Interleukin-4 Modulates the Inflammatory Responsein Ifosfamide-Induced Hemorrhagic Cystitis

Francisco Yuri Bulcão Macedo,1 Lívia Talita Cajaseiras Mourão,1 Helano Carioca Freitas,1

Roberto C. P. Lima-Júnior,1 Deysi Viviana Tenazoa Wong,1 Reinaldo Barreto Oriá,2

Mariana L. Vale,1 Gerly Anne Casto Brito,2 Fernando Q. Cunha,3 and Ronaldo A. Ribeiro1,4,5

Abstract—We investigated whether interleukin-4 (IL-4) is present and capable of reducing inflam-matory changes seen in ifosfamide-induced hemorrhagic cystitis. Male Swiss mice were treated withsaline or ifosfamide alone or ifosfamide with the classical protocol with mesna and analyzed bychanges in bladder wet weight (BWW), macroscopic and microscopic parameters, exudate, andhemoglobin quantification. In other groups, IL-4 was administered intraperitoneally 1 h beforeifosfamide. In other experimental groups, C57BL/6 WT (wild type) and C57BL/6 WT IL-4 (−/−)knockout animals were treated with ifosfamide and analyzed for changes in BWW. Quantification ofbladder IL-4 protein by ELISA in control, ifosfamide-, and mesna-treated groups was performed.Immunohistochemistry to tumor necrosis factor-alpha (TNF-α) and interleukin-1 beta (IL-1β) aswell as protein identification by Western blot assay for inducible nitric oxide synthase (iNOS) andcyclooxygenase-2 (COX-2) was carried out on ifosfamide- and IL-4-treated animals. In other ex-perimental groups, antiserum against IL-4 was given 30 min before ifosfamide. In IL-4-treatedanimals, the severity of hemorrhagic cystitis was significantly milder than in animals treated withifosfamide only, an effect that was reverted with serum anti-IL-4. Moreover, knockout animals forIL-4 (−/−) exhibit a worse degree of inflammation when compared to C57BL/6 wild type. Exoge-nous IL-4 also attenuated TNF-α, IL-1β, iNOS, and COX-2 expressions in ifosfamide-treated bl-adders. IL-4, an anti-inflammatory cytokine, attenuates the inflammation seen in ifosfamide-inducedhemorrhagic cystitis.

KEY WORDS: IL-4; ifosfamide; bladder; hemorrhagic cystitis; TNF-α; IL-1β; iNOS; COX-2.

INTRODUCTION

Ifosfamide is a chemotherapeutic agent primarilyindicated in some hematological conditions such aschronic lymphocytic leukemia and as adjunctive therapy

for malignant lymphomas and non-Hodgkin’s lym-phoma. Also, it can be given for the treatment ofrefractory tumors that do not respond to conventionaltherapy, such as soft tissue sarcoma and refractory germcell tumors [1]. Within the liver, ifosfamide is cleavedinto its active compound isofosforamide mustard, andacrolein, an inactive metabolite excreted by the urinarysystem [2].

Hemorrhagic cystitis (HC) is a complication thatlimits the use of ifosfamide. It has been proposed thatsuch toxicity is attributable to the renal excretion ofacrolein. Urothelial damage occurs upon direct contactwith acrolein, which causes edema, ulceration, neo-vascularization, hemorrhage, necrosis, and expression ofinflammatory enzymes [3].

Prophylactic mesna (2-mercaptoethanesulfonicacid) has shown some efficacy in becoming the drug of

1 Department of Physiology and Pharmacology, Faculty of Medicine,Universidade Federal do Ceará, Rua Cel. Nunes de Melo, 1127,60430-270 Fortaleza, Ceará, Brazil

2 Department of Morphology, Faculty of Medicine, UniversidadeFederal do Ceará, Fortaleza, Ceará, Brazil

3 Department of Pharmacology, School of Medicine of Ribeirão Preto,University of São Paulo, Ribeirão Preto, SP, Brazil

4 Department of Clinical Oncology, Cancer Institute of Ceará,Fortaleza, Ceará, Brazil

5 To whom correspondence should be addressed at Department ofPhysiology and Pharmacology, Faculty of Medicine, UniversidadeFederal do Ceará, Rua Cel. Nunes de Melo, 1127, 60430-270Fortaleza, Ceará, Brazil. E-mail: yuribulcao@gmail.com

0360-3997/12/0100-0297/0 # 2011 Springer Science+Business Media, LLC

Inflammation, Vol. 35, No. 1, February 2012 (# 201 )DOI: 10.1007/s10753-011-9319-3

297

1

choice within a short period of time [4]. It binds toacrolein inside the urinary system, detoxifying it, byexcretion of an inactive substance. In the absence ofadequate uroprotection with mesna, HC becomes dose-limiting, with an average incidence of 20–40% [5].However, despite positive results after the prophylacticuse of mesna for preventing HC, bladder protection isnot always achieved, particularly when a lesion hasalready been established [6]. The occurrence of cystitisdespite mesna was observed both experimentally [7] andclinically [8]. These facts enhance the importance ofstudies that investigate the mechanisms involved inbladder lesions resulting from alkylating agent therapies.

