°C until se ctioning. Eosinophils w e re de te cte d by histochemical visualization of cyanide-re sistant e osinophil peroxidase (EPO) ac tivity. 2 4 -2 6 Briefly, cryose ctions (10 m m ) w ere incubated for 8 min at room tempe rature in PBS buffer (pH 7.4 ) suppleme nte d w ith 3,3-diaminobe nzidine te trahydrochloride (60 mg /100 ml, Sigma ), 30% H 2 O 2 (0.3 ml/100 ml), and NaCN (120 mg /100 ml). Slide s w ere then rinsed in w ate r and mounte d in Kaise r's me dium (Merck, Darmstadt, Ge rmany ). Eosinophils w ere identifie d by the ir dark-brow n reaction product. For e valuation of the number of eosinophils in p ulmonary tissue , 40 randomly selec te d areas (0.04 mm 2 each ) in one lung section from e ach animal w ere ex amined. The number of eosinophils in the 40 are as w as counte d at a magnification of 400, and the me an w as ex pressed as eosinophils /unit area. Ce ll counts w e re made in a blinde d fashion. For asse ssme nt of gene ral airw ay morphology sections w ere staine d w ith hae matox ylin and e rythrosin. Lung tissue spe cime ns and tracheobronchial lymph node s imme rsed in buffere d 4% p araformaldehyde (pH 7.2 ), dehydrate d, and embe dded in paraffin, w ere used for immunohistoche mical vis ualization of MIF-ex pre ssing ce lls. Se ctions w ere incubated overnight in 4°C in a mois t chamber w ith a 1: 800 dilution of the rabbit antimurine MIF se rum. The anti-MIF se rum use d in the present study has pre viously bee n use d for immunohistoche mistry. 1,2 NRS at a dilution of 1:800 or PBS w ere used in control se ctions. The site of the antigenantibody re ac tion w as re veale d by application of fluoresce in is othioc yanate-c onjugated sw ine antise ra dire cted at rabbit immunoglobulins (DAKO, Glostrup, De nmark ) dilute d 1: 80 for 1 h at room te mperature .

Histological analysis of LPS-induced airway inflammation
Lung tissue specimens obtained 4 h and 24 h after LPS or SAL challenge w ere imme rsed in buffere d 4% paraformalde hyde (pH 7.2 ), dehydrate d, and embe dded in paraffin. One sec tion (6 m m ) per animal (n = 3 from e ach group ) w ere stained w ith haematox ylin and erythrosin to ex amine the ex te nt of pulmonary inflammation. Othe r se ctions (one pe r animal, n = 3 from e ach group ) w ere use d for immunohistochemic al demonstration of neutrophils in lung tissue. Se ctions w ere incubated ove rnight in 4°C in a moist chamber w ith a 1: 200 dilution of rabbit antise ra direc te d at human myeloperox idase (MPO, DAKO, Glostrup, De nmark ). The site of the antige n-antibody reaction w as re ve aled by applic ation of fluoresc ein is othiocyanate-c onjugated sw ine antisera direc te d at rabbit immunoglobulins (DAKO, Glostrup, De nmark ) for 1 h at room te mperature. In c ontrol sec tions, omitting the primary antibody, only slight ye llow ish auto-fluoresce nce w as found.

Analysis of cells in peripheral blood and BALF, and measurement of TNF-a and MIP-1 a levels in BALF and plasma
Animals w e re anaesthe tize d w ith an i.p. injection of pentobarbital. The chest w as opene d and a blood sample w as c ollec te d via the still beating heart. A tracheal c annula w as inserte d via a midce rvic al inc ision. The airw ays w ere lavaged once (LPS-challenge d mice ) or tw ic e (OVA-ex posed mic e ) w ith 0.7 ml of PBS (Life Te chnologie s, Paisle y, UK). The BALF w as imme diately c entrifuged (10 min, 4°C, 160 ´g ). Ce ll pellets w ere resuspe nded in 250 m l PBS for total and differential cell c ounting and the supernatants w ere rapidly frozen. Differential counting w as performed on May-Grünw ald-Giemsa stained c ytospins and blood sme ars. Be tw e en 200 and 500 c ells w ere c ounted on each cytospin, and 100 cells w e re c ounted on each blood smear. Comme rcial ELISA kits (R&D syste ms, MN, USA) w ere used to me asure levels of TNF-a and MIP-1 a in the BALF of LPS and SAL challenge d mic e. TNF-a le vels w ere also me asure d in plasma obtaine d from LPS challenged mic e. The limit of dete ction w as 5.1 pg /ml for TNF-a , and 1.5 pg /ml for MIP-1 a .

