Restricted VH/VL usage and limited mutations in gluten-specific IgA of coeliac disease lesion plasma cells

Abstract

Coeliac disease (CD), an enteropathy caused by cereal gluten ingestion, is characterized by CD4(+) T cells recognizing deamidated gluten and by antibodies reactive to gluten or the self-antigen transglutaminase 2 (TG2). TG2-specific immunoglobulin A (IgA) of plasma cells (PCs) from CD lesions have limited somatic hypermutation (SHM). Here we report that gluten-specific IgA of lesion-resident PCs share this feature. Monoclonal antibodies were expression cloned from single PCs of patients either isolated from cultures with reactivity to complex deamidated gluten antigen or by sorting with gluten peptide tetramers. Typically, the antibodies bind gluten peptides related to T-cell epitopes and many have higher reactivity to deamidated peptides. There is restricted VH and VL combination and usage among the antibodies. Limited SHM suggests that a common factor governs the mutation level in PCs producing TG2- and gluten-specific IgA. The antibodies have potential use for diagnosis of CD and for detection of gluten.

Figure 1. Survival of intestinal PCs in culture.

(a) Concentration of IgA in supernatants after 0, 1, 2, 3 and 4 weeks culture of SCSs (4-5000 cells ml⁻¹) grown with (F+) or without fibroblasts (F−). (b) Concentration of IgA after 0 and 2 days in supernatants from cultures of fibroblasts and PCs when PCs were added either as isolated IgA⁺ PCs (IgA⁺ PCs) or as part of SCSs (80-100 IgA+PCs ml⁻¹). IgA was detectable only in cultures of PCs as part of SCSs. The bars represents mean values of three separate culture wells and error bars represent s.e.m. (ND denotes non-detectable). (c) Representative flow cytometry plots of SCSs after 4 weeks of co-culture with fibroblasts. Viable PCs identified as large (left), viable CD27⁺CD3⁻CD14⁻ (middle) cells as indicated by gating strategy. The PC population divides into IgM⁺ and IgM⁻ PCs, where the IgM⁻ are mainly IgA⁺ PCs (right).

Figure 2. Supernatant reactivity to CT-gliadin and TG2 by ELISA.

IgA reactivity to antigen in representative cultures of SCSs from (a) subjects with UCD. (b) The ratio of culture supernatants with IgA reactivity to TG2 versus CT-gliadin, where the dots represent different subjects with UCD (n=8). Horizontal bar indicates mean value. (c) IgA reactivity to antigen in one representative of two tested non-CD controls (Ctr). (d) The background level was defined by signal in supernatants of cultures of fibroblasts only. In the headline are shown the number of cells of SCSs per culture well and antigen used as coating antigen in ELISA. Each column represents one well.

Figure 3. ELISA reactivity of 19 hmAbs expression cloned from IgA⁺ PCs in culture.

The antigens used for ELISA coating were CT-gliadin and PLQPEQPFP, and BSA for control as shown in headlines. Values are given as median and range of triplicates of hmAbs tested. The same hmAbs are tested for reactivity to all three antigens, and each hmAb is given the same position in each of the three graphs. The names of the hmAbs are depicted in 3c. Horizontal bars denote cutoff defined by negative control hmAb (NEG CTR). (a) Nine of 19 hmAbs were reactive to CT-gliadin. (b) One of nine hmAbs reactive to CT-gliadin was also reactive to BSA. (c) Five of eight hmAbs reactive to CT-gliadin but not BSA were also reactive to PLQPEQPFP.

Figure 4. Staining of intestinal PCs with tetramers of synthetic gluten peptides or TG2 in flow cytometry.

(a) Representative plots of PCs from SCSs stained with APC-conjugated streptavidin in complex with biotinylated PLQPEQPFP peptide. The IgA-negative events are mainly IgM⁺ PCs. The four plots represent two control patients (Ctr) and two subjects with UCD. (b) Frequency of IgA⁺ PCs stained positive with peptide in percentage of total number of IgA⁺ PCs. The peptide used as antigen is given in headline. Each dot represents one subject, and mean values are indicated by horizontal bars. ‘Background SA’ represents cells stained with APC-streptavidin alone. (c) Representative plots of one subject with UCD of IgM⁺ PCs and IgA⁺ PCs from SCSs stained with biotinylated peptides in complex with APC-conjugated strepavidin and biotinylated TG2 in complex with PE-conjugated streptavidin. (d) Ratio of PCs stained positive with TG2 compared with synthetic gliadin peptides. IgM⁺ PCs and IgA⁺ PCs of the same sample are connected by lines.

Figure 5. AlphaLISA characterization of gliadin-reactive hmAbs.

(a) Antibody reactivity to constant concentration of biotinylated PLQPEQPFP in the presence of competing antigens. The starting concentrations of competing antigens are CT-gliadin 10 μg ml⁻¹, cell lysate 1 ml cells ml⁻¹, TG2 200 μg ml⁻¹, LPS 250 μg ml⁻¹, CpG 100 μg ml⁻¹ and Jo-1 200 μg ml⁻¹. Only CT-gliadin blocked the signal, while cell lysate or a mixture of TG2, LPS, CpG and Jo-1 did not. (b) Reactivity to deamidated gliadin versus native gliadin. Representative plots of two hmAbs binding to constant concentration of biotinylated PLQPEQPFP in competition with various concentrations native PLQPQQPFP or the deamidated PLQPEQPFP was measured. The concentrations of competitive peptide are depicted on x axis.

Figure 6. Epitopes of gliadin-specific hmAbs.

(a) Reactivity of hmAbs to PLQPEQPFP and 33-mer. Reactivity of each hmAb is connected by line. Antibodies are grouped according to their reactivity pattern with reactivity to only one of the two peptides (left, right) or cross-reactive with both peptides (middle). (b) Reactivity of two representative hmAbs to constant concentration of biotinylated PLQPEQPFP in competition with four different competitive synthetic gliadin peptides. The amino-acid sequences of competitive peptides are shown and the concentrations of competitive peptides are depicted on x axis. Amino-acid sequences shared with the PLQPEQPFP selecting peptide are underlined.

Figure 7. AlphaLISA anti-PLQPEQPFP immunoglobulin inhibition assay.

Reactivity of hmAbs UCD1114 1F03 and UCD1143 3B02 to constant concentration of biotinylated PLQPEQPFP in the presence of sera from either UCD or control subjects (Ctr). The signal was blocked by sera from UCD patients but not from control subjects.

Figure 8. VH and VL usage and SHMs.

(a) VH/VL usage of 38 gliadin-specific hmAbs from PCs either isolated by culture or by flow sorting. (b) Frequencies of VH region somatic mutations (y axis) per sequence in the populations indicated on x axis. ‘Gliadin-specific IgA’ represents hmAbs from PCs sequestered after in vitro culture (open circles) and IgA⁺ PCs isolated with synthetic gliadin peptides (black dots). ‘Control IgA’ represents sequences derived from intestinal IgA⁺ PCs sorted as TG2-negative and PLQPEQPFP/33-mer-negative from two subjects with untreated CD (UCD1130 and UCD1143). Each point represents a single IgA sequence. (c) Comparison of mutations in VH3-23 genes. Horizontal bars indicate median values and P-values were obtained by Student’s t-test. NS>0.05. ***P<0.001.