Bacterial ClpB heat-shock protein, an antigen-mimetic of the anorexigenic peptide α-MSH, at the origin of eating disorders

Abstract

The molecular mechanisms at the origin of eating disorders (EDs), including anorexia nervosa (AN), bulimia and binge-eating disorder (BED), are currently unknown. Previous data indicated that immunoglobulins (Igs) or autoantibodies (auto-Abs) reactive with α-melanocyte-stimulating hormone (α-MSH) are involved in regulation of feeding and emotion; however, the origin of such auto-Abs is unknown. Here, using proteomics, we identified ClpB heat-shock disaggregation chaperone protein of commensal gut bacteria Escherichia coli as a conformational antigen mimetic of α-MSH. We show that ClpB-immunized mice produce anti-ClpB IgG crossreactive with α-MSH, influencing food intake, body weight, anxiety and melanocortin receptor 4 signaling. Furthermore, chronic intragastric delivery of E. coli in mice decreased food intake and stimulated formation of ClpB- and α-MSH-reactive antibodies, while ClpB-deficient E. coli did not affect food intake or antibody levels. Finally, we show that plasma levels of anti-ClpB IgG crossreactive with α-MSH are increased in patients with AN, bulimia and BED, and that the ED Inventory-2 scores in ED patients correlate with anti-ClpB IgG and IgM, which is similar to our previous findings for α-MSH auto-Abs. In conclusion, this work shows that the bacterial ClpB protein, which is present in several commensal and pathogenic microorganisms, can be responsible for the production of auto-Abs crossreactive with α-MSH, associated with altered feeding and emotion in humans with ED. Our data suggest that ClpB-expressing gut microorganisms might be involved in the etiology of EDs.

Figure 1

Proteomic identification of molecular mimicry between E. coli K12 proteins and α-MSH. (a) 2D GE of E. coli cytoplasmic proteins. (b, c) Immunoblots of E. coli proteins detected with rabbit anti-α-MSH IgG, preadsorbed (c) or not (b) with α-MSH. Circles in red surround the spots specifically recognized by α-MSH IgG which were used for protein identification. Circles in blue indicate nonspecific spots. Proteins identified in the spots 1–4 are isoforms of ClpB. (d) α-MSH and ClpB amino-acid sequence alignments using the Stretcher program. (e) Western blot of the recombinant ClpB, revealed with anti-α-MSH IgG. Lanes 1 and 2, 20 and 10 μg of ClpB, respectively. IP, isoelectric point.

Figure 2

ClpB immunization in mice. ClpB-immunized mice (ClpB+Adj) were compared with mice receiving adjuvant (Adj), PBS or controls (Ctr). (a) Body weight changes during 32 days of the study. Food intake and feeding pattern were studied during the last 2 weeks in the BioDAQ cages. Mean daily food intake (b), meal size (c) and meal number (d) during last 10 days of the study. (e) Food intake during 24 h after injection of α-MSH (100 μg kg⁻¹ body weight, i.p.) or PBS. (f) Plasma levels of ClpB-reactive IgG before and after adsorption with 10⁻⁶ M α-MSH. (g) Affinity of anti-ClpB IgG shown as the dissociation equilibrium constants (KD values). (h) Plasma levels of α-MSH-reactive total IgG. (i) Affinity (KD) of anti-α-MSH IgG. (j) cAMP assay in human embryonic kidney-293 cells overexpressing MC4R after stimulation by α-MSH alone or together with IgG (0.5 mg ml⁻¹) pooled from ClpB-immunized or from Adj-injected mice. (k) The cAMP assay was performed with IgG depleted from anti-α-MSH IgG. (a) Two-way repeated measurement analysis of variance (ANOVA) before α-MSH injection (100 μg kg⁻¹ body weight, i.p.), P<0.0001, Bonferroni post tests *^a at least, P<0.05 ClpB group vs Ctr; *^b at least, P<0.05 Adj group vs Ctr.; *^c, P<0.05, Student's t-test ClpB group vs PBS; and *^d, P<0.05, Student's t-test ClpB group vs Ctr. (b) ANOVA P=0.0002, Tukey's post tests ***P<0.001, **P<0.01, ^#P<0.05, Student's t-test. (c) ANOVA P=0.007, Tukey's post tests **P<0.01. (e) Student's t-test, *P<0.05. (f, g) ANOVA P<0.0001, Tukey's post tests ***P<0.001 ClpB +Adj vs other groups, paired t-test ^##P<0.01, ^###P<0.001. (h) ANOVA P=0.0002, Tukey's post tests ***P<0.001, *P<0.05; (i) Kruskal–Wallis test P=0.003, Dunn's post test **P<0.01, (mean±s.e.m., n=8). (j) ANOVA P=0.005, Tukey's post test *P<0.05; ANOVA P=0.04, Student's t-test ^#P<0.05, ^aClpB vs α-MSH, ^bClpB vs Adj. (mean±s.e.m.; j, n=6, k, n=3).