Statistics
Data are ex p re sse d as mean ± SEM. To calculate significance le vels betw e en treatment groups, the Student's t-te st w as use d throughout the study. ELISA values below detection limits w e re assigned the value of the de te c tion limit. Probabilitie s < 0.05 w ere used as the generally acc epte d level of statistic al signific anc e for diffe renc es betw ee n me an value s.

Additional validation of the present anti-MIF sera
Ex posure of a high dose of endotox in is know n to cause plasma ex travasation and subse que nt loss of circ ulating plasma volume . 27 ,28 In order to asce rtain a preserved activity of the anti-MIF serum, the hae matocrit w as determine d 90 min afte r i.p. LPS challenge as a measure of plasma ex travasation. The haematocrit w as significantly incre ased in LPS-challe nged mic e compared w ith SAL-challenged mice (53.1 ± 4.0% vs. 45.2 ± 6.2%, P< 0.001 ). This response w as inhibite d in mic e treate d w ith anti-MIF serum before LPS challenge (45.7 ± 2.1% vs.50.8 ± 5.0% in c orresponding NRS-treate d mic e, P< 0.01 ).

Effect of anti-MIF treatment on allergen-induced airway inflammation
To assess the role of MIF in alle rgic airw ay inflammation, lung tissue eosinophilia and ce llular composition of BALF w ere determine d in anti-MIF-tre ated and NRStre ate d mic e 8 h after last alle rge n ae rosol ex posure. The numbe r of eosinophils in lung tissue w as similar in allerge n ae rosol ex posed anti-MIF-treated and NRStre ate d mice (Fig. 1 ). Likew ise, total ce llular conte nt (data not show n ) and the perc entage of eosinophils, ne utrophils, lymphocyte s, and mac rophages in BALF did not diffe r significantly betw e en anti-MIF-tre ate d and NRS-treated mic e after allergen aerosol ex posure (Fig. 2 ).
Untreate d SAL and OVA ex posed mice w e re also include d in the study, to check that the prese nt inflammation w as spec ifically induc ed by OVA challenge . Histologic analysis of lungs take n from OVA- Immunohistoche mistry w as used to visualize the MIF-ex pressing c ells in the pulmonary infiltrates and tracheobronchial lymph nodes of OVA-challenge d mic e. A majority of the le ukocytes in the pe rivasc ular and pe ribronchial pulmonary infiltrates w e re MIFpositive (Fig. 3a ). Intere stingly, large , intensely MIFpositive ce lls w ith de ndritic shape w e re see n in the supe rfic ial cortex of the tracheobronchial lymph node s in OVA-challe nged mice (Fig. 3b ). Lymphocytes, mainly located in the c ortex , w ere also staine d for MIF, although less inte nse ly (Fig. 3c ). The ex ac t identity of the MIF-positive c ells w as not further e valuate d in the prese nt study. No staining ex cep t for a yellow is h auto-fluore sc enc e (compare Fig. 3a and d ) w as observe d in control sec tions w here NRS or PBS w as used instead of the anti-MIF serum (not show n ). A majority of the leukocytes in the pulmonary infiltrates (arrows) of OVA-challenged mice were stained with anti-MIF serum (a). Interestingly, large, intensely MIF-positive cells with dendritic shape (arrow) were seen in the superficial cortex of the tracheobronchial lymph nodes in OVA-challenged mice (b). Lymphocytes, mainly located in the cortex, were also stained for MIF, although less intensely (c). In deeper portions of the lymph nodes many cells were unstained (c). A multifocal perivascular and peribronchial MPO-positive (neutrophilic) distribution (arrows) was seen at 4 h after intratracheal LPS-challenge (d). In (a) and (d) a slight yellowish auto-fluorescence, mainly located in the lung parenchyma, is observed. B = bronchus, V = blood vessel. Original magnification 3 250.
Histologic al analysis of lungs take n 4 hours after LPS challenge demonstrated a mode rate ne utrophilia perivasc ularly and p eribronchially (Fig. 3d ). Neutrophils w ere also de te c te d in alveolar w alls and space s.

