Contributors to the British Paediatric Rheumatology Study Group are as follows: Dr. M. Abinun, Dr. M. Becker, Dr. A. Bell, Professor A. Craft, Dr. E. Crawley, Dr. J. David, Dr. H. Foster, Dr. J. Gardener-Medwin, Dr. J. Griffin, Dr. A. Hall, Dr. M. Hall, Dr. A. Herrick, Dr. P. Hollingworth, Dr. L. Holt, Dr. S. Jones, Dr. G. Pountain, Dr. C. Ryder, Professor T. Southwood, Dr. I. Stewart, Dr. H. Venning, Dr. L. Wedderburn, Professor P. Woo, and Dr. S. Wyatt.
Mutation screening of the macrophage migration inhibitory factor gene: Positive association of a functional polymorphism of macrophage migration inhibitory factor with juvenile idiopathic arthritis
Article first published online: 27 SEP 2002
Copyright © 2002 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 46, Issue 9, pages 2402–2409, September 2002
How to Cite
Donn, R., Alourfi, Z., De Benedetti, F., Meazza, C., Zeggini, E., Lunt, M., Stevens, A., Shelley, E., Lamb, R., the British Paediatric Rheumatology Study Group, Ollier, W. E. R., Thomson, W. and Ray, D. (2002), Mutation screening of the macrophage migration inhibitory factor gene: Positive association of a functional polymorphism of macrophage migration inhibitory factor with juvenile idiopathic arthritis. Arthritis & Rheumatism, 46: 2402–2409. doi: 10.1002/art.10492
- Issue published online: 27 SEP 2002
- Article first published online: 27 SEP 2002
- Manuscript Accepted: 13 MAY 2002
- Manuscript Received: 20 MAR 2002
- Arthritis Research Campaign, UK
To determine if polymorphisms of the macrophage migration inhibitory factor (MIF) gene are associated with juvenile idiopathic arthritis (JIA).
Denaturing high-performance liquid chromatography was used to screen the MIF gene in 32 UK Caucasian controls and 88 UK Caucasian JIA patients. Ninety-two healthy UK Caucasian controls were then genotyped for each of the polymorphic positions identified. A panel of 526 UK Caucasian JIA patients and 259 UK Caucasian controls were subsequently genotyped for a single-nucleotide polymorphism (SNP) identified in the 5′-flanking region of the gene, using SNaPshot ddNTP primer extension and capillary electrophoresis. The functional significance of this polymorphism was also studied using luciferase-based reporter gene assays in human T lymphoblast and epithelial cell lines.
A tetranucleotide repeat CATT(5–7) beginning at nucleotide position −794 and 3 SNPs at positions −173 (G to C), +254 (T to C), and +656 (C to G) of the MIF gene were identified. No JIA-specific mutations were found. Allele and genotype frequencies differed significantly between the controls and the JIA patients for the MIF-173 polymorphism. Individuals possessing a MIF-173*C allele had an increased risk of JIA (34.8% versus 21.6%) (odds ratio 1.9, 95% confidence interval 1.4–2.7; P = 0.0002). Furthermore, the MIF-173* G and C variants resulted in altered expression of MIF in a cell type–specific manner. Serum levels of MIF were also significantly higher in individuals who carried a MIF-173*C allele (P = 0.04).
The −173-MIF*C allele confers increased risk of susceptibility to JIA. Our data suggest a cell type–specific regulation of MIF, which may be central to understanding its role in inflammation.
Growing interest surrounds the macrophage migration inhibitory factor (MIF) molecule (1). Originally described as a T cell–derived cytokine, MIF is now known to be expressed by many different cell types (for review, see ref. 2). MIF has proinflammatory, enzymatic, and hormonal activities (1, 3). The molecule is induced by low concentrations of glucocorticoids, which is unusual for a proinflammatory cytokine, and it then acts to counterregulate the antiinflammatory actions of glucocorticoids. At higher glucocorticoid concentrations, MIF secretion is suppressed. Therefore, MIF, together with endogenous glucocorticoids, is central to determining the magnitude, and possibly the chronicity, of an immune inflammatory response (1–3).