Our group has shown the involvement of cytokinessuch as tumor necrosis factor alpha (TNF-α) andinterleukin-1 beta (IL-1β) in the pathogenesis ofhemorrhagic cystitis playing a role in some inflamma-tory events [9]. Nitric oxide (NO) also participates inthis process [10]. Either anti-TNF-α or anti-IL-1β seraadministration significantly reduced the vesical inflam-matory events as well as the rise in inducible nitric oxidesynthase (iNOS) expression and activity [9], in whichinduction is mediated by the action of platelet-activatingfactor (PAF) [10]. Dexamethasone, a corticosteroid thatinhibits TNF-α, IL-1β, PAF, and iNOS expressions,proved to be effective in preventing inflammatorychanges induced by ifosfamide in rat bladders [7]. Itwas also demonstrated by our group that cyclooxyge-nase-2 (COX-2), an enzyme seen in most inflammatorystates, is expressed in animals with hemorrhagic cystitis,and that TNF-α and IL-1β could induce this expressionbased on the fact that TNF-α and IL-1β inhibitors nearlyabolished COX-2 expression [11]. An increase in COX-2 enzyme was also demonstrated when acrolein wasinjected directly into the bladders of the rats [3].

Interleukin-4 (IL-4) has marked inhibitory effectson the expression and release of the proinflammatorycytokines. It is able to block or suppress the monocyte-derived cytokines, including IL-1β, interleukin-6 (IL-6),interleukin-8 (IL-8), and TNF-α [12, 13], macrophagecytotoxic activity, and macrophage-derived NO produc-tion [14]. In addition to the inhibition of the productionof inflammatory cytokines, IL-4 has been reported toinhibit the induction of COX-2, with a consequentreduction in the production of prostaglandins [15, 16].It has been shown that 4 h following cyclophosphamide-induced cystitis (similar to ifosfamide), there was asignificant increase in IL-4 mRNA compared with IL-4mRNA from control urinary bladder, a finding thatpersisted 48 h later [17].

Given the role of TNF-α, IL-1β, iNOS, and COX-2in the pathogenesis of ifosfamide-induced hemorrhagiccystitis and additionally the capacity of IL-4 to inhibitthe expression of these proinflammatory molecules inseveral pathological conditions, the overall aim of thepresent study is to evaluate whether IL-4 is present uponifosfamide-induced cystitis and whether it is able toreduce the inflammatory findings in hemorrhagic cys-titis. In addition, we investigate whether IL-4 is capableof inhibiting two inflammatory cytokines (TNF-α andIL-1β) and two inflammatory enzymes (iNOS and COX-2) present in this pathology.

MATERIALS AND METHODS

Animals. Male swiss mice (25–30 g, six mice pergroup) were kept in a temperature-controlled room, withaccess to food and water ad libitum, and under 12 h ofdark–light cycles until use. The Local Ethics Committeefor Animal Experiments approved the entire protocolfollowing the rules approved by the Declaration ofHelsinki and the Guide for Care and Use of LaboratoryAnimals, National Institutes of Health (Bethesda, MD,USA). Also, 25- to 30-g C57BL/6 WT (wild type) and IL-4 (−/−) knockout mice (strain IL-4tm1Nnt/J, fromUniversityof São Paulo, Ribeirão Preto, Brazil) were housed in thesame conditions and taken to the testing room at least 1 hbefore experiments. The mice were used only once.

Drugs. The following materials were obtained fromthe indicated sources. Murine recombinant IL-4 wasprovided by the National Institute for Biological Stand-ards and Control (South Mimms, Hertfordshire, UK).Ifosfamide (Holoxane®, 1 g; ASTA Medica AG,Frankfurt Germany), Mesna (Mitexan®, 200 mg; ASTAMedica AG, Frankfurt Germany), and Evans blue(Sigma Chemical Co.) were dissolved in sterile saline.

Hemorrhagic Cystitis Model. Ifosfamide (400 mg/kg)was injected intraperitoneally (ip), and the animals werekilled 12 h later. The bladders were removed and emptiedof their urinary contents. The Swiss mice received either anip injection of saline (1 ml/200 g of rat weight) orifosfamide. In other experimental groups, the Swiss micewere treated 1 h before ifosfamide with IL-4 (0.4; 2; 10 ng,ip.). C57BL/6 WT and IL-4(−/−) mice received eithersaline or ifosfamide. The classical protocol using threedoses of mesna (20% of the whole ifosfamide dose, onedose immediately before ifosfamide, 4 h and 8 h after,respectively) for ifosfamide-induced cystitis preventionwas used as a positive control. The bladder wet weight

298 Macedo, Mourão, Freitas, Lima-Júnior, Wong, Oriá, Vale, Brito, Cunha and Ribeiro

(BWW) was measured and expressed as milligrams per20 g of body weight (mean±SEM).

Macroscopic Evaluation. Bladders were grosslyexamined for edema and hemorrhage. According toGray’s scoring criteria [18], edema was consideredsevere (3+) when fluid was seen externally and internallyin the bladder walls, moderate (2+) when confined to theinternal mucosa, mild (1+) between normal and moder-ate, and normal (0) when no edema was observed.Hemorrhage was scored as follows: 3+, intravesicalclots; 2+, mucosal hematomas; 1+, telangiectasia ordilatation of bladder vessels; and 0, normal.

Microscopic Evaluation. Bladders were fixed in for-malin at 10%, embedded in paraffin, and processed forhematoxylin and eosin (Reagen) staining. Histopatho-logical analysis was performed by a blinded person whowas unaware of the treatments and group divisions, andscored as follows: (0) normal epithelium and absence ofinflammatory cell infiltration and ulceration, (1) mildchanges involving reduction of urothelial cells, flatteningwith submucosal edema, mild hemorrhage, and fewulcerations, and (2) severe changes including mucosalerosion, inflammatory cell infiltration, fibrin deposition,hemorrhage, and multiple ulcerations.