Effect of anti-MIF treatment on LPS-induced airway inflammation
To ex amine the role of MIF in LPS-induc ed airw ay inflammation, c ellular profile in BALF, and neutrophilia in lung tissue and blood w ere de te rmined in anti-MIF-treated and NRS-tre ated mice . At 4 h and 24 h after intratrache al instillation of a low dose LPS total ce llular c ontent (data not show n ) and the p ercentage of ne utrophils, lymphocyte s, and mac rophages in BALF w ere similar in LPS-challenge d anti-MIF-tre ate d and NRS-treated mic e ( Fig. 4a and b ). Consis te nt w ith the findings in BALF, the number of PMN in blood did not diffe r signific antly be tw ee n anti-MIF-treate d and NRS-treate d animals at either time points (Fig. 4a and  b ). Also, no obvious differe nce in lung tissue neutrophilia w as obse rved be tw e en anti-MIF-tre ate d and NRS-treate d animals at 4 h and 24 h after LPS challe nge (data not show n ).
Since MIF has be en re ported to modulate the ex pre ssion of TNF-a and chemokines in mode ls of endotox aemia and ac ute lung injury, 11 ,1 6 TNF-a and MIP-1 a levels w ere me asured in LPS-challenge d anti-MIF-treated and NRS-treated mic e. At 4 h afte r LPS instillation mic e tre ated w ith anti-MIF se rum ex hibite d similarly inc reased le ve ls of TNF-a and MIP-1 a in BALF as NRS-treate d mic e (Fig. 5 ). Equally low levels of TNF-a in BALF w ere observe d in anti-MIF-tre ate d and NRS-tre ated mice at 24 h after LPS challe nge (12.5 ± 2.6 p g /ml and 15.9 ± 4.8 pg /ml, re spe ctively). Plasma le ve ls of TNF-a w e re below detection limit in both anti-MIF-treated and NRS-tre ated mic e at 4 h after LPS challe nge (data not show n ).

Effect of intratracheal anti-MIF treatment on LPS-induced airway inflammation
LPS-challenge d mice tre ated w ith topic al intratrache al anti-MIF serum or NRS ex hibited a similar neutrophilia in BALF at the 4-h time point (86.5 ± 1.4% and 82.4 ± 2.0% neutrophils in BALF, re spe ctively ). Also 24 h after LPS challe nge no significant diffe renc e in BALF neutrophilia w as obse rved be tw een the intratracheally treated mice . Anti-MIF-treate d and NRStre ate d mic e ex hibited 68.0 ± 1.6% and 71.0 ± 4.8% ne utrophils in BALF, respe ctive ly. The BALF neutrophilia in mice rece iving intratra cheal doses of anti-MIF serum or NRS togethe r w ith LPS w as somew hat inc re ased w he n c ompare d w ith that observe d in BALF from animals rec eiving LPS challenge only. In accord, SAL-challenge d mice treated intratrache ally w ith NRS demonstrated a mild neutrophilia at both the 4-and 24-h time points (35.9 ± 13.0 and 39.4 ± 11.8 % ne utrophils in BALF, re spe ctively).