Multiple positive disease associations have recently been described with increased expression of MIF protein appearing to be critical to disease pathogenesis (4–9). Specifically, with respect to inflammatory arthritis, the use of monoclonal antibodies to block the actions of MIF results in an inhibition of the onset of rat adjuvant arthritis (6) and delays the development of collagen-induced arthritis (10). MIF has also been shown to reverse the antiinflammatory effects of glucocorticoids in murine antigen-induced arthritis (11). Elevated levels of MIF have been described in the serum and synovium of adults with rheumatoid arthritis (6, 12). The proinflammatory actions of MIF include up-regulation of the cytokines tumor necrosis factor α, interleukin-1β (IL-1β), IL-2, IL-6, IL-8, and interferon-γ. MIF also induces the local release of nitric oxide and matrix metalloproteinases 1 and 3 from synovial fibroblasts from patients with rheumatoid arthritis (13). In addition, Sampey et al (14) have also demonstrated MIF regulation of phospholipase A2 and cyclooxygenase 2 expression from fibroblast-like synoviocytes obtained from rheumatoid synovium. Despite all of these findings, little is known about the genetic control of the MIF gene or how polymorphisms within it may contribute to MIF expression in health or disease.
Juvenile idiopathic arthritis (JIA), previously called juvenile chronic arthritis in Europe and called juvenile rheumatoid arthritis in the US, is the most common chronic arthritic condition of childhood. JIA is a general term that encompasses 7 clinically distinct presentations (15). Despite the apparent heterogeneity of the differing JIA phenotypes, a unifying feature of all JIA is chronic inflammation in synovial joints, and as such, MIF is a relevant candidate gene for investigation in patients with JIA. Furthermore, Meazza et al (16) recently described significantly increased levels of MIF in sera and synovial fluid from children with JIA. This increase was most marked for children with systemic-onset JIA (16). Subsequently, Meazza et al also showed that higher synovial fluid levels of MIF are associated with a shorter clinical response to intraarticular corticosteroids (17).
We previously screened a 1-kb 5′-flanking region of the MIF gene, identified a novel single-nucleotide polymorphism (SNP) at −173, and showed a positive association with systemic-onset JIA (18). In the present study, we have extended our research into MIF and JIA by mutation screening the entire MIF gene and by studying the function of the MIF-173 polymorphism both in vitro and in vivo.
PATIENTS AND METHODS
Blood samples were obtained, with consent, from 526 UK Caucasian JIA patients seen by pediatric rheumatologists at different UK hospitals. Samples were sent to the British Paediatric Rheumatology Group's National Repository for JIA (Epidemiology Unit, Arthritis Research Campaign, Manchester, UK). All children included in this study satisfied the International League of Associations for Rheumatology classification for JIA (15). The control panel comprised 259 unrelated healthy UK Caucasian subjects, of whom 167 were blood donors and 92 were selected from the registers of general practitioners as part of a population-based study (19). Genomic DNA was extracted from all the samples using DNAce MaxiBlood Purification System kits (Bioline, London, UK).
Screening for potential SNPs in the MIF gene and in 1 kb of the 5′-flanking region using denaturing high-performance liquid chromatography.
The MIF gene and 1 kb of the 5′-flanking region (20) (GenBank accession no. L19686; European Molecular Biology Laboratory identification HSMIF) was divided into 5 overlapping fragments for mutation detection by denaturing high-performance liquid chromatography using a Transgenomic WAVE machine (Transgenomic, Crewe, UK). All fragments were <700 bp, ensuring >99% sensitivity and specificity (Table 1). To identify potential SNPs and to gain an estimate of the SNP frequency in a normal population, we examined genomic DNA from 32 healthy unrelated Caucasian subjects. In addition, a group of 88 JIA patients was screened to determine the presence of JIA-specific MIF mutations. The composition of the JIA panel was as follows: 9 had systemic-onset JIA, 24 had persistent oligoarticular, 12 had extended oligoarticular, 18 had rheumatoid factor (RF)–negative polyarticular, 7 had RF-positive polyarticular, and 8 had enthesitis-related JIA, and 10 had juvenile psoriatic arthritis.