Evan’s Blue Extravasation Assay. Vesical vascularpermeability was evaluated by the Evan’s blue extrav-asation technique. Following the same previous groupdivisions (n=6), 2.5% Evans blue (25 mg/kg) wasinjected intravenously via the retro orbital plexus30 min before the animals were sacrificed. Bladderswere then excised, dissected, and placed into glass tubescontaining a formamide solution (1 mL/bladder) at 56°Covernight to extract the stain. The total extracted dyewas determined by measuring the absorbance change at630 nm (ELISA). At the same time, an absorbanceconcentration curve was determined. The BWW wasmeasured as a parameter of vesical edema and expressedas micrograms per 20 g of body weight (mean±SEM).

Hemoglobin Assay. The presence of hemorrhage inthe bladder was determined by a hemoglobin assay usingthe cyanomethemoglobin method (Bioclin, Belo Hori-zonte, MG, Brazil). This standard hemoglobin kit(Bioclin) contains the color reagent for hemoglobindetection (Drabkin’s reagent). The whole bladder washomogenized in Drabkin’s reagent (100 mg tissue permilliliter of reagent). Shortly afterwards, the sampleswere centrifuged at 10,000×g for 10 min. The super-natants were then removed, filtered using a 0.22-μmfilter, and centrifuged at 10,000×g for 10 min. Absorb-ance was measured at 540 nm (ELISA), and hemoglobin

concentration was read off of a standard curve andexpressed as micrograms Hb per 100 mg wet tissue.

Preparation of ELISA Samples for IL-4. Adult Swissmice were sacrificed as above, and the bladder (n=6 foreach time point; n=8 for control) was rapidly dissected andweighed. Individual bladders were solubilized in a T-PERtissue protein extraction reagent (1 g tissue/20 ml; eBio-science, CA, USA) with protease inhibitor cocktail tablets.Bladder tissue was disrupted with a homogenizer and thencentrifuged (10,000 rpm for 5 min). The supernatants wereused for IL-4 quantification. Total protein was determinedby the mouse IL-4 protein assay reagent kit (eBioscience)and expressed as picograms per microgram of bladder totalprotein (mean±SEM).

Effect of Antisera Against Interleukin-4 upon Ifosfamide-Induced Cystitis. Fifty microliters of antiserum againstmurine interleukin-4 (monoclonal IgG antibodies tomurine IL-4: BVDG and 11B11 (Professor F. Liew,University of Glasgow, UK) or of a control (preimmune)serum was diluted in 250 μl of saline and then injectedinto the peritoneal cavities of mice; 30 min later, theanimals received an ip injection of ifosfamide, andchanges on BWW were measured and expressed asmilligrams per 20 g of body weight (mean±SEM). Thecontrol monoclonal antibody was a purified unrelatedIgG raised against ovalbumin in our laboratory.

Immunohistochemistry for IL-β and TNF-α. Immuno-histochemistry was performed using the streptavidin–biotin–peroxidase method [19] in formalin-fixed, paraf-fin-embedded tissue sections (4 μm thick), mounted onpoly-L-lysine-coated microscope slides. The sections weredeparaffinized and rehydrated through xylene and gradedalcohols. After antigen retrieval, endogenous peroxidasewas blocked (15 min) with 3% (v/v) hydrogen peroxideand washed in phosphate-buffered saline (PBS). Sectionswere incubated overnight (4°C) with primary rabbit anti-body (polyclonal goat anti-mouse) diluted 1:200 and1:400, respectively, for TNF-α and IL-1β in PBS plusbovine serum albumin (PBS–BSA). The slides were thenincubated with biotinylated goat anti-rabbit IgG diluted1:200 for TNF-α and IL-1β in PBS–BSA. After washing,the slides were incubated with avidin–biotin–horseradishperoxidase conjugate (Strep ABC complex by Vectastain®ABC Reagent and peroxidase substrate solution) for30 min, according to the Vectastain protocol. TNF-α andIL-1β were visualized with the chromogen 3,3′diamino-benzidine. Negative control sections were processedsimultaneously as described above, but with the firstantibody being replaced by PBS–BSA 5%. None of thenegative controls showed immunoreactivity. Slides were

299Interleukin-4 in Ifosfamide-Induced Hemorrhagic Cystitis

counterstained with Harry’s hematoxylin, dehydrated in agraded alcohol series, cleared in xylene, and coverslipped.

Assessment of Positively Stained Cells. Each tissue wasexamined for TNF-α and IL-1β expressions, comparedwith staining in experiment-matched negative controlsand graded on a scale of 0 to 3+ viz. that follows: absentexpression (0), low expression (+1), moderate expres-sion (2+), and high expression (3+) by two observersblinded to the treatment on two separate occasions. Thefollowing sites were analyzed: urothelium, subepithelialcells, and parietal bladder cells.

Western Blot for iNOS and COX-2. The animals had asample of their bladder removed to determine the amountsof iNOS and COX-2 present. Specimens were immediatelystored at −70°C until required for the assay. Briefly, thetissues were homogenized in 0.2 ml of lysis buffercontaining protease inhibitors (50 mM Tris–HCl[pH 8.5]; 50 mM NaCl; 0.1 mM EDTA; 1% Tween 20;1 mM [each] dithiothreitol, leupeptin, aprotinin, andphenylmethylsulfonyl fluoride). Total protein (50 μg pro-tein/well) was resolved on 10% (iNOS) or 12.5% (COX-2)sodium dodecyl sulfate–polyacrylamide gel electrophore-sis and transferred to a nitrocellulose membrane (Hybond-ECL, Amersham Pharmacia Biotech, Amersham, UK).The membranes were blocked with 5% skimmed milk/Tris-buffered saline with 0.1% Tween 20 for 14 h at 4°Cfollowed by an incubation period of 1 h at room temper-ature with the primary antibodies (rabbit polyclonal anti-iNOS, 1:500; anti-COX-2, 1:400, or anti-β-Actin, 1:1,000;Santa Cruz Biotechnology, Santa Cruz, CA). The blotswere washed, followed by incubation with horseradishperoxidase-conjugated secondary antibody (donkey anti-rabbit immunoglobulin G, 1:1,000; Santa Cruz Biotech-nology, Santa Cruz, CA) for 1 h at room temperature. Themembranes were washed, incubated with electrogeneratedchemiluminescence (Amersham Pharmacia Biotech), andexposed to Hyperfilm ECL (Amersham Pharmacia Bio-tech) to develop the western blot. Densitometry analyseswere performed by ImageJ 1.4 software (National Instituteof Health, USA). Data were expressed as the relativedensity of iNOS/β-Actin and COX-2/β-Actin bands.