Discussion
This study de monstrates that anti-MIF tre atment does not have any major effe cts on the eosinophil-ric h airw ay inflammation occ urring in a murine model of alle rgic asthma. Similarly, anti-MIF treatme nt did not change the neutrophilic inflammatory response se en after instillation of a low dose of LPS into the low er airw ays of mice. Although w e c annot ex clude the possibility that MIF may regulate other indice s of pulmonary inflammation than measure d in this study, the pre se nt data do not support the view that MIF is critic ally involved in pulmonary e osinophilic or neutrophilic inflammatory c onditions.
Human asthma is charac te rize d by peribronchial inflammatory infiltrates, mainly c onsisting of e osinophils, T-lymphocyte s, and macrophage s. 1 5,2 9 Give n that all these immune ce lls ex p re ss MIF in significant quantities, and that MIF has proinflammatory e ffec ts, MIF has be en imp lic ated in developme nt of asthma and other inflammatory airw ay diseases. 9,1 4 Moreover, IL-5, a cytokine c onside red pivotal for the rec ruitment of eosinophils to the airw ays in both human asthma and alle rgic mice, induce s MIF se cretion by c ultured eosinophils. 9 Indeed, abnormally high le vels of MIF have be en dete cted in BALF from asthmatic subjects. 9 In similarity to human asthma, the pre sent allergic mic e ex hibit pe ribronchial and perivasc ular infiltrates of eosinophils, T-lymphocytes, and macrophages. 2 5,2 6 The pre se nt study also show e d the prese nce of MIF-positive ce lls in the pulmonary infiltrate s. In addition, an intriguing distribution of MIF-positive c ells w as observe d in the trache obronchial lymp h node s of the allergic mice , a loc ation w here e sse ntial immune re sponse s to antigens take plac e. He nce , this allergic model w ould be w e ll suite d for ex ploration of anti-inflammatory effic ac y of anti-MIF active c ompounds.
It w as hypothe size d that anti-MIF tre atment might inhibit the allergic re sponses in the present asthma mode l both by the e nhanc ement of anti-inflammatory effects of endoge nous c orticoste roids no longer counte r-regulated by MIF, and by diminishe d production of MIF-inducible proinflammatory cytokines. How e ver, immunoneutralization of MIF during the period of alle rgen ae rosol challenge did not influenc e the magnitude of pulmonary eosinophilia and c ellular composition of inflammatory c ells obtaine d by BAL. He nce , the present study faile d to support the view that MIF is a c ritical regulator of pulmonary eosinophilic inflammation.
Instillation of a low dose of LPS into the airw ays of mic e proved suffic ie nt for produc ing c onsiste nt, predominantly neutrophilic pulmonary inflammation w ith moderate ne utrophilia perivascularly and peribronchially, but also involving alve olar w alls and space s. These fe ature s are re minisc ent of the pathology of COPD. 2 9 In addition, chronic ex posure to endotox in, as a c omp one nt of organic dust in occ upational settings, has bee n related to de ve lopme nt of COPD-like conditions. 3 0 -3 2 The pote ntial rele vanc e of the pre sent model to human COPD may further be supported by studies suggesting that inhaled LPS, as a c onstitue nt of c igarette smoke , is of importance for the de ve lopme nt of COPD. 33 The present de monstration of LPS-induced inc reases in BALF TNF-a and MIP-1 a le ve ls is also of interest in relation to COPD, sinc e these tw o mediators are pote ntially important for pulmonary neutrophil rec ruitment. 18 ,34 ,35 The se findings together w ith an inc re ase in the pe rcentage of lymphoc yte s in BALF and lack of eosinophilia, support the possibility that the present LPS challenge produce d a pote ntially useful model of COPD. 2 9 In c ontrast to most c ytokines, MIF mRNA and prote in are ex pressed constitutively in a varie ty of ce ll type s, such as monocytes /mac rop hages, T-ce lls, airw ay epithe lial ce lls, and pituitary endocrine ce lls. Proinflammatory stimuli, including LPS, are know n to inc re ase MIF mRNA ex pression above the leve l prese nt constitutiv ely. 8 ,12 LPS administrated syste mically (i.p.) in high dose s has be en used in e arlie r studie s to elucidate the role of MIF in endotox ae mia and ac ute lung injury. 1,11 ,12 ,1 6 For ex ample, treatme nt of mic e w ith anti-MIF se rum c onferred full p rote ction to the lethal effec ts of LPS (17.5 mg /kg ) administrated i.p. 1 Consiste nt w ith these pre vious findings and c onfirming the validation of the employe d anti-MIF serum, the present study show ed that anti-MIF treatme nt inhibits plasma ex travasation in re sponse to a high syste mic dose of LPS.
In a rat model of acute lung injury, Makita e t a l. 1 1 demonstrate d that anti-MIF treatme nt reduce d the number of ne utrophils pe r alveolus and the BALF ne utrophilia, induc ed by a high dose of LPS given systemic ally. In mic e, i.p. administration of LPS (1-20 mg /kg ) does not induce transpulmonary neutrophil migration and infiltration of ne utrophils into the alveolar space, but only ne utrophil sequestration w ithin the lung vasculature . 36 -3 8 In this study, a fraction of the previously employe d i.p. dose s of LPS (250-5000-fold less ) w as give n loc ally into the airw ays. Refle cting the low dose and route of administration, plasma levels of TNF-a w ere below dete ction limit at 4 h after LPS challenge . In app arent contrast to the important role of MIF in host re sponse s to high systemic doses of LPS, anti-MIF treatment did not change the ne utrophil-rich inflammatory response induc ed by this mode of LPS ex posure. To ex plain these data it is sugge ste d that the importanc e of MIF in different models of LPS-induced host re ac tions may vary depe nding on the dose and /or the administration route of LPS. In ac cord, it has previously bee n show n that the mechanisms behind host responses to LPS may be completely different in mode ls using high or low doses of LPS. 39 Ge ne tically MIF-defic ie nt mice are also resis tant to the lethal e ffec ts of a high syste mic dose of LPS, but susceptible to a combination of a low dose of LPS and D-galactosamine . 16 The de monstration that MIF is not involved in the prese nt pulmonary ne utrophilic inflammation may re duc e the promise of anti-MIF compounds as future COPD drugs.
In c onc lusion, the pre sent data suggest that MIF is not c ritically important for allergic e osinophilic, or LPS-induce d ne utrophilic inflammation, in airw ays of mic e. If translatable to human disease c onditions, 4 0 these findings do not support the notion that MIF inhibitors w ill be effe ctive against eosinophilic or ne utrophilic respiratory disease s, such as allergic asthma and COPD.