|Primers for MIF fragment (5′ to 3′) and probe sequence for SNP (5′ to 3′)||Annealing temperature|
|MIF fragment (nucleotide position)|
|A1 (−1076 to −726)||Forward: CTG-CAG-GAA-CCA-TA-CCC-AT||60°C|
|A2 (−730 to −176)||Forward: TTA-CCA-TTA-GTG-GAA-AAG-ACA-TT||61°C|
|A3 (−235 to +130)||Forward: ACT-AAG-AAA-GAC-CCG-AGG-C||60°C|
|B1 (+5 to +625)||Forward: AGT-GGT-GTC-CGA-GAA-GTC-AGG||67°C|
|B2 (+623 to +1091)||Forward: CTG-AGC-CAC-CCG-CTG-AGT||60°C|
|MIF −173, GC||AGC-CGC-CAA-GTG-GAG-AAC-AG||–|
|MIF +254, TA||AGG-AAG-AGG-GGG-GTG-CCC-ACC-GGA-CGA-GGG-GT||–|
|MIF +656, CG||CAC-CCG-CTG-AGT-CCG-GCC-TCC-TCC-CC||–|
The DNA was amplified and analyzed using the appropriate acetonitrile gradient, as determined with WAVE-Maker software, version 4.0 (Transgenomic). Heteroduplex and homoduplex conformation patterns were easily distinguishable. Appropriate samples were then selected for sequencing. Polymerase chain reaction (PCR) products were sequenced in both directions using the BigDye Terminator Cycle Sequencing kit (PE Applied Biosystems, Foster City, CA). Primers used for sequencing were, in the main, the same as those used for the initial WAVE screening PCR (Table 1), except for fragment A1, where an additional internal primer was required (5′-GCA-GTA-TTA-GTC-AAT-GTC-TCT-TT-3′).
Determination of the frequency of the MIF polymorphisms and the extent of linkage disequilibrium in UK controls.
A panel of control subjects (n = 92) was genotyped for the CATT repeat and for SNPs at positions −173, +254, and +656 of the MIF gene, as determined by the initial mutation screening of the 32 normal individuals. The panel of 92 controls included the 32 normal individuals used for the mutation screening and formed part of the larger control panel of 259 subjects used for the association study with JIA patients.
Genotyping for the CATT repeat. Genomic DNA (50 ng) was amplified by PCR in a total reaction volume of 10 μl containing 5 pmoles of both the forward and reverse primers (forward primer 5′-TTG-CAC-CTA-TCA-GAG-ACC-3′; reverse primer 5′-TCC-ACT-AAT-GGT-AAA-CTC-G-3′). The forward primer was prelabeled with a FAM fluorescent dye. Four nanomoles of each of the 4 dNTPs, 0.2 units of Taq polymerase (Bioline), 1.5 mM MgCl2 buffer, 1× KCl buffer, and 1 mM betaine were included in the PCR mix. The PCRs were performed in 96-well microtiter plates on a Tetrad thermal cycler (MJ Research, Waltham, MA). Forty PCR cycles were carried out, each with denaturation for 1 minute at 95°C, primer annealing for 1 minute at 54°C, and extension for 45 seconds at 72°C. A final extension step was conducted at 72°C for 5 minutes. Amplified product was pooled with the TAMRA 350 size standard (PE Applied Biosystems).
Gel electrophoresis was performed on a 0.4-mm 6% polyacrylamide gel on a PE Applied Biosystems model 377 DNA sequencer. Gels were run at 1,200V for 2 hours. Semiautomated genotyping was carried out using GeneScan analysis and Genotyper 3.6 software. All genotyping was checked manually. Appropriate samples were then selected for sequencing. PCR products were sequenced in both directions using the BigDye Terminator Cycle Sequencing kit. Primers used for sequencing were the same as those used for the PCR amplification.
MIF SNP genotyping. The −173, +254, and +656 SNPs were each typed using the SNaPshot (PE Applied Biosystems) ddNTP primer extension method and capillary electrophoresis. The primers used were identical to those described in Table 1; the primer pair for fragment A3 was used for the MIF −173 SNP, B1 primers for the +254 SNP, and the B2 primers for the SNP at position +656. The probes used for the identification of each SNP are detailed in Table 1.