Statistical Analysis. The results were reported asmean±SEM for quantitative variables (BWW, vascularpermeability, hemoglobin and IL-4 quantification, andwestern blot) or as median values (macroscopic, histo-pathological data, and immunohistochemistry) of at leastfive determinations. Statistical significance (p<0.05) wasassessed by analysis of variance (ANOVA) followed byBonferroni’s test. For macroscopic, microscopic, andimmunohistochemistry data, we used statistical evalua-

tion by Kruskal–Wallis nonparametric analysis of var-iance followed by the Dunn test.

RESULTS

IL-4 Production is Increased After Ifosfamide Adminis-tration. Intraperitoneal injection of ifosfamide (400 mg/kg) induced an increase of 60% on bladder IL-4 proteinas measured by ELISA when compared to control group(animals treated with saline only, Fig. 1a).

Increase on BWW and Evan’s Blue Assay is Reduced withIL-4. Intraperitoneal injection of ifosfamide (400 mg/kg)induced a marked increase in BWW 12 h after itsadministration (172% when compared to control group, p<0.05). The administration 1 h before ifosfamide injectionwith IL-4 ip (2 and 10 ng/animal; Fig. 1b) significantlyprevents the increase in bladder wet weight induced byifosfamide, around 27% and 39%, respectively. When thedose of 0.4 ng was used, there was no significant reductionin BWW. Similar findings were seen when Evan’s bluetechnique was used. Edemawas decreased by 29% and 24%(compared to saline group, p<0.05) in both groups treatedwith IL-4 (2 and 10 ng/animal, respectively; Fig. 1c).

Increase on Bladder Hemoglobin Concentration isReduced with IL-4. Ifosfamide ip (400 mg/kg) induced amarked increase in bladder hemoglobin concentration12 h after its administration (171% compared to controlgroup, p<0.05). The administration 1 h before ifosfa-mide injection with IL-4 ip (2 and 10 ng/animal; Fig. 1d)significantly prevents the increase in hemorrhageinduced by ifosfamide measured by tissue hemoglobinconcentration (47% and 61%, respectively). When thedose of 0.4 ng was used, there was no significantreduction in hemoglobin concentration.

Fig. 1. a Quantification of IL-4 protein upon ifosfamide-induced cys-titis. b Effect of treatment with IL-4 in different doses on bladder wetweight in ifosfamide-induced hemorrhagic cystitis. c Quantification ofvascular permeability to Evan’s blue on ifosfamide-induced cystitis andeffects of IL-4. d Quantification of bladder hemoglobin in ifosfamide-induced cystitis and effects of IL-4. e Effect of antiserum against IL-4on ifosfamide-induced hemorrhagic cystitis (CS: pre immune serum). fEffect on bladder wet weight in ifosfamide-induced hemorrhagic cys-titis on knockout mice for the IL-4 (−/−) gene when compared to thewild-type C57BL/6 WT. The results are reported as mean±SEM (n=6per group). *p<0.05 vs. control group (C, treated with saline alone)and #p<0.05 vs. ifosfamide alone by ANOVA and Bonferroni’s tests.MMM (classical protocol for the prevention of cystitis using threedoses of mesna).

b

300 Macedo, Mourão, Freitas, Lima-Júnior, Wong, Oriá, Vale, Brito, Cunha and Ribeiro

a b

C - MMM0.00

0.02

0.04

0.06

*

Ifosfamide (400 mg/kg)

Inte

rleuk

in-4

(p

g/ μ

g bl

adde

r to

tal p

rote

in)

0

20

40

60

#

#

*

IL-4

Ifosfamide (400 mg/kg)

C MMM - 0.4ng 2ng 10ng

Bla

dder

Wet

Wei

ght (

mg/

20g)

c d

C MMM - 0.4 ng 2 ng 10 ng0

50

100

150

200

*

##

IL-4

Ifosfamide 400 mg/kg

Eva

n's

Blu

e (μ

g/20

g)

C MMM - 0.4 ng 2 ng 10 ng 0

30

60

90

IL-4

*#

#

Ifosfamide 400 mg/kg

Hem

oglo

bin

(μg

. 10- 3 /

g bl

adde

r)

C CS - Anti IL-40

10

20

30

40

50

60

70

*

#

*

Ifosfamide 400 mg/kg

Bla

dder

Wet

Wei

ght (

mg/

20g)

e f

C57BL/6 WT C57BL/6 WT IL-4 (-/-)0

10

20

30

40

Saline

*

#

Ifosfamide 400 mg/kg

Bla

dder

Wet

Wei

ght (

mg/

20g)

301Interleukin-4 in Ifosfamide-Induced Hemorrhagic Cystitis

Potentiation by Antiserum Against IL-4 of the CystitisInduced by Ifosfamide. Antiserum against IL-4 (50 μl, ip),but not a preimmune (control) serum (50 μl, ip), injected30 min before ifosfamide (400 mg/kg, ip) induced amarked increase in BWW 12 h after its administration(28% compared to animals treated with ifosfamide only,p<0.05; Fig. 1e).

