TG2-ω18merEE fusion protein is expressed in intestinal epithelial cells in doxycycline-treated double transgenic mice.

(A) Schematic overview of the transgene construct made to generate ROSA26-Tgm2-ω18merEE; H2B-eGFP transgenic mice, referred to as Tgm2- ω18merEE single transgenic (STG) mice. The cDNA sequence encoding murine transglutaminase 2 (TG2), with retention of intron 7, was fused to cDNA encoding the 18mer deamidated gluten peptide QPEQPFPQPEQPFPWQPQ (ω18merEE) and linked to a nuclear histone H2B-eGFP reporter via P2A peptide sequences, all placed under control of the tetracycline-responsive promoter (TREtight) and targeted in reverse orientation into the Rosa26 locus. (B) Illustration of doxycycline (dox)-induced expression of the transgenic TG2-ω18merEE fusion protein in intestinal epithelial cells (IECs) using the Tet-On system. In IEC-rtTA STG mice, expression of the reverse tetracycline-controlled transactivator (rtTA) is driven by the Villin1 promoter, active in IECs. Crossing IEC-rtTA STG mice with Tgm2- ω18merEE STG mice generates double transgenic (DTG) mice, in which rtTA binds and activates the TREtight promoter in the presence of dox, thereby inducing TG2- ω18merEE expression in IECs. (C) Induction of eGFP and TG2-ω18merEE protein in small intestinal tissue from DTG mice 24 h after intraperitoneal administration of dox. eGFP (green) was detected by direct epifluorescence, and tissue was stained with anti-TG2 (red) anti-DGP (white) and DAPI (blue). (D, E) DTG, Tgm2-ω18merEE STG or IEC-rtTA STG mice were administered dox via drinking water (2 mg/ml) or normal drinking water (no dox) for 3 days followed by 24 h on normal drinking water. IECs were isolated from the small intestine and lysed by three freeze-thaw cycles. (D) Western blot detection of TG2-ω18merEE fusion protein in IEC lysates. Top: Anti-TG2 antibodies detect transgenic TG2-ω18merEE protein (∼82 kDa, upper band) and endogenous TG2 (∼78 kDa, lower band). The band at ∼42 kDa represents β-actin, which was included as loading control. Bottom: Anti-deamidated ω18mer (ω18merEE) antibodies detect the ω18merEE peptide in the transgenic TG2-ω18merEE fusion protein. Recombinant TG2-ω18merEE protein was loaded as a control. Data are representative of three independent experiments. (E) TG2-specific B cells and non-cognate B cells were isolated from Tgm2-/- HLA-DQ2.5 14E06 mice and HLA-DQ2.5 mice, respectively, labeled with CellTrace Violet (CTV), and incubated with IEC lysate from dox-treated DTG mice for 10 minutes on ice. B cells were washed and co-cultured with CFSE-labeled gluten-specific CD4+ T cells, isolated from HLA-DQ2.5 TCR-glia-ω2 mice, for 3 days at 37°C. Proliferation of T cells (left) and B cells (right) as indicated by dilution of proliferation-tracking dyes is shown as the mean of culture duplicates ± SD. Data are representative of three independent experiments. Source Data 1. Uncropped Western Blots shown in Figure 1D. Figure Supplement 1. Gating strategy for Figure 1E.

TG2-ω18merEE is detected in the intestinal lumen of dox-treated DTG mice.

DTG mice received dox (50 mg/kg) by i.p. injection, and were euthanized 4 h, 12 h, or 3, 5 or 7 days after treatment. Small-intestinal lavage fluids were collected, and the TG2-ω18merEE fusion protein was extracted by immunoprecipitation using an antibody recognizing the deamidated ω-gliadin peptide, followed by Western blot visualization using a polyclonal anti-TG2 antibody. Data are representative of two independent experiments. Source Data 2. Uncropped Western Blot.

IEC-derived TG2-ω18merEE fusion protein drives anti-TG2 responses in a T-cell dependent manner.

