Macrophage Migration Inhibitory Factoris Increased in the Urine of Patients WithUrinary Tract Infection: Macrophage MigrationInhibitory Factor-Protein Complexes in Human UrineKatherine L. Meyer-Siegler,* Kenneth A. Iczkowski and Pedro L. VeraFrom the Research and Development Service, Bay Pines Veterans Affairs Medical Center (KLMS, PLV), Bay Pines, Department ofSurgery, University of South Florida (KLMS, PLV), Tampa and Pathology Service, Malcolm Randall VA Medical Center (KAI),Gainesville, Florida

Purpose: MIF is a proinflammatory cytokine present in preformed stores in human urothelium. In animal models of bladderinflammation, including bacterial cystitis, MIF is up-regulated in the bladder and released from the bladder as a highmolecular weight complex. We compared urine MIF amounts in patients with UTI to that in patients without UTI, and weexamined and identified MIF-protein complexes in urine.Materials and Methods: Using enzyme-linked immunosorbent assay we compared MIF levels in the urine of 14 patientswith UTI to levels in 16 controls with no UTI. Western blotting under native, denaturing and reducing conditions was doneto examine MIF complexes found in urine. Mass spectrometry identified MIF associated proteins in urine, while co-immunoprecipitation confirmed the associations.Results: Mean urine MIF amounts � SEM determined by enzyme-linked immunosorbent assay were significantly greater in14 patients with UTI compared to that in 16 controls (1.96 � 0.40 vs 0.59 � 0.09 ng/mg creatinine, p �0.01). Western blottingunder denaturing conditions showed several high molecular weight complexes (100 to 165 kDa) that increased in UTI urineas well as typical, monomeric MIF (12 kDa). Mass spectrometry identified associated MIF proteins, including ceruloplasmin,albumin and uromodulin. Co-immunoprecipitation confirmed mass spectrometry findings and also identified MIF interactionwith �-2-macroglobulin.Conclusions: Increased urine MIF amounts in patients with bacterial cystitis support our experimental evidence showinga role for MIF in pelvic visceral inflammation. The novel finding of an association of MIF with other urine proteins suggestthat the physiologically relevant form of MIF may be an MIF-protein complex.

Key Words: bladder, urinary tract infections, macrophage migration-inhibitory factors, cystitis, inflammation

UTIs vary in pathogenesis and severity. However, themajority of UTIs are caused by gram-negative bac-teria, in particular Escherichia coli, which is respon-

sible for 70% to 90% of all UTIs.1 UTIs and the associatedinflammatory response are mediated by host recognition ofspecific pathogen produced molecules, such as LPS, a majorcomponent of the gram-negative bacterial cell wall.2 Afterthey are recognized by the host, these microbe specific mol-ecules elicit a rapid response. In the bladder interactionbetween LPS and toll-like receptor 4 expressed on urothelialumbrella cells initiates the innate immune response induc-tion and subsequent release of proinflammatory mediators.3

MIF is a ubiquitous proinflammatory cytokine implicatedas an inducer and modulator of various inflammatory disor-ders, and a central regulator of the innate immune system.4

Submitted for publication June 16, 2005.Study received Bay Pines Veterans Affairs Medical Center Insti-

tutional Review Board approval.Supported by the Veterans Affairs Merit Review Program and the

Bay Pines Foundation (KLM-S and PLV) and VA MREP (KAI).* Correspondence: Bay Pines Veterans Affairs Medical Center,

10,000 Bay Pines Blvd., Bay Pines, Florida 33744 (telephone: 727-

398-6661, extension 5236; FAX: 727-398-9467; e-mail:Katherine.Siegler@med.va.gov).