Briefly, for each SNP studied, a total of 20 ng of genomic DNA was amplified in a 10-μl final PCR reaction volume containing 5 pmoles of each primer, 0.08 nmoles of dNTPs, 1× KCl buffer, and 0.6 units of Taq polymerase. All reactions were performed in 384-well microtiter plates on a Tetrad thermal cycler. The DNA was denatured at 95°C for 5 minutes followed by 40 cycles at 95°C for 45 seconds, the appropriate annealing temperature for 45 seconds (Table 1), and at 72°C for 45 seconds. The final extension was carried out at 72°C for 5 minutes. The PCR product was incubated with 1 unit each of shrimp alkaline phosphatase (Amersham, Buckinghamshire, UK) and ExoI (New England Biolabs, Beverly, MA) at 37°C for 1 hour and at 72°C for 15 minutes. The extension reactions each utilized 1 μl of PCR template and 0.25 pmoles of probe (Table 1). Twenty-five cycles of the extension reaction were carried out, each cycle consisting of 96°C for 10 seconds, 50°C for 5 seconds, and 60°C for 30 seconds. Six microliters of this extension product was incubated with 1 unit of calf intestine alkaline phosphatase (Amersham).
One microliter of this product was pooled with 5 μl of deionized formamide and electrophoresed on an ABI 3100 analyzer. Samples of known sequence were included to optimize the SNaPshot running conditions. The results were analyzed using GeneScan analysis and Genotyper 3.6 software.
MIF-173 genotyping in UK JIA patients and controls.
The −173 MIF polymorphism was also used to genotype a total of 259 UK Caucasian controls and 526 UK Caucasian JIA patients. The primers, probes, and conditions were identical to those detailed above.
Determining the functional significance of the MIF-173 G-to-C polymorphism.
Generation of luciferase reporter gene constructs. Specific primers were designed to generate an 860-bp fragment of the 5′-flanking region (nucleotide positions −775 to +84 relative to the transcription start site), avoiding the known tetranucleotide repeat elements. The fragment was prepared with restriction enzyme sites added (Bam HI forward primer 5′-GGA-TCC-TCT-CTT-GAT-ATG-CCT-GGC-ACC-TGC-TAG 3′ and Xho I reverse primer 5′-CTC-GAG-ACC-AGG-AGA-CCC-GCG-CAG-AGG-CAC-AGA-3′). The PCR product was generated from DNA samples of known genotype from 2 JIA patients, using the Advantage HF-2 Taq polymerase kit (Clontech, Palo Alto, CA). The PCR product was initially inserted into the TOPO 2.1 TA cloning vector (Invitrogen, Groningen, The Netherlands) following the manufacturer's instructions, then into the relevant sites of the pGL3 luciferase reporter plasmid (Promega, Madison, WI). All constructs were fully sequenced to confirm the presence of MIF −173*G or −173*C and the absence of PCR errors.
Potential transcription factor binding sites.
To analyze the 860-bp construct for potential transcription factor binding sites, the MatInspector software (http://www.molbiol.ox.ac.uk/documentation/MatInspector.html) was used. This program relies on the TRANSFAC database (21) of transcription factors to create a profile of all potential DNA binding factors that may bind to a specific sequence.
Transient transfection of A549 and CEM C7A cells.
A549 (lung epithelial) cells were obtained from ECACC (Salisbury, UK), and cultured in DMEM with Glutamax (Gibco, Paisley, UK) supplemented with 10% fetal calf serum. No antibiotics were used. Cells were transfected using Lipofectamine Plus (Life Technologies, Paisley, UK) and were divided into treatment and control groups, so that within each experiment, all groups were from a single transfection. Typically, 105 cells were transfected in a 10-cm tissue culture dish. Cells were routinely cotransfected with a CMV-β-gal plasmid, and the results were expressed as corrected light units by dividing the luciferase assay light units by the optical density at 570 nm from the β-galactosidase assay using chlorophenol red galactosidase as substrate. Cells were transfected with 2.5 μg of MIF-luc. In all experiments, 2 μg of CMV-β-gal was added.
The human T lymphoblast cell line CEM C7A was obtained from Dr. G. Brady (University of Manchester, Manchester, UK), and cells were cultured in RPMI 1640 plus Glutamax, supplemented with 10% fetal calf serum. Cells were suspended at 107 cells/ml, and 0.8 ml was added to a 4-mm path-length electroporation cuvette (Bio-Rad, Hercules, CA). Ten micrograms of plasmid DNA was added, and the cells were electroporated at 260V and 1,050 μF in an Easyject1 electroporator (Eurogentec, Luik, Belgium). Within each transfection, the results were expressed relative to the normalized activity seen with the frequent MIF-173*G allele.