(A) Schematic overview of the experiment. Naive gluten-specific CD4+ T cells, isolated from HLA-DQ2.5 TCR-glia-ω2 mice, were transferred to HLA-DQ2.5 DTG or HLA-DQ2.5 Tgm2-ω18merEE STG recipient mice. The following day, T cells were activated by intragastric (i.g.) administration of deamidated gluten peptides (DGP, 100 µg or 500 µg) and cholera toxin (CT, 10 µg). Three days later, naive TG2-specific B cells isolated from Tgm2-/- HLA-DQ2.5 14E06 KI mice were transferred and dox (50 mg/kg) was administered i.p. 2 hours after B-cell transfer. Serum was collected at the indicated time points. The small intestine was harvested 14 days (14d) after dox injection for collection of lavage fluids and isolation of lamina propria lymphocytes. (B) Anti-TG2 IgG and IgA in serum, and anti-TG2 IgA in small-intestinal (SI) lavage in DTG mice treated with dox or PBS, or in Tgm2-ω18merEE STG mice. The control group represents dox-treated DTG mice that did not receive i.g. immunizations of DGP and CT after T-cell transfer (unprimed). Area under the curve (AUC) was determined as a function of serum-to-buffer ratio measured as peaks above baseline, ignoring peaks < 10% of maximum. (C) Anti-TG2 IgA-producing plasma cells (PCs) expressed as a percentage of total IgA-producing PCs in the SI lamina propria in the groups as described under (B). (B-C) Data are shown with median and are pooled from five independent experiments. *p < 0.05, **p < 0.01, and ****p < 0.0001 by one-way ANOVA.

Strong expansion of TG2-specific B cells in response to IEC-derived TG2-ω18merEE fusion protein in Peyer’s patches.

CD45.2 HLA-DQ2.5 DTG or CD45.2 HLA-DQ2.5 Tgm2-ω18merEE STG mice received naive gluten-specific CD4+ T cells, isolated from HLA-DQ2.5 TCR-glia-ω2 mice, by adoptive transfer and were subsequently immunized i.g. with DGP (500 µg) and CT (10 µg). Naive CD45.1 TG2-specific B cells, isolated from CD45.1 Tgm2-/- HLA-DQ2.5 14E06 KI mice, were then transferred, followed by i.p. administration of dox (50 mg/kg) or PBS, as outlined in Figure 3A. Mice were euthanized 7 days after dox administration. (A) Representative plots of TG2-specific CD45.1+ B cells in Peyer’s patches (PPs; top) and mesenteric lymph nodes (MLNs; bottom). TG2-specific B cells are recognized by an anti-14E06 antibody targeting the 14E06 KI BCR. (B) Graphical representation of transferred TG2-specific B cells as a percentage of all B220+ B cells in PPs and MLNs. Data are pooled from four independent experiments. (C) Frequency of class switched (IgD-) TG2-specific B cells in PPs. Data are pooled from four independent experiments. (D) Immunoglobulin isotype distribution among IgD- TG2-specific B cells in PPs. A representative scatter plot showing IgG versus IgA and a summary graph with pooled data shown as mean and SD from two independent experiments with a total of five mice included. N.d. not detected.. (E) Frequency of TG2-specific B cells in PPs with a germinal center (GC) phenotype, identified as CD95⁺ CD38⁻ Ki-67⁺ Bcl-6⁺. (A-E) Data are shown with median. Ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 as determined by one-way ANOVA. Figure Supplement 2. Flow data on activation and differentiation status of TG2-specific B cells in MLN, corresponding to Figure 4C,D,E. Figure Supplement 3. Gating strategy.

Gluten-specific T cells expand and adopt a TFH-phenotype upon expression of TG2-ω18merEE following transfer of TG2-specific B cells.

Recipient mice received adoptively transferred gluten-specific CD4⁺ T cells, followed by immunization with DGP (500 μg) and CT (10 μg), adoptive transfer of TG2-specific B cells, and subsequent treatment with dox (50 mg/kg) or PBS. Mice were euthanized 7 days after dox administration. (A) Graphical representation of presence of gluten-specific TCR-glia-ω2 T cells as a percentage of all CD4+ T cells in PPs and MLNs. (B) Representative plot (left) and graphical summary (right) of the percentage of T cells with a TFH-like phenotype (CXCR5+, PD-1+) among the CD44+ CD62L- gluten-specific CD4+ T cells in PPs. (C) Representative histogram plot (left) and graphical summary (right) showing the percentage of Bcl-6⁺ cells among gluten-specific T follicular helper (TFH) cells. Bcl-6 expression within CXCR5+PD-1+-gated TFH cells (blue) relative to the CXCR5-PD-1- population (red). (A-C). Data are shown with median and are pooled from four independent experiments. Ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 as determined by one-way ANOVA. Figure Supplement 2. Supplementary data on activation and differentiation status of TG2-specific B cells and gluten-specific CD4+ T cells in gut-draining and peripheral lymph nodes. Figure Supplement 3. Gating strategy.