0022-5347/06/1754-1523/0THE JOURNAL OF UROLOGY®

Copyright © 2006 by AMERICAN UROLOGICAL ASSOCIATION

1523

MIF affects the host response to LPS by modulating toll-likereceptor 4 amounts.5 We have previously documented that1) urothelium is a rich source of preformed MIF,6 2) intralu-minal LPS induces MIF up-regulation in the rat bladder,7 3)experimental bladder inflammation results in MIF releaseinto the intraluminal fluid8 and 4) sequestering releasedMIF decreases bladder inflammation, suggesting that in-traluminal release of urothelial MIF induces proinflamma-tory responses associated with induction of the innateimmune response.9

MIF proinflammatory effects have been well described.10

However, baseline MIF amounts are found in the serum andurine of normal patients without underlying inflammatoryconditions.10,11 A number of possibilities that are not mutu-ally exclusive may explain the fact that substantial amountsof MIF are present in the serum and urine of normal indi-viduals. 1) MIF amounts may need to attain a certainthreshold to initiate proinflammatory effects. Increased MIFlevels have been reported in a number of inflammatory10,11

and neoplastic12 conditions, supporting the view that in-creased MIF levels may be necessary for proinflammatoryeffects. 2) MIF receptor, recently identified as CD74,13 is

inactivated or present in decreased amounts under nonin-

Vol. 175, 1523-1528, April 2006Printed in U.S.A.

DOI:10.1016/S0022-5347(05)00650-6

MACROPHAGE MIGRATION INHIBITORY FACTOR IN CYSTITIS1524

flammatory conditions and, thus, it is unable to initiate theproinflammatory effects of MIF. We recently reported up-regulation of CD74 in the rat bladder and urothelium afterbladder inflammation, lending support to this theory.14 3)MIF may exist in different forms, eg active vs inactive, thatmediate proinflammatory effects. To our knowledge thephysiologically relevant form of MIF is unknown. We re-cently observed that during inflammation MIF is releasedinto the rat bladder lumen as a high molecular weight com-plex including acute phase proteins, which may modulateMIF activity.15 Additionally, most human serum MIF is in ahigh molecular weight complex (180 kDa).16 Therefore, MIFtypes and changes in MIF complexes associated with inflam-matory conditions may be the key to determining the biolog-ically relevant MIF form(s) and may provide insight into theMIF mechanism of action.

Based on our experimental findings in animal models ofcystitis we hypothesized that MIF may increase in the hu-man bladder in response to LPS and, thus, have an impor-tant role in bacterial cystitis. In addition, we proposed thatMIF is released in the urine complexed to other proteins.Therefore, we studied urine MIF amounts associated withbacterial cystitis by analyzing urine samples from patientswith and without documented UTIs, and we identified pro-teins that interact with MIF in human urine.

MATERIALS AND METHODS

This study was approved by the Bay Pines Veterans AffairsMedical Center Institutional Review Board. Urine sampleswere obtained with waiver of informed consent under sec-tion 46.116(d) of the Department of Health and HumanServices human subject regulations at 45 CFR 46. All urinespecimens were analyzed by ELISA and Western blot.

Comparison of MIF Urine Amounts in Patients Withand Without Bacterial CystitisA total of 30 urine samples remaining following routineevaluation were randomly collected and stored at �80C withthe addition of a general protease inhibitor cocktail (SigmaChemical Co., St. Louis, Missouri). The greater proportion ofmale urine samples collected reflects the patient population,which was 94.5% male. UTI in 9 males and 5 females wasdetermined by positive bacterial cultures (greater than 105

cfu/ml). Ten male and 6 female patients without UTI hadnormal urine chemistry negative for nitrites and leukocytes,and no documented chronic inflammatory diseases. Patientswith bladder or prostate cancer were excluded. MIFamounts in urine samples (100 �l) were determined using aDuoSet™ ELISA system. MIF amounts were normalized tototal urine creatinine and data are expressed as the meanMIF ng/mg creatinine � SEM.