Increase in BWW is Intensified in IL-4(−/−) KnockoutMice. Ifosfamide (400 mg/kg, ip) induced a markedincrease in BWW in C57BL/6 WT (50% when com-pared to control group 12 h after ifosfamide admin-istration, p<0.05). In IL-4(−/−) knockout mice treatedwith ifosfamide, a significant increase of 44% in BWWwas seen when compared to C57BL/6 WT animalstreated with ifosfamide, and 94% when compared to thecontrol group (p<0.05) as shown in Fig. 1f.

Western Blot for iNOS and COX-2. In control animals,it was observed that both iNOS and COX-2 were presentin very low levels. However, after ifosfamide admin-istration, there was a significant rise in both iNOS (57%when compared to control group, p<0.05) and COX-2(125% when compared to the control group, p<0.05).IL-4 administered prior to ifosfamide caused a decreasein both iNOS and COX-2 levels, being most remarkablein the latter one, as seen in Fig. 2a, b.

Macroscopic and Microscopic Changes Induced by IL-4on Bladders of Animals Injected with Ifosfamide. Cystitisobserved 12 h after ifosfamide administration was charac-terized macroscopically by the presence of severe edema,receiving a score of 3 (3–3), and by marked hemorrhagewith mucosal hematomas and intravesical clots, receiving ascore of 3 (2–3), being significantly (p<0.05) different fromthe control group, which received scores of 0 (0–0) for bothedema and hemorrhage. Treatment with IL-4 in a dose-dependent manner reduced the intensity of cystitis (p<0.05),as indicated by Fig. 3 and Table 1. According to Gray’shistopathological criteria, 12 h after ifosfamide administra-tion, there was an obvious histological evidence of cystitis:extensive mucosal erosion with ulceration, fibrin deposition,hemorrhage, edema, and leukocyte infiltration of mainlylymphomononuclear cells and neutrophils, receiving a scoreof 2 (2–2), which were reduced with treatment 1 h beforeifosfamide with IL-4 (2 ng and 10 ng/animal).

Immunohistochemistry for IL-1β and TNF-α in UrinaryBladders. In control animals, IL-1β and TNF-α expressionswere not seen within urothelial cells. Twelve hours afterifosfamide administration, IL-1β and TNF-α expressionswere observed and characterized with a high immunoreac-tivity mainly in epithelial, parietal, and subepithelial cells andbeing significantly (p<0.05) different from the control group,

a b

Saline - 2 ng 10 ng

0.0

0.2

0.4

0.6

0.8

IL-4

**

# #

Ifosfamide (400 mg/kg)

CO

X-2

/β -A

ctin

rat

io

0.0

0.5

1.0

1.5

2.0

Saline - 2 ng 10 ng

IL-4

Ifosfamide (400 mg/kg)

**#

iNO

S/β

β -A

ctin

rat

io

iNOS

actin actin

COX2

Fig. 2. a Western blot for iNOS on ifosfamide-induced cystitis and effects of IL-4. b Western blot for COX-2 on ifosfamide-induced cystitis andeffects of IL-4. The results are reported as the relative density of iNOS/β-Actin and COX-2/β-Actin bands. *p<0.05 vs. control group (treated withsaline alone) and #p<0.05 when compared to vehicle (treated with ifosfamide alone) by ANOVA and Bonferroni’s tests.

302 Macedo, Mourão, Freitas, Lima-Júnior, Wong, Oriá, Vale, Brito, Cunha and Ribeiro

which received a score of 0 (0–0) in all cellular groupsanalyzed. Treatment 1 h before ifosfamide administrationwith IL-4 (10 ng ip) reduced the intensity of IL-1β and TNF-α expressions (p<0.05), mainly within parietal and sub-epithelial cells. (Fig. 4 and Tables 2 and 3, respectively).

DISCUSSION

IL-4 has a number of biological activities and anti-inflammatory effects by suppressing the production ofIL-1β, TNF-α, IL-8, and IFN-γ [14, 20] and through the

inhibition of the iNOS and COX-2 enzymes responsiblefor the production of nitric oxide and prostaglandins,respectively [21]. In the present study, we explored thepotential anti-inflammatory effect of IL-4 in the modelof ifosfamide-induced hemorrhagic cystitis.

Previous reports from our laboratory have shownthat proinflammatory cytokines such as platelet-activat-ing factor (PAF), TNF-α, and IL-1β have beenimplicated in hemorrhagic cystitis [10]. We also dem-onstrated the participation of interleukin-1β (IL-1β) andthe tumor necrosis factor-α (TNF-α) pathway in theinduction of nitric oxide (NO) production as animportant mediator of the urothelial lesion induced byalkylating agents. NO was demonstrated to be the finalmediator of urothelial damage and hemorrhage in thattype of hemorrhagic cystitis. TNF-α and IL-1β havebeen shown to be important mediators of NO synthesissince treatment with anti-TNF-α and anti-IL-1βdecreased the vesical damage as well the rise ofinducible NO synthase expression and activity [9].

It has also been shown that COX-2 is present inboth ifosfamide- and acrolein-induced hemorrhagiccystitis. Moreover, COX-2 expression seems to beregulated by TNF-α and IL-1β since blockade withTNF-α and IL-1β inhibitors decreased significantlyCOX-2 expression [11]. Recently, it has been shownthat interleukin-11 partially prevents ifosfamide-inducedexperimental hemorrhagic cystitis [22] by possiblyinhibiting TNF-α, IL-1β, and nitric oxide release [23].Another study showed that IL-4 is present after cyclo-sphophamide-induced hemorrhagic cystitis either acutelyor chronically [17]; however, it did not address the anti-inflammatory mechanisms of IL-4.

