High abundance of plasma cells secreting transglutaminase 2-specific IgA autoantibodies with limited somatic hypermutation in celiac disease intestinal lesions

Abstract

Celiac disease is an immune-mediated disorder in which mucosal autoantibodies to the enzyme transglutaminase 2 (TG2) are generated in response to the exogenous antigen gluten in individuals who express human leukocyte antigen HLA-DQ2 or HLA-DQ8 (ref. 3). We assessed in a comprehensive and nonbiased manner the IgA anti-TG2 response by expression cloning of the antibody repertoire of ex vivo-isolated intestinal antibody-secreting cells (ASCs). We found that TG2-specific plasma cells are markedly expanded within the duodenal mucosa in individuals with active celiac disease. TG2-specific antibodies were of high affinity yet showed little adaptation by somatic mutations. Unlike infection-induced peripheral blood plasmablasts, the TG2-specific ASCs had not recently proliferated and were not short-lived ex vivo. Altogether, these observations demonstrate that there is a germline repertoire with high affinity for TG2 that may favor massive generation of autoreactive B cells. TG2-specific antibodies did not block enzymatic activity and served as substrates for TG2-mediated crosslinking when expressed as IgD or IgM but not as IgA1 or IgG1. This could result in preferential recruitment of plasma cells from naive IgD- and IgM-expressing B cells, thus possibly explaining why the antibody response to TG2 bears signs of a primary immune response despite the disease chronicity.

Figure 1

TG2-specific ASCs are localized in the duodenal mucosa of CD patients. (a) TG2-specific and total IgA secretion by CD138⁺ ASCs as compared to CD138^neg cells (plated at 5-fold higher amount) sorted from an untreated CD patient. To exclude the possibility that detected antibodies were released from deposits, neoproduction of IgA was demonstrated by freeze-thaw treatment of ASCs. Data are representative of two independent experiments. Bars represent standard error. (b-h) Immunofluorescence staining of cryosections from biopsy specimens of CD patients and controls. (b) Section from an untreated CD patient stained with soluble biotinylated TG2 (green) and anti-CD138 (red); detail in panel c. (d) Exceptionally high frequency of TG2-specific ASCs in the lesion of another patient; detail in panel e. (f) Section from a CD patient stained with anti-IgA (green) and soluble biotinylated TG2 (red). (g) TG2 and anti-CD138 staining of a section from a healthy control. (h) Section from an untreated CD patient stained with a control biotinylated protein (BSA). Nuclei were counterstained with DAPI (blue). Data in b-h are representative of experiments performed on samples from several patients and controls. Scale bars, 50 μm.

Figure 2

TG2-specific ASCs are identified and sorted by flow cytometry for quantification and cloning of antibody variable genes. (a) Representative plots of ASCs stained with TG2 multimers from samples of patients with untreated or GFD-treated CD and non-CD controls. (b) Frequency of TG2⁺ ASCs in 33 analyzed samples. Each white dot represents a different donor. Antibodies were cloned from patients represented by black dots. Red dots represent two patients at early GFD timepoint with slow recovery and clinical observations still compatible with active disease. The blue dot represents a patient on 12 months of GFD, seronegative but with poor diet compliance and poor mucosal recovery. (c) Expression of intracellular Ki-67 and HLA-DR in ASCs (gate on large cells; ASCs defined as CD138⁺ cells). A proliferating Epstein-Barr Virus (EBV) immortalized B cell line was included as positive control for the intracellular Ki-67 staining. (d) Reactivity to recombinant human TG2 measured by ELISA of 63 hmAbs cloned from single TG2⁺ plasma cells of four patients with active CD, 15 hmAbs cloned from single TG2^neg plasma cells from patient 3, and 50 hmAbs cloned from naïve cells from five non CD subjects^,. For panel b: red bars indicate means and P values indicated were generated by Student’s t test. *, P < 0.05. ***, P < 0.001.

Figure 3

The highly restricted repertoire of TG2-specific ASCs suggests a unique origin for this autoantibody response. (a) Analysis of the heavy chain variable regions used comparing TG2-specific mucosal ASCs (by subject: n = 15, 3, 36, and 6) to the TG2-negative fraction (n = 30, and 38) and to ASCs from healthy mucosa of one subject (n = 69) or compared to our historical data on the repertoires of the intestinal anti-rotavirus plasma cells from 3 subjects, peripheral blood anti-influenza ASCs from 20 subjects and naïve cells^, from 5 subjects. Numbers of sequences in each pool are indicated in the pie charts. (b) Frequency of VH5-51 gene usage by subject. (c) Frequencies of VH region somatic mutations per sequence within the populations indicated. (d) Comparison of hmAb avidities as measured by the area-under-the-curve (AUC) of ELISA plots (original plots are in Fig. 2d). For panels a-d: red bars indicate means and P values indicated were generated by Student’s t test. *, P < 0.05. **, P < 0.01. ***, P < 0.001.

Figure 4

TG2 covalently crosslinks antibodies of the IgD isotype and its enzymatic activity is not inhibited by anti-TG2 hmAbs. (a) Western blot showing the effect of enzymatically active TG2 on a hmAb (679-14-E06) engineered with IgA1, IgG1, IgM or IgD constant regions. (b) Incubation of enzymatically active TG2 with a FITC-labeled gliadin peptide (DQ2.5-glia-α2(EQ) peptide, PQPELPYPQPQLPY) in the presence of the amine donor 5-BP generates transamidation and deamidation products, which are detected in a capillary electrophoresis assay. Details about the experimental aspects of the assay are reported in Supplementary Fig.11. (c) Effect of the mouse anti-TG2 murine mAb CUB7402 (positive control), mouse IgG1 isotype control, control hmAbs not reactive to TG2 (n=10) and anti-TG2 hmAbs (n=47) on the enzymatic activity of TG2. Residual activity on the Y axis is defined as the relative activity of each hmAb as compared with a non TG2-reactive isotype control hmAb, the activity of which is set to 1.