MIF-Protein Complexes in Urine of Patients Withand Without Bacterial Cystitis: Western BlottingAnalysis and Mass SpectroscopyUrine proteins (30 �g) were separated by native, denaturingand reducing gel electrophoresis conditions according tomanufacturer protocols (tris-glycine or NuPAGE® Bis-Trisgels) and transferred to PVDF. MIF specific bands weredetected and quantified using a biotinylated polyclonal an-

tibody (BAF289, R and D Systems, Minneapolis, Minnesota),

as described previously.15 WB intensity was determined byKodak Image Station digital imaging (Kodak, Rochester,New York). Data are expressed as mean MIF band intensi-ty/mg creatinine � SEM.

Duplicate urine samples from 3 patients with UTI wereseparated by denaturing NuPAGE® electrophoresis on thesame gel with 1 set of samples analyzed by WB and itsduplicate stained with Coomassie stain (Bio-Rad Laborato-ries, Hercules, California). Coomassie stained MIF complexbands were excised and subjected to mass spectroscopy anal-ysis (Protein Chemistry Core Laboratory, University of Flor-ida, Gainesville, Florida), as previously described.15

Co-Immunoprecipitation of MIF Associated ProteinsProtein complexes were isolated from urine samples (3.5 mgtotal protein) with appropriate primary polyclonal antibodyat a concentration of 5.0 �g/ml (anti-MIF, AB-289-PB, R andD Systems and anti-�-2-macroglobulin, sc-8514, Santa CruzBiotechnology, Santa Cruz, California) using Protein G aga-rose beads (Kirkegaard and Perry Laboratories, Gaithers-burg, Maryland) according to the manufacturer protocol.Controls included incubation of precleared samples withnonspecific IgG (Sigma Chemical Co.) and samples pro-cessed without the addition of antibody in the immunopre-cipitation reaction. Precipitated complexes were separatedby electrophoresis under reducing conditions and bandswere detected as described previously.15 Following immuno-precipitation with anti-MIF antibody ceruloplasmin (anti-human ceruloplasmin, G37, Biomeda, Foster City,California) and uromodulin (anti-human uromodulin, BT-590, Biomedical Technologies, Stoughton, Massachusetts)were detected by WB. MIF was detected following �-2-mac-roglobulin immunoprecipitation using biotinylated MIFpolyclonal antibody.

Statistical AnalysesStatistical analysis was performed using Prism 4 (GraphPadSoftware, San Diego, California). Regression with interpola-tion was used to evaluate MIF amounts by ELISA. ELISA orWB results were compared by the Student t tests with sig-nificance considered at p �0.05. Data are reported as themean � SEM.

RESULTS

Bacterial Cystitis Induced MIF ComplexesMIF WB of urine proteins from patients with UTI separatedunder native conditions resulted in a prominent high molec-ular weight band in the range of 150 to 500 kDa (fig. 1). Aless intense band at the same molecular weight range wasobserved on WB of urine from patients without UTI. Undernative WB conditions a 12 kDa band indicative of mono-meric MIF was not detected (fig. 1). However, under dena-turing conditions WB of the same sample resulted in 3prominent high molecular weight bands of 165, 135 and 100kDa, respectively, as well as the 12 kDa monomeric MIFband (fig. 1). MIF staining intensity was greater in samplesfrom patients with UTI. Under reducing conditions only asingle MIF staining band at 12 kDa was evident in all urinesamples. MIF band staining intensity was greatest in urine

samples from patients with UTI (fig. 1).

MACROPHAGE MIGRATION INHIBITORY FACTOR IN CYSTITIS 1525

Thus, in agreement with our earlier findings in the rat15

MIF in human urine is found in high molecular weightcomplexes. In all urine specimens a significant portion(65.7% � 4.5%) of monomeric MIF (12 kDa) interacted withother molecules via noncovalent bonds since denaturing con-ditions revealed a prominent 12 kDa MIF band. Finally,reducing conditions showed that the higher molecularweight bands observed in urine at 165, 135 and 100 kDainteracted with MIF through covalent bonds since thesecomplexes were disrupted only in conditions that allow thedissociation of disulfide bonds.

