A role for gut-associated lymphoid tissue in shaping the human B cell repertoire

Abstract

We have tracked the fate of immature human B cells at a critical stage in their development when the mature B cell repertoire is shaped. We show that a major subset of bone marrow emigrant immature human B cells, the transitional 2 (T2) B cells, homes to gut-associated lymphoid tissue (GALT) and that most T2 B cells isolated from human GALT are activated. Activation in GALT is a previously unknown potential fate for immature human B cells. The process of maturation from immature transitional B cell through to mature naive B cell includes the removal of autoreactive cells from the developing repertoire, a process which is known to fail in systemic lupus erythematosus (SLE). We observe that immature B cells in SLE are poorly equipped to access the gut and that gut immune compartments are depleted in SLE. Thus, activation of immature B cells in GALT may function as a checkpoint that protects against autoimmunity. In healthy individuals, this pathway may be involved in generating the vast population of IgA plasma cells and also the enigmatic marginal zone B cell subset that is poorly understood in humans.

Figure 1.

A subpopulation of transitional B cells is recruited to the GALT. (A and B) Healthy peripheral blood and Peyer’s patch B cells were analyzed by flow cytometry for the expression of CD19, CD20, CD27, CD24, CD38, CD10, and IgD. FACS plots display the percentages of T1, T2, and naive B cells, and histograms display CD10 and IgD expression for each subset, healthy control blood (A) and Peyer’s patches (B). Plots show one of four experiments with similar results. (C) Confocal microscopy of human Peyer’s patch. CD10⁺ (red), IgD⁺ (green) transitional B cells are highlighted by arrows at the periphery of the germinal center. GC, germinal center; MZ, marginal zone. One of three experiments with similar results is shown. (D) Immunohistochemistry of GALT identifies CD10⁺ (pink) and IgD⁺ (brown) transitional B cells in the periphery of the germinal center. One of three experiments with similar results is shown. (E) Immunohistochemistry of GALT showing CD10⁺ cells (brown) scattered in the mantle zone of GALT. One of three experiments with similar results is shown. (F and G) T1, T2, and naive B cells were identified by flow cytometry (as in A and B) in F. Perfusates from normal liver grafts and matched donor blood (G) are shown. One of three experiments with similar results is shown. (H) Ratios of the %T2 to %T1 subpopulations in normal blood, matched liver donor blood, liver graft perfusates, and normal Peyer’s patches. Error bars show mean ± SEM. Statistical test: one-way ANOVA. n = up to 15. ***, P < 0.0001; **, P < 0.003. (I) Immunohistochemistry showing MAdCAM (brown) and CD20 (pink). MAdCAM is seen on Peyer’s patch vessels (marked by arrowheads). One of two experiments with similar results. (J and K) Healthy control blood, liver perfusates, and matched donor blood samples were analyzed by flow cytometry for the expression of CD19, CD20, CD27, CD24, CD38, and β7 integrin. One of three experiments with similar results is shown. (J) FACs plots display β7 expression on T1, T2, and naive B cells (K) Summary data showing percentages of naive, T1, and T2 B cells in normal peripheral blood and liver graft perfusates that express β7 integrin. n = 7–16. Error bars show mean ± SEM. Statistical test: one-way ANOVA. ***, P < 0.0001; **, P < 0.0043.

Figure 2.

GALT transitional B cells are activated in vivo and in vitro in response to intestinal bacteria. (A) Mononuclear cells were isolated from Peyer’s patches and analyzed directly ex vivo by flow cytometry for their expression of CD19, CD10, IgD, and phospho-BTK, phospho-Syk, and phospho-ERK. Representative histograms displaying the degree of BTK, Syk, and ERK phosphorylation for CD19⁺CD10⁺IgD⁺ transitional (TS) and CD19⁺CD10⁺IgD⁻ germinal center (GC) B cells. Shown is one of three experiments with similar results. (B) PBMCs were stimulated polyclonally for 30 min with CpG, anti-IgM, and heat-inactivated intestinal bacteria as a positive control for B cell kinase phosphorylation, or left unstimulated. After stimulation, cells were analyzed by flow cytometry for the expression of CD19, CD20, CD27, CD24, CD38, and phospho-BTK, phospho-Syk, and phospho-ERK. Histograms display the degree of BTK, Syk, and ERK phosphorylation for T1, T2, and naive B cells. Histograms show one of three experiments. (C) FACS-sorted T1 cells (CD19⁺ CD20⁺ CD27⁻ CD24⁺⁺ CD38⁺⁺), T2 cells (CD19⁺ CD20⁺ CD27⁻ CD24⁺ CD38⁺), and naive mature B cells (CD19⁺ CD20⁺ CD27⁻ CD24⁻ CD38⁻) isolated from normal blood were incubated with combinations of heat-inactivated intestinal bacteria, CpG, and anti-IgM for 30 min, followed by intracellular staining for phospho-BTK. Mean values + SD from three independent experiments are shows. (D and E) Autoantibody secretion by in vitro stimulated peripheral blood and Peyer’s patch B cells. T1, T2, and naive mature B cells were FACS sorted (sort strategy as in Fig. S2 A) from normal blood and Peyer’s patches (one experiment with pooled cells from three donors) and were incubated with heat-inactivated intestinal bacteria for 8 d. Supernatants were screened for secreted autoantibodies by Hep-2 ELISA. Graphs show Hep-2 binding IgG (D) and IgA (E) autoantibodies. Data are standardized so that reactivity of supernatant from blood naive cells is 1 for each isotype. (F and G) IGHV genes were amplified by PCR from FACS-sorted TS and GC B cells isolated from Peyer’s patches of seven different donors (gated and sorted as shown in Fig. S2 B) from two sorts. (F) Pooled data of the number of mutations in GC or TS IGHV genes. Mann-Whitney test, P = 0.0001. (G) Number of mutations in GC or TS IGHV genes segregated by donor. Horizontal bars show the mean number of mutations. (H and I) An example of mutations in clonally related germinal center B cells. (H) Examples of IGHV sequencing of germinal center B cells showing four clonally related sequences deriving from a common ancestor. (I) Graphical explanation of the relation of the clones in H to a common ancestor and their germline precursor.