Luciferase activity in cell lysates was detected using a Berthold tube luminometer (model LB 9501; Berthold, Wildbad, Germany). Chlorophenylthio (cpt) cAMP was obtained from Sigma (Poole, UK) and was used at a final concentration of 250 μM. Forskolin (Sigma) was used at a final concentration of 5 × 10−6M.
Serum MIF measurement.
MIF was measured in the serum of controls (n = 53) with an enzyme-linked immunosorbent assay based on the use of a monoclonal antibody to MIF for coating and a biotinylated antibody to MIF for detection (both from R&D Systems, Minneapolis, MN). Recombinant human MIF (R&D Systems) was used as a standard. The assay was performed according to the instructions provided by the manufacturer. The detection limit of the assay was 31.25 pg/ml.
The genotype frequencies between the patient and control groups were compared using Fisher's exact test with Monte Carlo simulations (SPSS software version 8.0; SPSS, Chicago, IL). Allele frequencies were compared using the chi-square test (Stata Corporation, College Station, TX). The odds ratio (OR) and 95% confidence interval (95% CI) for the C phenotype were calculated using the Stata software. Linkage disequilibrium was determined using EH Plus software (22).
All reporter gene data points represent the mean ± SEM results of multiple independent experiments. Statistical significance was analyzed using analysis of variance and t-tests as appropriate. Wilcoxon's 2-sample rank sum test (Stata) was used to analyze the serum MIF levels by phenotype in the normal subjects.
A tetranucleotide repeat CATT(5–7), beginning at nucleotide position −794, and 4 SNPs at positions +24, −173, +254, and +656 were identified from the initial mutation screening carried out in 32 normal individuals. The nucleotide change at position +24 (A to T) occurred only once in the 32 controls (0.03% frequency) and not at all in the JIA patients; therefore, additional genotyping of this position was not performed. Ninety-two UK Caucasian controls were genotyped for the tetranucleotide repeat and for the SNPs at positions −173, +254, and +656. The allele and genotype frequencies for these polymorphisms are shown in Table 2.
|MIF polymorphism||Genotype; frequency, no. (%)||Allele; frequency, %|
|CATT(5–7) (n = 90)||5,5||5,6||5,7||6,6||6,7||7,7||5||6||7|
|7 (7.8)||35 (38.9)||0 (0)||38 (42.2)||10 (11.1)||0 (0.0)||27.2||67.2||5.6|
|−173 (n = 88)||GG||GC||CC||G||C|
|67 (76.1)||21 (23.9)||0 (0)||88.1||11.9|
|+254 (n = 88)||TT||TC||CC||T||C|
|69 (78.4)||19 (21.6)||0 (0)||89.2||10.8|
|+656 (n = 83)||CC||CG||GG||C||G|
|61 (73.5)||22 (26.5)||0 (0)||86.7||13.3|
Analysis across all 4 loci showed strong linkage disequilibrium to exist (P < 10−16). Specifically, the mutant alleles at the 3 SNP loci were in strong linkage disequilibrium with each other. The rarer length of 7 CATT repeats was also found to be in strong linkage disequilibrium with each of the rare SNP alleles.
Mutation screening of the 88 JIA patients revealed no additional informative polymorphism positions.
UK JIA association data for MIF-173.
No evidence of departure from Hardy-Weinberg equilibrium was seen in the 259 controls (P = 0.75). Overall, the frequency of the 3 possible −173 genotypes differed significantly between the patients and the controls, with an increase in both the heterozygotes and the homozygous mutants observed in the patients (P < 0.0001 by Fisher's exact test) (Table 3). Individuals possessing a MIF-173*C allele have an increased risk of JIA (34.8% versus 21.6%) (OR 1.9, 95% CI 1.4–2.7; P = 0.0002). When the MIF-173 genotype frequencies were compared between the different JIA subgroups, no significant differences were found (P = 0.63) (Table 4).