Identification of Proteins Complexed With MIFWe have previously identified acute phase proteins as com-ponents of high molecular weight MIF complexes in a ratcystitis model.15 Denaturing polyacrylamide gel electro-phoresis MIF WB analysis of human urine revealed MIFcomplexes at 165, 135 and 100 kDa (figs. 1 and 2, A). Thesehigh molecular weight bands were excised from Coomassiestained gels, analyzed by mass spectroscopy and identifiedas the acute phase proteins ceruloplasmin and albumin aswell as the predominant urine protein uromodulin (fig. 2, B).

Co-immunoprecipitation studies using anti-MIF followedby WB using antiserum to proteins identified by mass spec-troscopy confirmed the interaction of MIF with ceruloplas-min and uromodulin (fig. 3, A). MIF immunoprecipitationwith ceruloplasmin WB analysis identified 2 bands at 135and 64 kDa, respectively, while uromodulin WB of MIFimmunoprecipitation revealed an intense band at 86 kDa.Mass spectroscopy data based algorithms also identifiedcomplement factor H and C-reactive protein as potentialMIF binding partners. No MIF complexes with either ofthese 2 proteins were identified, confirming specific MIFinteraction with ceruloplasmin and uromodulin (data notshown).

We expected to identify MIF interaction with proteins inthe �-2-macroglobulin superfamily since we have previously

FIG. 1. MIF in 30 urine samples obtained after routine urinalysisand separated under native, denatured and reduced conditions be-fore Western blot analysis for MIF. Note representative bandingpattern observed with nonUTI (N1 and N2) and UTI (U1 and U2)urine samples under 3 electrophoresis conditions. Under nativeconditions broad band between 150 and 500 kDa was detected.Denaturing conditions resulted in 4 prominent bands of 165, 135,100 and 12 kDa, respectively. Under reducing conditions single 12kDa band was observed.

identified interaction between MIF and �-1-inhibitor-3, a

rodent specific �-2-macroglobulin superfamily protein.15

�-2-Macroglubulin WB of proteins co-immunoprecipitatedwith MIF failed to show �-2-macroglobulin. However, MIFWB of �-2-macroglobulin precipitated protein identified afaint 12 kDa band (fig. 3, B), indicating that these 2 proteinsinteract in human urine.

Comparison of MIF Urine Amounts in Patients Withand Without Bacterial CystitisELISA detected significantly increased MIF amounts in theurine of patients with UTI compared to that in patientswithout UTI (1.96 � 0.40 vs 0.59 � 0.09 ng/mg creatinine,p �0.01, fig. 4, A). In all samples creatinine was within thereported normal range (10 to 300 mg/dl) with mean creati-nine equal to 115.0 � 21.27 and 110.3 � 8.46 mg/dl in thenonUTI and UTI groups, respectively. In addition, analysisof specific bands in WB under denaturing conditions showedthat UTI urines significantly increased in the ceruloplasmin(4.8-fold increase in UTI samples) and in the 12 kDa MIFmonomeric band (0.7-fold increase in UTI when normalizedto total creatinine) (each p �0.01, fig. 4, B).

FIG. 2. Identification of urine MIF complex proteins. A, Coomassieand MIF Western blot of urine separated by denaturing polyacryl-amide gel electrophoresis shows molecular weight of correspondingCoomassie stained bands (values on left) and size of molecularweight markers (values on right). Coomassie stained bands corre-sponding to bands detected by Western blot were excised from gels.B, Excised Coomassie stained bands were analyzed by mass spec-troscopy to determine MIF complex band. Band, apparent molecu-lar weight calculated by polyacrylamide gel electrophoresis.Reference Number, National Center for Biotechnology Informationreference number for identified proteins. Protein (MW) Sample Pep-tide, identified protein and its corresponding molecular weight(MW) with mass spectroscopy identified peptide included for refer-ence. Score, probability based Mowse score (molecular weight

search). Peptides Matched, number of peptides matched that iden-tified protein.