Figure 3.

Expression of β7 integrin on TS B cells is impaired in SLE. (A) PBMCs isolated from healthy donors and patients with SLE were analyzed by flow cytometry for the expression of CD19, CD20, CD27, CD24, CD38, and β7. Graph displays the percentages of β7⁺ healthy or SLE T1, T2, and mature B cells. ***, P < 0.001 one-way ANOVA + Bonferroni’s multiple comparison test. (B) PBMCs isolated from SLE and RA patients were analyzed by flow cytometry as in A. Graph shows percentages of β7⁺ SLE and RA B cell subsets segregated into patients treated with or without hydroxychloroquine. Statistics: Kruskal-wallis + Dunn’s multiple comparison test. (C) Peripheral blood B cells from healthy individuals and SLE patients were cultured for 7 d with indicated treatments (10 µg/ml anti-IgM, 1 µM CpG, 100 nM RA, and 1 µM Ro41). Graph shows β7 expression analyzed by flow cytometry. Means + SEM of three independent experiments are shown. P = 0.04, Mann-Whitney-test. (D) Intestinal biopsies were taken from eight healthy and eight SLE patients (up to eight biopsies per patient). The biopsies were stained by immunohistochemistry for CD3, IgD, or CD20 and were scored according to whether sections contained no lymphoid tissue (diffuse lymphoid infiltrates), diffuse lymphoid aggregates only (aggregates), or structured lymphoid tissue (GALT). GALT in biopsies from healthy individuals versus SLE P < 0.004, χ². (E) Representative histological pictures of a healthy biopsy displaying GALT and an SLE biopsy showing a diffuse lymphoid aggregate lacking GALT features. CD20, IgD, and CD3 are stained brown. Bar, 75 µm. (F) Biopsies in D were stained for the presence of IgA⁺ plasma cells identified by their morphology and cytoplasmic IgA. Graphs show mean plasma cell density. Each data point is a mean of several biopsies from each donor, P = 0.03, Mann-Whitney-Test. (G) PBMCs isolated from SLE patients and healthy controls were analyzed by flow cytometry for CD3 and β7 expression. Graph displays the mean percentage of CD3⁺ T cells expressing β7. P = 0.0078, Mann-Whitney-Test. (H) Sections of intestinal biopsies isolated from the Ileum and colon of healthy controls and patient with SLE were stained with CD3. Graph shows mean percentages of intraepithelial lymphocytes (IEL) as percentage of total epithelial cells and IEL. *, P = 0.03; **, P = 0.004, Mann-Whitney Test. (I) Representative examples of immunohistochemistry of intestinal biopsies from healthy controls or SLE patients, stained for CD3 (brown) or CD4 (brown) and CD8 (pink). Bar, 20 µm. (J) Percent of β7-expressing CD19⁺ blood B cells after 7-d culture with indicated treatments (α-IgM 10 µg/ml; CpG 1 µM; RA 100 nM; E50/100/200 = β-estradiol 50/100/200 pg/ml). Analysis was performed by flow cytometry. Shown are means + SEM of three independent experiments. *, P = 0.02, two-tailed Student’s t test. (K) Flow cytometry analysis of β7-expressing CD19⁺ blood B cells after 7-d culture with indicated treatments (αIgM 10 µg/ml; CpG 1 µM; RA 100 nM; E50/100/200 = β-estradiol 50/100/200 pg/ml). The figure shows representative FACS plots of one of three experiments with similar results. (L) Transitional B cells were FACS sorted from the peripheral blood of healthy controls and patients with SLE based on their expression of CD19, IgD, and CD38. Sorted CD19⁺IgD⁺CD38⁺ transitional B cells were analyzed for RARalpha transcripts by RT-PCR, n = 6. The graph shows mean relative copy number for each group. No difference was observed. (M) Flow cytometric analysis of ALDH enzyme activity in blood myeloid dendritic cells (lineage⁻ HLA-DR⁺ CD11c⁺) from lupus and control blood samples using the Aldefluor reagent system. Mean percentages of myeloid DCs that were ALDH⁺ are shown. n = 5. P = 0.3, two-tailed Mann-Whitney-test.