|Genotype||JIA patients (n = 526)||Controls (n = 259)|
|MIF-173 GG||343 (65.2)||203 (78.4)|
|MIF-173 GC||166 (31.6)||54 (20.8)|
|MIF-173 CC||17 (3.2)||2 (0.8)|
|Genotype||Systemic onset (n = 122)||Oligoarticular||Polyarticular||Enthesitis related (n = 32)||Juvenile PsA (n = 34)|
|Persistent (n = 143)||Extended (n = 73)||RF negative (n = 94)||RF positive (n = 28)|
|MIF-173 GG||76 (62.3)||95 (66.4)||44 (60.3)||63 (67.0)||17 (60.7)||23 (71.9)||25 (73.5)|
|MIF-173 GC||41 (33.6)||40 (28.0)||28 (38.4)||30 (31.9)||10 (35.7)||8 (25.0)||9 (26.5)|
|MIF-173 CC||5 (4.1)||8 (5.6)||1 (1.4)||1 (1.1)||1 (1.1)||1 (3.2)||0 (0.0)|
Reporter gene analysis.
MIF reporter genes were constructed as described above. The construct was designed to exclude the repeat region at −794, allowing the function of the −173 polymorphism to be analyzed. Initial experiments confirmed that the segment of the MIF 5′-flanking region did harbor promoter activity. Under basal conditions, luciferase activity was induced 10-fold by introduction of the MIF DNA upstream of the promoterless luciferase reporter gene (pGL3-luc) in A549 cells.
In A549 cells, a series of cotransfections including MIF-luc and CMV-β-gal were performed. The MIF-173G-luc had significantly greater luciferase activity (2-fold [±0.3]) compared with the MIF-173C-luc (P = 0.013) (Figure 1). This finding was consistent through multiple plasmid purifications. Since the human MIF promoter contains a putative cAMP response element, and since the rat gene has been shown to be cAMP responsive, the activity of forskolin on promoter activity was examined. Forskolin had no effect on MIF-luc basal promoter activity in either of the alleles tested.
The MIF gene is widely expressed, and since a link to chronic inflammatory arthritis was suggested by the genetic association studies detailed above, MIF promoter activity was measured in a human T lymphoblast cell line (CEM C7A). There was a consistent and significant increase in the promoter activity of the MIF-173C-luc compared with the MIF-173G-luc. Because the difference between the two was modest, as expected from the sequence comparison showing only a single nucleotide difference, these experiments were repeated on 9 occasions in multiples of 3 or 6 for a total of 39 separate transfections. Since the transfection efficiency differed by only 7% within each experiment, no correction was applied to the luciferase results (23, 24). Within each experiment, the results were normalized for the mean luciferase value of the MIF-173G-luc to control for differences in transfection efficiency between separate experiments. The MIF-173C-luc had a corrected luciferase activity of 1.4-fold (±0.9), as compared with 1-fold (±0.24) for the MIF-173G-luc (P = 0.013) (Figure 1). There was no induction of the MIF reporter genes in the presence of cpt cAMP. However, as a positive control experiment, the cpt cAMP was found to cause a 3-fold increase when it was transfected with a human prolactin reporter gene (data not shown).
Analysis with the MatInspector software (21) showed that the MIF-173*C allele created a potential activator protein 4 (AP-4) transcription factor binding site. This was the only difference documented between the MIF −173*C and −173*G sequences.
MIF levels in healthy controls.
The concentration of MIF protein (in ng/ml) was determined in the sera of 53 normal Caucasian control subjects of known MIF-173 genotype. Subjects with a MIF-173*C allele had significantly higher serum levels of MIF protein compared with subjects with the MIF-173*GG genotype (P = 0.04) (Figure 2).
Of the polymorphic sites identified in MIF, we concentrated mainly here on the −173 position. The G-to-C transition at −173 appears to be the most likely functional polymorphism identified so far, especially since the presence of the mutant C allele creates an AP-4 transcription factor binding site. It remains possible that the CATT repeat element may influence some aspects of MIF expression. Equally, intronic polymorphisms may have some functional capacity. Most notably, they can alter splicing. While this is possible, it would be unexpected for MIF, since transcripts of identical length have been seen for all the tissues studied thus far (20).
We found an increased risk of susceptibility to JIA attributable to carriage of the MIF-173*C allele. This increased risk was not limited to any one JIA subgroup. When the frequency of the MIF-173 alleles and genotypes were analyzed between the JIA subgroups, no significant differences were seen. This suggests that the effect of MIF-173*C is common to a shared feature of the JIA subgroups. This could be the ongoing inflammation that occurs within the synovial joints of JIA patients.