Gating strategy for in vitro T-cell and B-cell proliferation assay, supporting Fig 1 E.

Gating of proliferated cells at day 3 of a co-culture assay of CTV-labeled TG2-specific B cells and CFSE-labeled gluten-specific CD4+ T cells in response to IEC-lysates of dox-treated DTG mice. Cells are first gated on lymphocytes (FSC-A x SSC-A scatter) and single cells (FSC-A x FSC-H scatter). T cells are first identified as B220- and CTV- before gating on CD4 and the transgenic human Vβ3 TCR. B cells are identified as CD4- and CFSE- before gating on B220+.

Supplementary data on transferred antigen-specific B cells and T cells, supporting Figures 4 and 5.

Recipient DTG mice received gluten-specific T cells and by intragastric administration of DGP (500 µg) and CT (10 µg) 24 hours after cell transfer. After 3 days, TG2-specific B cells were transferred and recipient DTG mice received dox (50 mg/kg i.p.), as illustrated in Fig 3 A. (A) Percentage of class-switched (IgD-) TG2-specific B cells in the MLNs. Data are pooled from four independent experiments. (B) Representative scatter plot (left) showing the IgG and IgA isotype distribution among IgD- TG2-specific B cells, with the accompanying graph (right) summarizing the distribution of all immunoglobulin isotypes. Data shown as mean and SD are pooled from two independent experiments with a total of five mice included. N.d. not detected. (C) Frequency of TG2-specific B cells in the MLNs exhibiting a germinal center phenotype (CD95⁺ CD38- Ki-67⁺ Bcl-6⁺). Data are pooled from four independent experiments. (D) Graphical representation of the percentage of activated (CD62L-, CD44+) gluten-specific TCR-glia-ω2 T cells in Peyer’s patches (PPs) and MLNs. Data are pooled from four independent experiments (E) Frequency of TG2-specific B cells (left), gated on B220+, and gluten-specific T cells (right), gated on CD4+ in peripheral (non gut-draining) lymph nodes. Data are pooled from two independent experiments. (F) Representative plot (left) showing gluten-specific T cells with a TFH-like phenotype (CXCR5+PD-1+), gated on activated T cells (CD44+CD62L-), with a graph (right) summarizing the percentage of gluten-specific T cells with this phenotype. Data are pooled from four independent experiments. (G) Representative histogram (left) showing Bcl-6 expression within of CXCR5+PD-1+-gated TFH cells in MLNs (blue) compared to the CXCR5- PD-1- population (red). Graph shows quantification of the percentage of Bcl-6+ cells of four independent experiments. Ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 as determined by one-way ANOVA.

Gating strategy of adoptively transferred T and B cells, supporting Figures 4, 5 and S2.

Gluten-specific T cells, isolated from TCR-glia-ω2 mice, were transferred to CD45.2 HLA-DQ2.5 DTG mice and 24 hours later primed by intragastric administration of DGP (500 µg) and CT (10 µg). After 3 days, CD45.1+ TG2-specific B cells isolated from Tgm2-/- HLA-DQ2.5 14E06 KI mice were transferred and DTG mice were treated with dox (50 mg/kg) via i.p. injections 2 hours after transfer. These plots are from PPs, and the gating strategies shown are representative of those used for both PPs and MLNs. (A) Viable cells were recollected in PPs and MLNs by gating on lymphocytes (FSC-A x SSC-A scatter), single cells (FSC-A x FSC-H scatter), and live cells (Zombie-NIR x FSC-A). (B) Adoptively transferred B cells were recollected in PPs and MLNs by gating on viable lymphocytes (A) and B cells (B220+). TG2-specific B cells were identified by a human IgG α14E06 followed by an anti-human IgG-PE. The isotype of class-switched cells w identified as IgD-IgG1+IgA+IgM+. Germinal center cells were determined as CD38-CD95+Ki-67+Bcl-6+ (C) Adoptively transferred T cells were identified in PPs and MLNs by gating on live lymphocytes (A) and CD3+CD4+ T cells. Gluten-specific T cells were identified by the human transgenic TCR Vβ3, activated T cells as CD44+ and CD62L-, and the TFH phenotype as CXCR5+ and PD-1+. Cells with a TFH phenotype (CXCR5+PD-1+) were determined to be Bcl-6+ (blue) by using CXCR5-PD-1- as a negative control (red).