MACROPHAGE MIGRATION INHIBITORY FACTOR IN CYSTITIS1526

DISCUSSION

MIF, a 12 kDa protein, is a proinflammatory cytokine associ-ated with the etiology of various inflammatory diseases.10,11

Increased urine MIF amounts are associated with urologicalinflammatory disease.17,18 The current study showed in-creased levels of MIF in the urine of patients with UTI andidentified several proteins that formed high molecular weightcomplexes with MIF in urine. Under denaturing conditionsthese large complexes resolve into 3 prominent bands of 165,135 and 100 kDa (and 12 kDa monomer MIF), which wereidentified by mass spectroscopy as ceruloplasmin, albumin anduromodulin, respectively, and confirmed by co-immunoprecipi-tation studies, followed by WB with the appropriate antiserum.

Our current findings confirm earlier observations in therat by showing that MIF is associated with other proteins,including acute phase proteins, in human urine. Althoughthese complexes were observed under denaturing condi-tions, they dissociated under reducing conditions to revealonly a single 12 kDa (monomeric) MIF band. Therefore, high

FIG. 3. Urine protein co-immunoprecipitation. A, urine proteinswere immunoprecipitated with MIF antibody, separated by reduc-ing polyacrylamide gel electrophoresis, transferred to PVDF andceruloplasmin or uromodulin detected using polyclonal antibody.Arrows indicate location of IgG heavy or light chains. B, urineproteins were immunoprecipitated with �-2 macroglobulin anti-body, separated by polyacrylamide gel electrophoresis, transferred

to PVDF and MIF detected using polyclonal antibody. Arrows indi-cate location of IgG light chains.

molecular weight complexes seen under denaturing condi-tions (eg MIF-ceruloplasmin) represent associations medi-ated by covalent interactions, whereas 12 kDa MIF observedunder denaturing but not native conditions represents mo-nomeric MIF associations with protein complexes throughnoncovalent interactions.

Typically MIF WB analysis is performed under reduc-ing conditions,10 which detects levels of dissociated MIFin samples but likely ignores different complexes formedby MIF or changes in these complexes that may provephysiologically significant. Although they were not de-scribed until recently,6,15 we predict that MIF-proteininteractions may be ubiquitous in physiological systems.In fact, most human serum MIF is in a high molecularweight complex (180 kDa).16

The physiological significance of MIF complexes was notinvestigated in the current study. However, the associationof MIF with several identified proteins in human urine rep-resents a novel finding and suggests that MIF activity maybe modulated by interaction with other proteins. Under de-naturing conditions revealing multiple MIF bands the levelsof MIF-ceruloplasmin and monomeric (12 kDa, nonco-valently bound) MIF were increased in patients with UTI.Ceruloplasmin is a class I positive acute phase protein, ofwhich the serum concentration increases by up to 50% dur-ing inflammation.19 MIF was also found in complexes withuromodulin and albumin but these complexes were notchanged by UTI.

To our knowledge the mechanism whereby MIF formscomplexes with other proteins is unknown. In vitro cultured

FIG. 4. Urine MIF amounts. A, ELISA in all 100 �l urine samplesidentified significant increase in total urine MIF in patients with vswithout UTI. Data are reported as ng/mg creatinine � SEM asdetected under standard ELISA conditions. B, Western blot showsMIF immunoreactive bands from denaturing gels from all urinesamples quantified by digital imaging by determining total inten-sity of bands corresponding to ceruloplasmin, uromodulin and albu-min, and total intensity of 12 kDa band. Total intensity of eachurine band was divided by total creatinine in mg loaded per well.Data are reported as band intensity per mg creatinine � SEM in 16patients without and in 14 with UTI. Asterisks indicate p �0.01.

