Suppressive activity of a macrolide antibiotic, roxithromycin on co-stimulatory molecule expression on mouse splenocytes in vivo.

The influence of roxithromycin (RXM) on the expression of co-stimulatory molecules, CD40, CD80 and CD86, was examined in vivo. When BALB/c mice were immunized intraperitoneally with two doses of dinitrophenylated ovalbumin (DNP-OVA) at 1 week intervals, intraperitoneal administration of RXM at 250 microg/kg once a day for 14 days strongly suppressed IgE contents in sera obtained from mice 22 days after the first immunization. In addition, RXM treatment of mice suppressed endogenous IL-4 contents in aqueous spleen extracts, which were enhanced by DNP-OVA immunization. We next examined the influence of RXM on co-stimulatory molecule expression on splenic lymphocytes. RXM treatment of the immunized mice caused suppression of CD40 expression, but this treatment did not affect CD80 and CD86 expression.


Introduction
Administration of macrolide antibiotic s such as e rythromycin and trole andomycin can favourably modify the clinical status of patie nts w ith inflammatory diseases. [1][2][3] Although inve stigation of the mechanisms of improve ment has suggested that it is not due to anti-mic robiologic al e ffec ts of the drugs, the prec ise me chanisms are not w e ll unde rstood. 4 -6 We have demonstrate d that oral administration of rox ithromycin (RXM) onc e a day for 21 days could suppress the ability of lymp hocyte s to produc e se veral types of inflammatory c ytokine s such as IL-1, IL-3 and IL-5 in re sponse to mitoge nic stimulation in v ivo and in vitro . 6 -9 Subse que ntly, Eyrand 1 0 and Forsgre n 11 have re ported that macrolide antibiotic s such as erythromyc in, troleandomycin and RXM inhibited chemotax is and generation of inflammatory me diators from polymorphonuclear leukocyte s w hen the c ells w e re cultured in v itro in the prese nce of macrolides.
It is ge nerally acc epted that antigen-spe cific immune responses are initiated after the collaboration of T-ce lls w ith antigen prese nting cells (APC). 1 2 Rec ently, it has be en found that an optimal T-ce ll activation requires another c ell-to-cell interaction. 12 ,1 3 The first signal transduction is due to the interaction betw e en T-c ell re ce ptor and major histoc omp atibility complex class II molec ule w ith antigenic determinant. The se cond signal is provide d by the direc t contac ts of c o-stimulatory mole cules on T-ce lls w ith their ligands on APC. 13 ,1 4 The signal through the binding of CD28 /CTLA4 on T ce lls w ith its ligands, CD80 and CD86 on APC is a crucial co-stimulatory pathw ay. 12 -1 4 Furthermore, e ngageme nt of the B-ce ll marker CD40 by its ligand CD40L is also rec ognized to play an important role in T-ce ll-dep endent isotype sw itching to IgE. 14 ,15 The re is much evidence that ex pression of the c ostimulatory mole cules CD40, CD80 and CD86 on peripheral blood le ukocytes from patients w ith inflammatory disease s w as upregulate d compare d w ith normal subje cts , 16 -18 suggesting the importanc e of the se molec ule s in the induc tion and the de ve lopme nt of the diseases. The re fore, the present study w as undertake n to answ er the unre solved questions regarding the favourable e ffec ts of macrolide antibiotics on inflammatory disease s by ex amining the influenc es of RXM on ex pre ssion of co-stimulatory molecules in mic e.

Materials and Methods
Mice BALB/c male mice , 5 w e eks of age, w e re p urchase d from Charles River Japan Inc. (Atsugi, Japan ).
Immunization BALB/c mice w ere immunized by intrape ritone al injection of 5.0 m g /ml of dinitrophe nylated ovalbumin (DNP-OVA) absorbed on 4 mg Al(OH) 3 , and booste d intraperitone ally w ith the same dose of antigen 7 days later.

Drugs and treatment
RXM w as kindly donated by EISAI Co. Ltd (Tokyo, Japan ) as a w ater insoluble pure p ow der. The age nt w as dissolved in me thyl alcohol at 50 mg /ml and dilute d w ith normal saline so as to give 1.0 mg /ml. This solution w as the n ste rilized by passing through a 0.22 m m filter and stored at 4°C until used. The immunize d mice w e re given 250 m g /kg /day of RXM intraperitone ally for 14 days starting 7 days after the first immunization and the non-immunize d mic e injected w ith the same dose of RXM for 2 w eeks. Sinc e our previous reports have show n that administration of 2% alcohol did not show any adverse effects on mouse immune re sponses, 6,7 ,9 the c ontrol mic e rece ive d saline in the same route.

Assay for serum total IgE
The blood w as obtaine d from the mice by cardiac puncture under e ther anesthe sia 22 days after the first immunization. After clotting, the se rum w as obtaine d and total se rum IgE le vels w ere measured by mouse IgE e nzyme immunoassay kits (YAMASA Co. Ltd, Chiba, Japan ) ac cording to the manufacture r's recomme nde d proce dures. The assay w as performe d in duplic ate and the re sults w e re ex pressed as the mean ng /ml ± SD of five individual mic e.

