INCREASED FREQUENCY OF THE UNCOMMON ALLELE OF A TUMOUR NECROSIS FACTOR ALPHA GENE POLYMORPHISM IN RHEUMATOID ARTHRITIS AND SYSTEMIC LUPUS ERYTHEMATOSUS

The frequency of the uncommon allele (TNF2) of a polymorphism in the promoter region of the tumour necrosis factor alpha (TN Fa) gene in patients with rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) was found to be 3 times that of the normal anglo-saxon population . In SLE patients, this allele was strongly associated with HLA-DR3 expression and was also more frequent in patients who did not have malar rash. Functional studies of normal monocyte cytokine production in vitro showed that this genotype was associated with increased IL-I a protein production but there were no differences in the production of TNFa protein.


INTRODUCTION
Rheumatoid arthritis (RA) and Systemic lupus erythematosus (SLE) are autoimmune diseases of unknown aetiology but with a genetic predisposition which is only partly accounted for by genes within the HLA region (Grennan and Dyer, 1988;Grennan et al. , 1994).Cytokine genes such as TNFa may be candidate genes responsible for some of the undefined genetic predisposition in autoimmune diseases.TNFa has been shown to contribute to genetic predisposition to murine lupus (Jacob, 1992) and recent treatment of RA patients with anti-TNFa antibodies has been shown promising results (Elliott et al., 1993).Restriction fragment length polymorphism analysis of the TNFa locus using the endonuclease Ncol has revealed an increase in the frequency of the 5.5kb fragment in SLE but not in RA (Tomita et  al., 1993 ; Fugger et al. , 1989).Recently, a polymorphism in the promoter sequence of the TNFa gene has been described (Wilson et aI. , 1992).The frequency of the uncommon alIele (TNF2) in the latter report was low (16 %) and concur with our results that show a frequency of 12% in the normal anglo-saxon population.Here we report that the frequency of the TNF2 allele is significantly increased in RA (31 '7c ) and SLE (30%) patient groups, and that this genotype may be associated with increased IL-l a production by monocytes.

Patients
Thirty four patients with classic or definite RA by the criteria of the American Rheumatism Association (ARA) (Ropes et ai., 1958) and 40 unrelated patients with SLE by the criteria of the ARA (Tan et ai., 1982) were studied.Patients with SLE were subgrouped according to the following parameters: malar rash, photosensitivity, renal proteinuria, antinuclear antibody (SSA), and HLA-DR3 type.In addition, 57 healthy anglo-saxon Caucasian controls were studied.

peR analysis of IL-J ra gene polymorphism
Genomic DNA was extracted from 20ml of EDT A blood by proteinase K digestion, phenol-chloroform extraction and ethanol precipitation using an established protocol (Wyman, R. and White, 1980).200ng of DNA in 50 f.llreaction buffer was subjected to PCR amplification for the I 07bp sequence containing the polymorphic site of the TNFa gene in a GeneAmp PCR System 9600 (Perkin Elmer) using the primers as described (Wilson et ai., 1992).The polymorphism is a base change from G to A at position -308 and the forward primer created an Ncol recognition site which could be used to detect the polymorphism.Primers were synthesised using a 391 DNA Synthesiser (PCR-Mate, Applied Biosystems).The reaction buffer consisted of lOmM Tris pH8.8 with 75mM KCl and 1.5mM MgCI 2 .
DNA was subjected to a preliminary cycle of 94 0 for 3 minutes, 60 ' C for one minute and 72'C for one minute.This was followed by 35 cycles of the following: 94°C for 30 seconds, for 60'C for 30 seconds and 72 e C for 60 seconds.The final cycle was 94 'c for 30 seconds, for 60'C 30 seconds and 72'C for 5 minutes.PCR products were digested with Ncol and separated by electrophoresis on 10% polyacrylamide gels and visualised by ethidium bromide staining.Allele 1 was recognized by a band 87bp in size representing the larger fragment of a cut at position 20 of the 107bp PCR fragment.Allele 2 was recognised as an uncut nucleotide of 107bp.Heterozygotes were recognised by the appearance of2 bands (87bp and I 07bp) clearly distinguishable on the gels.A molecular ladder ranging from 67bp to 500bp (pUCHpaII, Bresatec) was used to determine the , iz e of the PCR fragment s.
Monocytes were cultured at 106/ml in a total volume of 0.4 ml in tubes fitted with a loose cap in 5% CO 2 in air in a humidified incubator.After 20 hours the tubes were centrifuged at 200g for 10 minutes and the supernatants harvested.The cells were resuspended in 0.4 ml offresh culture medium and disrupted by sonication for 10 seconds at 50W using a Labsonic 2000 ultrasonic homogeniser (B.Braun Instruments, Sydney.Australia) with a 4mm titanium needle probe.This was used to estimate the cellassociated IL-I activity.Levels of cell-associated IL-6 and TN Fa were consistently low «0.3 ng/ml) and were not routinely measured.Samples were stored at -20 ' C for up to 4 weeks prior to assay.