bladder epithelial cells synthesize and release 12 kDa MIF

MACROPHAGE MIGRATION INHIBITORY FACTOR IN CYSTITIS 1527

(unpublished data).6 Thus, it is likely that noncovalentlybound MIF represents urothelial MIF that is increased dur-ing UTI and may form complexes in the bladder lumen. Infact, MIF up-regulation was observed in the bladder duringexperimental bacterial cystitis.7 High molecular weight MIFcomplexes may represent 1) covalent interactions occurringin the bladder lumen between 12 kDa MIF released by theurothelium and urine proteins, 2) kidney clearance of MIF-protein complexes, eg increased serum ceruloplasmin as aresult of inflammation,19 or 3) MIF-protein complexes re-leased by the bladder as a result of edema and plasmaprotein extravasation during inflammation. Although wecould not rule out the first 2 possibilities in the currentstudy, in animals with the bladder isolated from the kidneyswe still observed an increase in MIF-protein complexes re-leased into the lumen as a result of neurogenic inflamma-tion.15 Therefore, it is likely that edema and plasma proteinextravasation may result in the release of MIF-protein com-plexes into the lumen.

Our previous experiments showed that MIF is associatedwith �-1-inhibitor 3, a rat homolog of �-2-macroglobulin.15

The current study establishes that MIF complexes with hu-man �-2-macroglobulin. The physiological significance ofthis complex is unclear. Proteinase inhibitors are a majorclass of positive acute phase proteins that function to controlthe activity of proteinases released during the inflammatoryresponse.19 An additional function of many proteinase inhib-itors is as carriers of various growth factors, cytokines andhormones, which bind through diverse mechanisms. Espe-cially noteworthy in this regard is the �-2-macroglobulinfamily of protease inhibitors, in which investigation of othercytokines, such as nerve growth factor-�, platelet-derivedgrowth factor-�� and transforming growth factor-�, has doc-umented that interaction with �-2-macroglobulin regulatescytokine activity.20 The current study documents interac-tion between MIF and �-2-macroglobulin in urine. However,MIF bound to �-2-macroglobulin was not immunoprecipi-tated with an MIF specific polyclonal antibody, so that it islikely that this interaction blocks MIF antibody binding.Alternatively the MIF-�-2-macroglobulin complex may be atconcentrations low enough that make co-immunoprecipita-tion with anti-MIF difficult to detect.

CONCLUSIONS

Increased urine MIF correlates positively with bacterial UTIdiagnosis and supports our experimental data in animalmodels, in which MIF is released into the intraluminal spacedue to bladder inflammation to exert proinflammatory ef-fects on the bladder. In urine MIF was complexed to anumber of acute phase proteins, including ceruloplasmin,albumin and �-2 macroglobulin as well as the predominanturine protein, uromodulin. The physiologically relevantform(s) of MIF in urine and the functional significance ofMIF protein complexes remain to be determined. However,as we have documented previously, it is not exclusively free,uncomplexed MIF (monomer, dimer or trimer) since theseforms are not found under native conditions. The biochemi-cal nature of the binding of MIF to these proteins, the formof the protein that MIF binds and the physiological conse-quences of this binding are currently under investigation.Together these data suggest that MIF interaction with var-

ious proteins modulates the activity of this cytokine.

ACKNOWLEDGMENTS

Michael Bellino and Gary A. Smith, Jr. provided technicalassistance.

Abbreviations and Acronyms

ELISA � enzyme-linked immunosorbent assayLPS � lipopolysaccharideMIF � macrophage migration inhibitory factor

PVDF � polyvinylidene fluorideUTI � urinary tract infectionsWB � Western blot

REFERENCES

1. Hooton, T. M. and Stamm, W. E.: Diagnosis and treatment ofuncomplicated urinary tract infection. Infect Dis Clin NorthAm, 11: 551, 1997

2. Moore, K. N., Day, R. A. and Albers, M.: Pathogenesis of urinarytract infections: a review. J Clin Nurs, 11: 568, 2002

3. Backhed, F., Meijer, L., Normark, S. and Richter-Dahlfors, A.:TLR4-dependent recognition of lipopolysaccharide by epithe-lial cells requires sCD14. Cell Microbiol, 4: 493, 2002

4. Calandra, T. and Roger, T.: Macrophage migration inhibitoryfactor: a regulator of innate immunity. Nat Rev Immunol, 3:791, 2003