When MIF protein levels were determined in the sera of healthy individuals, those possessing a MIF-173*C allele were found to have significantly higher amounts of MIF, suggesting that individuals carrying the MIF-173*C allele produce higher amounts of MIF protein. Studies are ongoing to establish the relationship between MIF protein production and MIF-173 genotype in disease states.
The functional significance of the G/C at position −173 was studied using transient transfections in 2 different human cell lines. In the epithelial cell line, the G polymorphism was found to be correlated with increased MIF expression. In the T lymphoblast cell line, the reverse situation was found, with the MIF-173*C giving significantly increased MIF expression under basal conditions. These differences in expression are likely to be due to differences in transcription factor interaction with the MIF-173 element. Based on the promoter sequence analysis, AP-4 transcription factor is a particular candidate.
No induction with a cAMP analog could be shown for the MIF constructs in either cell line. This is different from the observation by Waeber et al (25) in their studies of a similar region of the MIF gene in mice, and it suggests that the cAMP element is not conserved across species or that the effects of cAMP may be mediated by regulatory elements outside the region covered by our construct.
Promoter polymorphisms in other cytokine genes have been found to have different functional capacities in a cell type–specific manner (26, 27). Our genetic and functional data are consistent with this, since it was the T lymphoblast CEM C7A transfections that showed that the MIF-173*C had increased MIF production, and a considerable body of evidence points to the role of T cells as pivotal contributors to persistent inflammation in JIA (28, 29).
The molecular mechanisms of the action of MIF are beginning to be defined. MIF plays a role in the innate immune response and is regulated by Toll-like receptor 4 and PU.1 transcription factor (30). Also, MIF has been shown to inhibit p53-dependent apoptosis. This inhibition coincides with the induction of arachidonic acid metabolism and cyclooxygenase 2 expression (31). MIF directly affects the transcriptional activity of AP-1–responsive genes via an interaction with Jab-1 (32). Mitchell et al (33) have previously reported that MIF regulates cytosolic phospholipase A2 activity via a pathway dependent on protein kinase A and extracellular signal–regulated kinase (33). These molecular studies enhance our understanding of the biologic roles of MIF.
A plethora of diverse disease associations have been described with increased MIF protein expression (4, 5, 34–38). It is encouraging to note that in addition to the association of MIF-173*C with JIA, as described herein, carriage of MIF-173*C has also been found to be associated with an increased risk of adult inflammatory arthritis (swelling in ≥2 joints for > 4 weeks) (39). Amoli et al (40), in a study of Spanish subjects, has also shown that MIF-173*C confers an increased risk of erythema nodosum in patients with biopsy-proven sarcoidosis. The polymorphisms we have identified should therefore assist in the understanding of the role of MIF in these conditions and in many other autoimmune and inflammatory diseases. In the future, targeting of glucocorticoid treatment or the success of specific MIF-directed therapies could also be influenced by an individual's MIF genotype.
- 16Increased levels of macrophage migration inhibitory factor in patients with juvenile idiopathic arthritis [abstract]. Arthritis Rheum 2001; 43 Suppl 9: S356., , , , .
- 18the British Paediatric Rheumatology Study Group. A novel 5′-flanking region polymorphism of macrophage migration inhibitory factor is associated with systemic-onset juvenile idiopathic arthritis. Arthritis Rheum 2001; 44: 1782–5., , , , and
- 25Transcriptional activation of the macrophage migration-inhibitory factor gene by the corticotropin-releasing factor is mediated by the cyclic adenosine 3′,5′-monophosphate responsive element-binding protein CREB in pituitary cells. Mol Endocrinol 1998; 12: 698–705., , , , , , et al.
- 27Inherited haplotypes of the interleukin-10 promoter differentially regulate gene transcription [abstract]. Arthritis Rheum 1999; 42 Suppl 9: S177., , , , , .
- 39MIF polymorphism is associated with susceptibility to inflammatory arthritis [abstract]. Rheumatology (Oxford) 2002; 41 Suppl 1: 126., , , , , , et al.
- 40Macrophage migration gene polymorphism is associated with sarcoidosis in biopsy proven erythema nodosum. J Rheumatol. In press., , , , , , et al.