Cytokines and cytokine ELlSAs
Sandwich ELISAs specific for IL-l a, IL-I~, IL-I ra, IL-6 and TNFa were developed and used as previously described (Danis et aI., 1991;1994).Recombinant human IFNy and recombinant human TN Fa were provided by Boehringer Ingelheim (Austria) , recombinant human GM-CSF by the Schering Corporation (Sydney, Australia), recombinant human IL-I ra by the Upjohn Company (Kalamazoo, USA) , and recombinant human IL-6 by Dr. Gordon Wong (Genetics Institute, USA).Recombinant human IL-l a and IL-I ~ were purchased from Boehringer-Mannheim (Sydney, Australia).

Statistical Analysis
The results are presented as genotypes and allele frequencies.The Chi-square test for statistical significance was performed on allelic frequencies.The Mann-Whitney U Test was used to compare cytokine protein production by monocytes from genotypically distinct normal healthy individuals.

RESULTS AND DISCUSSION
The TNFa genotype and allele frequencies in RA and SLE patients and controls are summarised in Table I.The frequency of allele 2 in RA patients (31 %) and in SLE patients (30%) was significantly higher than in controls (12%).The frequency of the expression of TNF2 in normal populations has previously been reported to be between 7% (Verjans et aI. , 1994) and 22% (Mansfield et aI., 1994).Increased frequency ofTNF2 in SLE patients has recently been reported by Wilson et al. (1994).When we looked at disease subgroups in our SLE patients we found significantly higher expression ofTNF2 in patients with HLA-DR3 haplotype, patients without malar rash, patients positive for the antinuclear antibody SSA and patients with renal proteinuria (Table 2).On the other hand, there was no difference in the frequency of expression ofTNF2 in SLE patients with photosensitivity compared with those with no photosensitivity.
The TNF locus is in close proximity to the HLA locus and it is difficult to determine whether the polymorphism at the TNF locus is having a direct effect or is in linkage disequilibrium with a polymorphism in the HLA region, especially in HLA-DR.Genetic susceptibility to RA is partly accounted for by HLA-DR4 (Grennan and Dyer, 1988) and genetic susceptibility to SLE is partly accounted for by HLA-DR3 (Grennan et al. ,.Previous work has shown TNF2 is associated with HLA-Al, B8 and DR3 alleles (Wil son et aI., 1993) and, at least in the context of Diabetes mellitus, it was not possible to sho\\ an independent association of TNF2 with disease (Pociot et aI., 1993).Our re s ult ~ confirm an association between TNF2 and HLA-DR3 in SLE patients.However.there Table 1.TNFa genotypes and allele frequencies in healthy and diseased populations.

Genotype
Allele Frequencies (%)  are also association s bet \\ ee n T\ F2 and other disease features in SLE such as malar rash, renal proteinuria and antinuclear antibody expression which are not associated with HLA-DR3 expression (Grenn an e t al. , 1994).
The functional signifi can ce of the polymorphism in the promoter region of the TNFa gene is still not kno \\ n.Th e regulation of TNFa protein production is primarily post-transcriptional.We ha\ e not been able to demonstrate an association between the TNF2 allele and TN Fa prot ein production by normal human monocytes stimulated with GM-CSF and IFNy ill \• ill'o.Other workers have failed to demonstrate an association between the TN F2 allele and TN Fa production in LPS-stimulated monocyte cultures (Pociot el al .. 1993).We also did not find any association between TNFa genotype and the production of IL-I~, IL-I ra or IL-6 proteins by monocytes (Table 3).However, there was a significant association between TNF2 and increased IL-l a production by monocyte s ill \'itro (Figure I).This may be due to enhanced TNFa production early in culture and more detailed kinetic studies are required to resolve this point.

CONCLUSION
Expression of the TNF2 allele is increased in autoimmune disease and upregulation ofIL-I a production may partly explain the functional significance of this polymorphism.
Figure I. Relationship between the production of IL-I a protein by human monocyte s and TNFa genotype.Purified human monocytes from hea lthy volunteers were stimulated with GM-CSF and INFy for 20 hours before the supernatants and cel l Iysates were co llected and assayed for cytok ine product ion.Each point represents the mean of two e stimation s of the lL-1 C1 protein production from a g iven indi vidual.Circles represent the A I A I genotype , and diamonds represent th e A I A2 genotype.Horizontal bars indicate median s.Data are expre ssed as total IL-I C1 production or as the secre ted and ce ll-associated components respectively.The significance levels were determined by th e Mann Whitn e\ U test.

Table 2 .
TN Fa genotypes and allele frequencies in SLE sub-groups.
Chi-square frequency probability compared with the normal control population is shown in brackets.

Table 3 .
Cytokine production by normal monocytes stimulated with GM-CSF+IFNy in relation to TNFa genotype.Median values fr olll T \ 'F I lIlJi \ iJuals (A I A I) are cOlllpared with TNF2 individuals (A 1 A2).Significance le\eh \I ere obtained Lhin g the Mann-Whitney U test.Monocytes were cultured at a Illillion per Illi in seru ill -free Illedi ulll for 20 hours as outlined in Materials and Methods.