Preparation of aqueous spleen extracts
The splee n w as re moved from mic e killed unde r e ther anesthe sia and store d in ice-cold PBS until proce sse d. The organ w as homogenize d in 0.5 ml PBS in an ic ecold w ate r bath for 3 min using a glass tissue homogenizer. The supernatant w as obtaine d by c entrifugation of the homoge nize d materials at 10 000 g for 1 h at 4°C and use d for aqueous splee n ex trac t.

Assay for IL-4
IL-4 c onc entration in aqueous sple en ex tract w as assaye d by commerc ially available mouse IL-4 ELISA kit (GENZYME Corp., Cambridge , MA, USA) ac cord-ing to the manufacture r's rec ommended proc edures. The ELISA w as done in duplicate and the re sults w ere ex pre sse d as the me an pg /ml ± SD of five individual mic e.

Preparation of spleen cell suspension
The splee n w as re moved from mic e killed unde r e ther anesthe sia and pre sse d through a 60 gauge ste e l me sh to give a single c ell suspension. After ce ntrifugation at 1000 r.p.m. for 10 min at 4°C, the pelleted ce lls w ere tre ate d w ith 0.15 M Tris -0.75% NH 4 Cl solution for 10 min to lyse red blood c ells. After filte ring through a 200 gauge stee l me sh, the residual c ells w e re w ashe d three time s and suspe nded in PBS at a c onc entration of 1 ´10 6 ce lls /ml.

Monoclonal antibodies (mAbs) and flow cytometry
The mAbs used in this study w ere anti-mouse CD16 / CD32 mAb, fluorescein isothiocyanate (FITC)-conjugate d anti-mouse CD40 mAb (hamste r IgM), phycoerythrin (PE)-conjugated anti-mouse CD80 mAb (hamster IgG), FITC-conjugated anti-mouse CD86 mAb (rat IgG2a ). The y w ere purchased from PharMinge n (San Die go, CA, USA). To block nonspe cific adhere nce of antibodie s to murine Fc rec eptors, sple en ce lls (1 ´10 6 ) w ere incubated w ith 1.0 m g of anti-mouse CD16 /CD32 mAb for 5 min at 4°C, w ashed, and then labe lle d w ith either anti-mouse CD80, CD86, or CD40 for 20 min in an ic e-cold w ate r bath. Afte r w ashing onc e, fluoresc ent staining w as analysed immediately by flow c ytometry on a FACScan (Bec ton Dickinson, Mountain Vie w, CA, USA). The fluoresce nt intensity of c ells w as ex pressed as the me an ± SD of four different ex perime nts. Spleen cells w ere also staine d w ith monoclonal immunoglobulin isotype standards (FITC-conjugate d hamste r IgM against trinitro p henol and PE-c onjugate d hamste r IgG against ke yhole limpet hae moc yanin ) purchase d from PharMinge n, and fluoresc ent staining w as analyse d in a similar manne r.

Statistical analysis
The statistical significance of the diffe renc e in the me an values betw e en tw o groups w as analysed by Mann -Whitne y U te st.

Influence of RXM on IgE and IL-4 production in mice
The first se t of ex pe riments w as c arried out to ex amine the influe nce of RXM on IgE and IL-4 production in immunized mic e. As show n in Table 1, RXM treatme nt of non-immunized mic e did not caused changes of IgE and IL-4 le ve ls: IgE and IL-4 contents in sera and aqueous splee n ex tracts from RXM-treated, non-immunized mice w e re nearly ide ntical to those from saline-treate d, non-immunized mic e (P > 0.05 ). On the other hand, RXM tre atment c ause d significant suppression of IgE and IL-4 contents in se ra and ex tracts, w hich w ere enhanced by DNP-OVA immunization. IgE levels in sera from saline-tre ate d, immunize d mice w e re signific antly de cre ase d from 223.8 ± 14.3 ng /ml to 72.4 ± 6.3 ng /ml by RXM tre atment (P < 0.001 ). IL-4 le vels in the ex trac ts w ere also decre ased from 81.5 ± 9.7 pg /ml to 31.5 ± 5.7 pg / ml by RXM treatment (P < 0.001 ).

Influence of RXM treatment on co-stimulatory molecule expression on spleen cells in vivo
This study w as de signed to ex amine the influe nce s of RXM on various profiles of CD40, CD80 and CD86 to be ex pressed on B-lymphocytes. In flow c ytometry on FACSc an, w e gate d and analysed on the lymphocyte position /population of sc attered dots of splee n cells in the display of compute r. Figure 1 show s one typical profile among results obtaine d in four ex pe riments of mic e immunized and non-immunized by DNP-OVA antigen. RXM treatment of non-immunize d mic e sc arce ly affec te d the ex pression p rofile s of the CD molecules on the gated sple nic lymp hocyte s (Fig. 1,  left). In immunize d mic e, how ever, RXM ex erte d remarkable suppre ssion of CD40 molec ule ex pre ssion on the gated lymphocyte population, w hich w as enhance d by DNP-OVA antigen (Fig. 1, right). The fluoresce nt intensity of splenic lymp hocyte s pre pare d from saline-treated mice w as 670.8 ± 89.65 and that from RXM-tre ated mice w as 344.56 ± 52.48. On the othe r hand, ex pre ssion of CD80 and CD86 mole cules w as not influenc ed by the 2-w e ek treatme nts w ith RXM (Fig. 1, right).