Figure 3

E. coli supplementation in mice. Effects of intragastric daily gavage (days 1–21) in mice with either E. coli K12 wild-type (WT), ClpB-deficient (ΔClpB) E. coli K12 or LB medium on body weight (a), food intake (b), meal size (c) and meal number (d). (e) PCR detection of a 180-base pair fragment of the bacterial ClpB DNA, first lane, molecular weight marker, second lane DNA from in vitro cultures of E. coli K12 WT, third lane DNA from in vitro cultures of E. coli K12 ΔClpB, and the remaining lanes DNA from mice feces collected at day 21. Plasma levels in optical density in enzyme-linked immunosorbent assay of anti-ClpB IgM (f) and IgG (g) before and after adsorption with 10⁻⁶ M α-MSH. Plasma levels of anti-α-MSH IgM (h) and IgG (i). (j) Affinity (equilibrium constant) of anti-α-MSH IgG. (a) Two-way repeated measurements analysis of variance (ANOVA), P=0.3, Bonferroni post test day 2, **P<0.01 control (Ctr) vs E. coli WT. (b) ANOVA days 1–2, P=0.0006, Tukey's post tests ***P<0.001, *P<0.05, E. coli WT vs ^aCtr and ^bLB. (c) Kruskal–Wallis (K–W) test third week P=0.0001, Dunn's post tests, ***P<0.001, **P<0.01, E. coli WT vs ^aCtr, ^bLB and ^CΔClpB. (d) ANOVA days 1–2, P=0.006, Tukey's post tests **P<0.01, *P<0.05, K–W test third week P<0.0001, Dunn's post tests, ***P<0.001, **P<0.01, E. coli WT vs ^aCtr, ^bLB and ^CΔClpB. (f) K–W test, before adsorption P=0.02, Dunn's post tests *P<0.05, ANOVA after adsorption, P<0.0001, Tukey's post tests **P<0.01, E. coli WT vs other groups. (g) ANOVA before adsorption, P=0.01, Tukey's post tests *P<0.05, E. coli WT vs other groups, paired t-test ^##P<0.01. (h) Student's t-test, E. coli WT vs other groups *P<0.05. (j) K–W test P=0.02, Dunn's post test *P<0.05, Mann–Whitney test, ^#P<0.05. (mean±s.e.m., n=8).

Figure 4

Anti-ClpB antibodies in ED patients. Plasma levels of anti-ClpB IgG (a) and IgM (b) in healthy women (control, Ctr) and in patients with AN, BN and BED. Plasma levels of ClpB IgG (c) and IgM (d) before and after adsorption with 10⁻⁶ M α-MSH. Percentage of α-MSH crossreactive anti-ClpB IgG (e) and IgM (f). (b) Student's t-test *P<0.05. (c, d) Paired t-tests, ***P<0.001, **P<0.01. (e) Kruskal–Wallis test P<0.0001, Dunn's post test, **P<0.01, Mann–Whitney test ^#P<0.05. (f) Analysis of variance P=0.02, Tukey's post test *P<0.05. (mean±s.e.m., Ctr, n=65, AN, n=27 BN, n=32 and BED, n=14).