In the current study, we have shown that IL-4production is enhanced after ifosfamide administrationwhen compared to control bladders. In a study

C Saline 0,4 ng 2 ng

Ifosfamide (400 mg/kg)

10 ng

IL 4

Fig. 3. Macroscopic analysis of representative bladders. C bladder ofanimals treated only with saline. No evidence of edema or hemorrhagewas seen, saline bladder of animals treated with ifosfamide only(400 mg/kg) 12 h before, showing a significant increase in size secondaryto edema and diffuse hemorrhage, IL-4 animals treated with interleukin-4in different doses 1 h before ifosfamide administration. A visibledose-dependent effect was noticed wherein animals treated with 2 ng or10 ng exhibited less hemorrhage and edema. The pictures were takenwith a 7 MP camera with a circular flash attached, macro mode andlens 1 cm away from the actual specimens.

Table 1. Macroscopic and Microscopic Alterations in Ifosfamide-Induced Hemorrhagic Cystitis and the Effect of IL-4 in Different Doses

Experimental groups Macroscopic analyses (edema) Macroscopic analyses (hemorrhage) Microscopic analyses

Control 0 (0–0) 0 (0–0) 0 (0–0)Ifosfamide (400 mg/kg) 3 (3–3)* 3 (2–3)* 2 (2–2)*IL-4 (0.4 ng) 2 (1–3) 2,5 (1–3) 2 (2–2)IL-4 (2 ng) 2 (1–2)** 2 (1–3)** 1 (1–2)**IL-4 (10 ng) 1 (1–2)** 2 (1–3)** 1 (1–2)**MMMa 0 (0–2)** 0 (0–2)** 0 (0–1)**

Results are reported as median and range (n=6/group)*p<0.05 when compared to the control group; **p<0.05 when compared to the ifosfamide groupa Treated animals with three doses of mesna using the classical protocol

303Interleukin-4 in Ifosfamide-Induced Hemorrhagic Cystitis

Fig. 4. Imunohistochemistry for TNF-α and IL-1β. a Bladder of animals treated only with saline and imunohistochemistry for TNF-α. b Bladder ofanimals treated with ifosfamide. Arrows showing positivity for TNF-α underneath the epithelium, mainly within submucosal cells. c Pretreatment withIL-4 showing attenuation of positive TNF-α cells after ifosfamide administration. d Bladder of animals treated only with saline. e Bladder of animalstreated with ifosfamide and imunohistochemistry for IL-1β. Arrows showing positivity for IL-1β underneath the epithelium, mainly within submu-cosal cells. f Pretreatment with IL-4 showing attenuation of positive IL-1β cells after ifosfamide administration.

304 Macedo, Mourão, Freitas, Lima-Júnior, Wong, Oriá, Vale, Brito, Cunha and Ribeiro

published by Malley and Vizzard [17], a significantalteration in urinary bladder cytokine mRNA forseveral cytokines in the bladder was demonstratedfollowing the administration of cyclophosphamide,another oxazaphosphorine similar to ifosfamide. Theseauthors showed an acute increase in urinary bladdercytokine mRNA for several factors examined, includ-ing IL-1β, IL-2, IL-4, IL-6, and TNF-α/β. Thecytokine levels gradually decreased 48 h and 10 daysafter cyclophosphamide-induced cystitis. In addition tothat, a significant increased level in proinflammatorycytokines, such as IL-1β, was observed even in chronicconditions. The same authors also observed that nochanges in urinary bladder IL-4 protein expressionwere observed at any time point examined after cyclo-phosphamide-induced cystitis compared with totalcontrol urinary bladder IL-4. Opposingly, Smaldone etal. [24] found a twofold increase in IL-4 tissue levels at24 h after cyclophosphamide injection. The importanceof each cytokine to the lesion in the urinary bladderwas not addressed by both observations, whereas it wassuggested that cytokines produced in the urinarybladder, alone or in combination with other cytokines,would play a pathophysiological role in cyclophospha-mide-induced cystitis.

In the present study, we showed that the intra-peritoneal administration of exogenous IL-4 inhibited, ina dose-dependent manner, and attenuated the inflamma-tory response to ifosfamide, demonstrated by thedecrease in both edema and hemorrhage. In addition,macroscopic and microscopic damages were both pre-vented with pretreatment with IL-4. As we have

previously elucidated the cascade of inflammatorymediators involved in the pathogenesis of experimentalifosfamide- and acrolein-induced hemorrhagic cystitis,we now hypothesized the participation of IL-4 as adownregulatory mediator of such inflammatory events.

In fact, in the present study, we showed that theanti-inflammatory effect of exogenous IL-4 is likelyto be a consequence of its capacity to inhibit theexpression of proinflammatory cytokines (TNF-α andIL-1β) and enzymes (iNOS and COX-2). Two linesof evidence support our observation: (1) the detec-tion of iNOS and COX-2 through western blot wassignificantly attenuated by IL-4 treatment and (2) theimmunohistochemistry detection of TNF-α and IL-1β was also significantly inhibited when IL-4 wasadministered.