5. Roger, T., David, J., Glauser, M. P. and Calandra, T.: MIFregulates innate immune responses through modulation ofToll-like receptor 4. Nature, 414: 920, 2001

6. Meyer-Siegler, K. L., Leifheit, E. C. and Vera, P. L.: Inhibitionof macrophage migration inhibitory factor decreases prolifer-ation and cytokine expression in bladder cancer cells. BMCCancer, 4: 34, 2004

7. Meyer-Siegler, K. L., Ordorica, R. C. and Vera, P. L.: Macro-phage migration inhibitory factor is upregulated in an endo-toxin-induced model of bladder inflammation in rats.J Interferon Cytokine Res, 24: 55, 2004

8. Meyer-Siegler, K. L. and Vera, P. L.: Substance P inducedrelease of macrophage migration inhibitory factor from ratbladder epithelium. J Urol, 171: 1698, 2004

9. Meyer-Siegler, K. L. and Vera, P. L.: Intraluminal antibodies tomacrophage migration inhibitory factor decrease substance Pinduced inflammatory changes in the rat bladder and pros-tate. J Urol, 172: 1504, 2004

10. Baugh, J. A. and Bucala, R.: Macrophage migration inhibitoryfactor. Crit Care Med, 30: S27, 2002

11. Donn, R. P. and Ray, D. W.: Macrophage migration inhibitoryfactor: molecular, cellular and genetic aspects of a key neu-roendocrine molecule. J Endocrinol, 182: 1, 2004

12. Meyer-Siegler, K. L., Bellino, M. A. and Tannenbaum, M.: Mac-rophage migration inhibitory factor evaluation comparedwith prostate specific antigen as a biomarker in patients withprostate carcinoma. Cancer, 94: 1449, 2002

13. Leng, L., Metz, C. N., Fang, Y., Xu, J., Donnelly, S., Baugh, J. etal: MIF signal transduction initiated by binding to CD74. JExp Med, 197: 1467, 2003

14. Meyer-Siegler, K. L. and Vera, P. L.: Substance P inducedchanges in CD74 and CD44 in the rat bladder. J Urol, 173:615, 2005

15. Vera, P. L., Iczkowski, K. A., Leng, L., Bucala, R. and Meyer-Siegler, K. L.: Macrophage migration inhibitory factor is re-leased as a complex with �1-inhibitor-3 in the intraluminal fluidduring bladder inflammation in the rat. J Urol, 174: 338, 2005

16. Meyer-Siegler, K. L., Iczkowski, K. A. and Vera, P. L.: Furtherevidence for increased macrophage migration inhibitory fac-

tor expression in prostate cancer. BMC Cancer, 5: 73, 2005

MACROPHAGE MIGRATION INHIBITORY FACTOR IN CYSTITIS1528

17. Lan, H. Y., Bacher, M., Yang, N., Mu, W., Nikolic-Paterson,D. J., Metz, C. et al: The pathogenic role of macrophagemigration inhibitory factor in immunologically induced kid-ney disease in the rat. J Exp Med, 185: 1455, 1997

18. Matsumoto, K., Maruyama, N., Maruyama, T., Ohnishi, Y.,Nonaka, S., Inoshita, A. et al: Elevated macrophage migra-tion inhibitory factor (MIF) levels in the urine of patientswith focal glomerular sclerosis. Clin Exp Immunol, 139: 338,

2005

19. Engler, R.: Acute-phase proteins in inflammation. C. R SeancesSoc Biol Fil, 189: 563, 1995

20. Gonias, S. L., Carmichael, A., Mettenburg, J. M., Roadcap,D. W., Irvin, W. P. and Webb, D. J.: Identical or overlap-ping sequences in the primary structure of human al-pha(2)-macroglobulin are responsible for the binding ofnerve growth factor-beta, platelet-derived growth factor-BB, and transforming growth factor-beta. J Biol Chem,

275: 5826, 2000