Discussion
Se ve ral studies have show n that mac rolide antibiotic s can favourably influe nce the clinical condition of ce rtain patients w ith allergic diseases. 1-3 Naturally, much efforts have been done to understand the me chanisms by w hich mac rolide antibiotic s modify the clinical status of allergic diseases. 4 -7 , 9 -11 Howe ve r, the pre cise mechanisms are not w ell defined.
The re is much e videnc e that allergic disease s are caused by IgE antibodies against spe cific allerge ns. Ishizaka has de sc ribe d that the prevention or suppre ssion of IgE antibody formation against allerge ns might be one of fundamental treatment of allergic disease s. 19 There fore, w e ex amined the influe nce of RXM on IgE production in vivo . The results obtaine d (Table 1 ) clearly de monstrate that 2-w e ek treatments w ith RXM could suppress IgE production in actively se nsitized mice.
Pre viously, w e found that RXM signific antly suppressed the enhancement of 3 H-thymidine uptake by human pe ripheral blood leukoc ytes induc ed by in v itro stimulation w ith T-cell mitogen (Conc anavalin A), but not w ith B-c ell mitogens. 7 We also reporte d the suppressive activity of RXM on the ability of T-c ells to produce several types of c ytokine s including IL-5 in v itro and in v ivo . 6 Although ex perime ntal and clinical data show that IgE synthesis in B-c ells is depende nt on a c omplex proce ss involving se veral ce llular and mole cular inte ractions, 15 ,20 -2 2 there is a established conce pt that IL-4 sec rete d by T-cells, espec ially Th2 type he lper T-ce lls, is the most important cytokine for IgE gene ration. Taken toge the r, the present re sults ( Table 1) sugge st that RXM treatme nt inhibits in vivo IL-4 sec retion in response to DNP-OVA immunization and results in suppression of IgE levels in immunize d mice. This suggestion is supported by the finding that treatme nt of mic e w ith RXM signific antly inhibits endoge nous IL-4 le vels, w hich w e re enhanced by DNP-OVA immunization.
The re are c irc umstantial evidences that allergenspe cific Th2 type he lpe r T-c ells play a pivotal role in the induction and development of many allergic diseases. 23 -25 And also, optimal T-ce ll ac tivation is rec ognized to re quire not only interaction betw e en antigen peptide bound to MHC mole cules of APC and the T-c ell rec eptor, but also additional signals, socalled co-stimulation. A numbe r of rece nt re ports have show n that ligation of CD28 on T-cells and CD80 or CD86 on APC is essential for activation of Th2 type Rox ith ro m y cin a n d co -s tim u la to r y m o le cu le  he lper T-ce ll and production of IL-4. 2 6,2 7 Additionally, inte rac tion betw e en CD40 on ac tivated B-ce lls and its ligand CD40L on T-ce lls has be en rep orte d as an important co-stimulatory signal for sw itch rec ombination to IgE synthesis in the p re sence of IL-4.
Therefore, w e nex t ex amine d the influence of RXM on CD40, CD80 and CD86 ex pression on sple en c ells. The pre sent results cle arly show ed that RXM treatme nt of mice strongly suppressed the ex p re ssion of CD40 co-stimulatory mole cule enhance d by DNP-OVA immunization, sugge sting that prote c tive effe ct of RXM against IgE hype r-produc tion is associate d w ith its suppre ssive e ffec t on co-stimulatory molecule ex pre ssion.
Although the pre se nt study may provide possible me chanisms by w hich macrolide antibiotics can modify favourably the clinical c ondition of allergic diseases, more in-depth analyses are needed to clarify the mode of ac tion of the age nts in the diseases. For ex ample, there are rese arch subjec ts conce rning CD4 vs. CD8 ex pre ssion in thymus and lymph node, and sIgD vs. sIgM ex p re ssion in splee n to show the ex istenc e of mature lymphocyte s. In addition, the suppression mechanisms of RXM regarding syntheses of prote ins such as IL-4 and CD40 mole cules ex amined here are not clear at present. How e ve r, FK-506 and rapamyc in, macrolide antibiotic s, ex e rt their immunosuppressive effe cts by formation of complex betw e en immunophilin and age nts w hich can inhibit gene transc ription. 28 ,29 It is possible that RXM binds to immunophilin, intracellular binding prote in, and the complex e s inte rfere s gene transcription, resulting in inhibition of prote in synthe sis. Another ex pe rime nts are also nee de d to clarify this point.