Additionally, we also observed that in knockoutanimals for both genes of IL-4 (IL-4−/−) and in wild-type animals pretreated with antiserum for IL-4, theinflammatory response in ifosfamide-induced hemor-rhagic cystitis was significantly intensified, a verystrong clue that the endogenous production of anti-inflammatory cytokines like IL-4 seems to play animportant role in attenuating the inflammatoryresponse due to ifosfamide-induced injury in thebladder. Wild-type animals that received exogenousIL-4 injection presented a marked inhibition of thevesical lesion, suggesting that ifosfamide triggerssuch a strong inflammatory response that overtakesendogenous IL-4 inhibitory effects. Even thoughthere are no reports comparing the effects ofifosfamide-induced hemorrhagic cystitis in different

Table 2. Immunohistochemistry for IL-1β and the Effects of IL-4 (10 ng, ip)

Experimental groups Urothelium Submucosa Basal membrane

Control 0 (0–0) 0 (0–0) 0 (0–0)Ifosfamide (400 mg/kg) 3 (2–3)* 3 (3–3)* 3 (3–3)*IL-4 (10 ng) 3 (1–3) 2 (0–2)** 2 (1–3)**

Results are reported as median and range (n=6/group)*p<0.05 when compared to the control group; **p<0.05 when compared to the ifosfamide group

Table 3. Immunohistochemistry for TNF-α and the Effects of IL-4 (10 ng, ip)

Experimental groups Urothelium Submucosa Basal membrane

Control 0 (0–0) 0 (0–0) 0 (0–0)Ifosfamide (400 mg/kg) 3 (3–3)* 3 (3–3)* 3 (2–3)*IL-4 (10 ng) 2,5 (2–3) 1 (1–3)** 1 (1–3)**

Results are reported as median and range (n=6/group)*p<0.05 when compared to the control group; **p<0.05 when compared to the ifosfamide group

305Interleukin-4 in Ifosfamide-Induced Hemorrhagic Cystitis

mice, it has been observed that these animals responddifferently to ifosfamide. The C57BL/6mice are moreresistant to ifosfamide toxicity when compared to theSwiss mice, which have a 100% mortality withifosfamide 800 mg/kg ip even with free access towater. For the C57BL/6mice, using the same doseand water deprivation, this mortality is around 40–50%. The reason why we used two different strainswas to compare if we could reproduce the inhibitoryeffect of IL-4 using the knockout methodology.

In contrast to IL-4 doses of 2 and 10 ng, a doseof 0.4 ng injected 12 h before ifosfamide was notable to inhibit cystitis. Inhibition was observed onlywhen IL-4 was injected in higher doses. A moreevident response was seen with 10 ng. IL-4 at thisdose inhibited TNF-α, IL-1β, iNOS, and COX-2expressions significantly. One possible explanationfor such observation is that IL-4 dose-dependentlywould be blocking iNOS or COX-2 expressionsindirectly through TNF-α and/or IL-1β inhibition.Along with that suggestion, one study published byNishisaka et al. [25] evidenced an inhibitory effectof IL-4 in NO release enhanced with IL-1 inchondrocyte culture. According to another studypublished by Hart and coworkers [26], increasingconcentrations of recombinant human IL-4 suppress,in a dose-dependent manner, macrophage productionof IL-Iβ, TNF-α, and PGE2, suggesting the anti-inflammatory properties of IL-4. However, the pro-tective mechanism of IL-4 in our model of hemor-rhagic cystitis merits further investigation.

CONCLUSION

In conclusion, the data described in this studyprovide evidence that IL-4 at least partially modu-lates negatively the inflammatory process seen inifosfamide-induced hemorrhagic cystitis. Along withthat, the lack of IL-4, as observed in knockout miceand after antiserum against IL-4 in wild-type ani-mals, may contribute to the potentiating inflammatoryeffect on this model of chemotherapy toxicity. IL-4appears to achieve its effects by inhibiting theproduction of inflammatory cytokines (e.g., TNF-αand IL-1β), the expression of iNOS and COX-2, andconsequently, the production of NO and prostaglan-dins, which are crucial mediators in ifosfamide-induced hemorrhagic cystitis [9–11].

ACKNOWLEDGMENTS

The authors gratefully acknowledge the technicalassistance of Maria Silvandira Freire, José Ivan Rodrigues,and Jand Venes Medeiros from FM-UFC-CE and ArturoRivera from University of Miami, USA and also wouldlike to thank the National Research Council of Brazil(CNPq) and Fundação Cearense de Amparo a Pesquisa(FUNCAP), Ceará-Brazil, for the financial support.

Financial Support. This project has received grantsfrom National Research Council of Brazil (CNPq),CAPES, and Fundação Cearense de Amparo àPesquisa (FUNCAP), Ceará-Brazil.

Conflict of Interests. None.

REFERENCES

1. Higgs, D., C. Nagy, and L.H. Einhorn. 1989. Ifosfamide: a clinicalreview. Seminars in Oncology Nursing 5: 70–77.

2. Dechant, K.L., R.N. Brogden, T. Pilkington, et al. 1991. Ifosfamide/mesna. A review of its antineoplastic activity, pharmacokineticproperties and therapeutic efficacy in cancer. Drugs 42: 428–467.

3. Macedo, F.Y., F. Baltazar, L.C. Mourao, et al. 2008. Induction ofCOX-2 expression by acrolein in the rat model of hemorrhagiccystitis. Experimental and Toxicologic Pathology 59: 425–430.

4. Brock, N., and J. Pohl. 1983. The development of mesna forregional detoxification. Cancer Treatment Reviews 10: 33–43.

5. Scheulen, M.E., N. Niederle, K. Bremer, et al. 1983. Efficacy ofifosfamide in refractory malignant diseases and uroprotection bymesna: results of a clinical phase II study with 151 patients. CancerTreatment Reviews 10: 93–101.

6. Shulman, L.N. 1993. The biology of alkylating-agent cellular injury.Hematology/Oncology Clinics of North America 7: 325–335.

