Raised levels of anti-glucose-6-phosphate isomerase IgG in serum and synovial fluid from patients with inflammatory arthritis

Abstract

Background: In K/BxN mice, anti-glucose-6-phosphate isomerase (GPI) antibodies (Abs) are arthritogenic, and their transfer into naive mice induces arthritis. Anti-GPI Abs develop in many human patients with RA and are associated with more severe forms of the disease.

Objective: To elucidate the serum and synovial fluid (SF) anti-GPI IgG profiles among different patient groups with a variety of arthritides.

Methods: Blood and SF obtained concomitantly from 91 patients with clinically well defined arthritis were tested for concentrations of total anti-GPI IgG, anti-GPI IgG subclasses, B lymphocyte stimulator (BLyS), and APRIL by ELISA.

Results: Anti-GPI IgG was detected in sera and SF of patients with many arthritic diseases, but was preferentially associated with inflammatory arthritis, in general, and RA, in particular. The anti-GPI IgG subclass usage was skewed and varied among the different arthritic disease groups. Inverse correlations between serum levels of BLyS and anti-GPI IgG and positive correlations between serum levels of APRIL and anti-GPI IgG were seen among immune based arthritic patients and patients with RA but not among non-immune based patients. No correlations were found in SF from any group of arthritic patients.

Conclusion: Raised circulating anti-GPI Abs are not unique to patients with RA but are present in many patients with inflammatory arthritis. The difference in anti-GPI IgG subclass usage among disease groups may influence effector function and disease outcome. The inverse correlation between serum BLyS and anti-GPI IgG levels suggests that anti-GPI B cells may be regulated differently from other autoantibody producing B cells. Anti-GPI Abs may serve a pathogenic function in humans by promoting the maintenance of existing disease.

Figure 1

Anti-GPI-specific IgG purified from the serum of a patient with RA with high anti-GPI titre. (A) Samples from different purification steps, including the RA serum before purification, the purified total IgG fraction, and the purified anti-GPI IgG fraction, were tested for their binding to rabbit GPI (rGPI), human GPI (hGPI), bovine serum albumin, and glutathione-S-transferase as control antigens in an ELISA. Mouse K/BxN serum and normal serum from a healthy subject (NS), were included as positive and negative controls. The fraction containing the purified anti-GPI IgG preparation was highly specific for hGPI and did not react with the other control antigens. (B) Titration of the human anti-GPI IgG fraction showed positive GPI reactivity down to 0.06 µg/ml of IgG, comparable to the mouse K/BxN serum, whereas no reactivity was found in the remaining RA IgG fraction after affinity purification.

Figure 2

Anti-GPI IgG concentration in sera and SF of patients with various arthritic diseases. Sera (A) and SF (B) obtained concomitantly from 37 patients with RA, 13 with OA, 6 with Tr, 14 with crystal induced arthritis (Crys), 11 with seronegative SpA, 10 with other forms of inflammatory arthritis (Other) and in 21 normal healthy control sera (NS) were tested for binding to recombinant hGPI by ELISA. Bound IgG was detected with an AP conjugated F(ab)₂ goat antihuman IgG-Fc and the concentration calculated according to a standard serum with known concentration of IgG. The cut off points for positivity were calculated as the 95% centile of the concentrations in healthy control subjects (>35.4 µg/ml). Immune based arthritis (Imm) includes patients with RA, SpA, SLE, undifferentiated inflammatory polyarthritis, and polymyalgia rheumatica. Non-immune based arthritis (Non) includes patients with OA, Tr, Crys, infectious arthritis, and adenocarcinomatous arthritis. The lines inside the boxes indicate the medians, the outer boxes indicate the 25th and 75th centiles; the bars extending from the boxes indicate the 10th and 90th centiles.

Figure 3

Distinct anti-GPI IgG subclass composition observed for patients with different arthritic diseases. Serum samples from 37 patients with RA, 12 with OA, 12 with Crys, 6 with Reiter's syndrome, and 21 healthy subjects (NS) were tested for binding to hGPI in an ELISA. Bound anti-GPI IgG was detected with subclass-specific (IgG1–4) Ab, and the subclass concentration determined and expressed as the median concentration in µg/ml of the different anti-GPI IgG subclasses.

Figure 4

Correlation between BLyS or APRIL levels and anti-GPI IgG concentration in serum and SF samples of arthritic patients. Correlations representing all 91 patients with arthritis (A-D), only patients with immune based arthritis (E-H), or only patients with RA (I-L).

Figure 5

Correlation between RF and BLyS, APRIL, or anti-GPI IgG levels in serum of patients with RA. Serum RF values plotted against serum BLyS (A), serum APRIL (B), and serum anti-GPI IgG (C) for the patients with RA for whom concurrent serum BLyS and RF levels were available.