7. Vieira, M.M., G.A. Brito, J.N. Belarmino-Filho, F.Y. Macedo, et al.2003. Use of dexamethasone with mesna for the prevention ofifosfamide-induced hemorrhagic cystitis. International Journal ofUrology 10: 595–602.

8. Lima, M.V.A., F.V. Ferreira, F.Y.B. Macedo, and R.A. Ribeiro.2007. Histological changes in bladders of patients submitted toifosfamide based chemotherapy even with mesna prophylaxis.Cancer Chemotherapy and Pharmacology 59: 643–650.

9. Ribeiro, R.A., H.C. Freitas, M.C. Campos, et al. 2002. Tumornecrosis factor-α and interleukin-1 β mediate the production ofnitric oxide involved in the pathogenesis of ifosfamide inducedhemorrhagic cystitis in mice. Journal d’Urologie 176: 2229–2234.

10. Souza-Filho, M.V.P., M.V.A. Lima, M.M.L. Pompeu, et al. 1997.Involvement of nitric oxide in the pathogenesis of cyclophospha-mide-induced hemorrhagic cystitis. The American Journal ofPathology 150: 247–256.

306 Macedo, Mourão, Freitas, Lima-Júnior, Wong, Oriá, Vale, Brito, Cunha and Ribeiro

11. Macedo, F.Y., F. Baltazar, P.R. Almeida, F. Távora, F.V. Ferreira,F.C. Schmitt, G.A. Brito, and R.A. Ribeiro. 2008. Cyclooxyge-nase-2 expression on ifosfamide-induced hemorrhagic cystitis inrats. Journal of Cancer Research and Clinical Oncology 134: 19–27.

12. Lord, C.J.M., and J.R. Lamb. 1996. TH2 cells in allergicinflammation: a target of immunotherapy. Clinical and Exper-imental Allergy 26: 756–765.

13. Tunon de Lara, J.M., Y. Okayama, A.R. Mceuen, et al. 1994.Release and inactivation of interleukin-4 by mast cells. Cells andcytokines in lung inflammation. Annals of the New York Academyof Sciences 725: 50–58.

14. Vannier, E., M.C. Miller, and C.A. Dinarello. 1992. Coordi-nated anti-inflammatory effects of interleukin4: interleukin 4suppresses interleukin 1 production but up-regulates geneexpression and synthesis of interleukin 1 receptor antagonist.Proceedings of the National Academy of Science 89: 4076–4080.

15. Seitz, M., P. Loetscher, B. Dewald, H. Towbin, M. Ceska,and M. Baggiolini. 1994. Production of interleukin-1 receptorantagonist, inflammatory chemotactic proteins, and prosta-glandin E by rheumatoid and osteoarthritic synoviocytes-regulation by IFN-γ and IL-4. Journal of Immunology 152:2060–2065.

16. Endo, T., F. Ogushi, and S. Saburo. 1996. LPS-dependentcyclooxygenase-2 induction in human monocytes is down-regulated by IL-13, but not by IFN-γ. Journal of immunology156: 2240–2246.

17. Malley, S.E., and M.A. Vizzard. 2002. Changes in urinary bladdercytokine mRNA and protein after cyclophosphamide-inducedcystitis. Physiological Genomics 9: 5–13.

18. Gray, K.J., U.H. Engelmann, E.H. Johnson, et al. 1986.Evaluation of misoprostol cytoprotection of the bladder with

cyclophosphamide (Cytoxan) therapy. Journal d'Urologie 133:497–500.

19. Hsu, S.M., and L. Raine. 1981. Protein A, avidin, and biotin inimmunohistochemistry. The Journal of Histochemistry and Cyto-chemistry 29: 1349–1353.

20. Fenton, M.J., J.A. Buras, and R.P. Donelly. 1992. IL-4 reciprocallyregulates IL-1 and IL-1 receptor antagonist expression in humanmonocytes. Journal of Immunology 149: 1283–1288.

21. Cunha, F.Q., S. Moncada, and F.Y. Liew. 1992. Interleukin-10 (IL-10) inhibits the induction of nitric oxide synthase by interferon-gamma in murine macrophages. Biochemical and BiophysicalResearch Communications 182: 1155–1159.

22. Mota, J.M., G.A. Brito, R.T. Loiola, F.Q. Cunha, and R.A.Ribeiro. 2007. Interleukin-11 attenuates ifosfamide-inducedhemorrhagic cystitis. International Brazilian Journal of Urology33: 704–710.

23. Trepicchio, W.L., M. Bozza, G. Pedneault, and A.J. Dorner.1996. Recombinant human IL-11 attenuates the inflammatoryresponse through down-regulation of proinflammatory cytokinerelease and nitric oxide production. Journal of Immunology 157:3627–3634.

24. Smaldone, M.C., Y. Vodovotz, V. Tyagi, D. Barclay, B.J. Philips,N. Yoshimura, M.B. Chancellor, and P. Tyagi. 2009. Multiplexanalysis of urinary cytokine levels in rat model of cyclophospha-mide-induced cystitis. Urology 73(2): 421–426.

25. Nishisaka, F., S. Sohen, H. Fukuoka, Y. Okamoto, M. Matukawa,et al. 2001. Interleukin-4 reversed the interleukin-1-inhibitedproteoglycan synthesis through the inhibition of NO release: apossible involvement of intracellular calcium ion. Pathophysiology7(4): 289–293.

26. Hart, P.H., R.L. Cooper, and J.J. Finlay-Jones. 1991. IL-4suppresses IL-1 beta, TNF-alpha and PGE2 production by humanperitoneal macrophages. Immunology 72(3): 344–349.

307Interleukin-4 in Ifosfamide-Induced Hemorrhagic Cystitis