Initiation of primary T cell-B cell interactions and extrafollicular antibody responses in an organized microphysiological model of the human lymph node

Abstract

Antibody production is central to protection against new pathogens and cancers, as well as to certain forms of autoimmunity. Antibodies often originate in the lymph node (LN), specifically at the extrafollicular border of B cell follicles, where T and B lymphocytes physically interact to drive B cell maturation into antibody-secreting plasmablasts. In vitro models of this process are sorely needed to predict aspects of the human immune response. Microphysiological systems (MPSs) offer the opportunity to approximate the lymphoid environment, but so far have focused primarily on memory recall responses to antigens previously encountered by donor cells. To date, no 3D culture system has replicated the engagement between T cells and B cells (T-B interaction) that leads to antibody production when starting with naïve cells. Here, we developed a LN-MPS to model early T-B interactions at the extrafollicular border built from primary, naïve human lymphocytes encapsulated within a collagen-based 3D matrix. Within the MPS, naïve T cells exhibited CCL21-dependent chemotaxis and chemokinesis as predicted. Naïve T and B cells were successfully skewed on chip to an early T follicular helper (pre-Tfh) and activated state, respectively, and co-culture of the latter cells led to CD38+ plasmablast cells and T cell dependent production of IgM. These responses required differentiation of the T cells into pre-Tfhs, physical cell-cell contact, and were sensitive to the ratio at which pre-Tfh and activated B cells were seeded on-chip. Dependence on T cell engagement was greatest at a 1:5 T:B ratio, while cell proliferation and CD38+ signal was greatest at a 1:1 T:B ratio. Furthermore, plasmablast formation was established starting from naïve T and B cells on-chip. We envision that this MPS model of primary lymphocyte physiology will enable new mechanistic analyses of human humoral immunity in vitro.

Figure 1.. Modeling approach for the human LN-MPS.

(a) Immunofluorescence image of a human tonsil slice (i) with magnified view of a T/B border zone (ii). FITC-anti-Podoplanin, white; AF647-anti-CD3, magenta; PE-anti-CD19, cyan. (b) Illustration of events involved in T cell—B cell interactions in humoral immunity. Naïve T and B cells receive stimulatory signals that result in cell differentiation towards a pre-Tfh and activated B state, respectively. Pre-Tfh cells provide help to activated B cells in a TCR-MHCII-dependent manner upon exposure to cognate antigen or superantigen such as staphylococcal enterotoxin B (SEB). Activated B cells eventually mature into antibody-secreting plasmablasts. (c) Schematic of T/B border chip (not to scale) illustrating chip capabilities. (ii) Zoomed schematic and (iii) image of the 3D culture lanes, where microposts enable filling of adjacent lanes with cell-laden hydrogels via surface tension.

Figure 2.. Naïve, CD4+ T cells were viable, responded to stimuli, and responded to chemokine CCL21 on chip.

(a-c) Naïve T cells were cultured on chip with IL-7. Representative image (b) and quantification (c) of T cell viability after 4 days, labeled with (Calcein AM, green, and Dapi, blue) for 3 donors. (d-g) A CCL21 gradient was established on chip, with CCL21 added to the left-hand media lane. Representative images (e) and quantification (f) of naive CD4+ T cells after migrating toward CCL21 for 1 hr and staining with Calcein AM (green). (g) Quantification of cell velocity 30 min after gradient set up (cells were unlabeled). (h-k) Naïve T cells were cultured without (i) and with (ii) a-CD3/CD28 (StemCell). Images of CD69+ signal (FITC-anti-CD69, green) for (i) naïve and (ii) activated T cells on-chip, and quantification (j) of CD69 signal after 48 hours (unpaired T test, **:p<0.005). (k) Quantification of IFN-γ secretion by activated or naïve CD4+T cells on-chip measured by ELISA of supernatants collected at day 5. Panels f, g, k analyzed with ordinary two-way ANOVA with Sidak’s multiple comparisons test w/single pooled variance, ns: p>0.05, *:p<0.05, ****:p<0.00005.

Figure 3.. The LN MPS replicates human pre-Tfh development

(a) Naïve T cells were cultured on-chip with anti-CD3/CD28 and a cytokine cocktail (IL-7, IL-12, and Activin A) to induce differentiation into a pre-Tfh state. (b) Overlays of immunofluorescence (AF647-anti-CXCR5, magenta) and brightfield images after 3 days culture for one representative donor, 71M, with or without Tfh-skewing cocktail (i, ii). (iii) Quantification of percent CXCR5 area across each image. Each dot shows the mean from one chip. Unpaired t test, p<0.005. (c) IL-21 concentration from the chip supernatants, quantified by ELISA. Unpaired t test, p<0.001. (d) Flow cytometry characterization of pre-Tfh markers from cells recovered from chips. (i) Flow plots with (ii) quantification of the percent of CD4+ T cells expressing CXCR5 and PD-1. (iii) Flow plots with (iv) quantification of percent of CD4+ T cells expressing CXCR5 and BCL-6. Each dot is pooled 20 chips/donor. Ratio paired T test, ns:p>0.05, *:p<0.02. Bar graphs show mean ± stdev.

Figure 4.. B cell activation on chip.

(a) Schematic of naïve (i) and activated (ii) culture conditions on-chip for purified naïve CD19+ B cells. (b) Composite (fluorescence overlayed over brightfield) microscopy images of naïve (i) and activated (ii) B cells on chip with viability stain (Calcein, green, Dapi, blue) after 48 hour culture. (c) Quantification of percent viability. (d) Quantification of Calcein AM brightness across whole image. **:p<0.005, Ordinary one-way ANOVA, Tukey’s multiple comparisons. (e) Images of naïve (i) and activated (ii) B cells on chip stained with AF647-anti-CXCR5 (magenta), with quantification (iii) of percent CXCR5 area across whole image, unpaired T test, **:p<0.005. (f) Images of naïve (i) and activated (ii) B cells on chip stained with FITC-anti-CD69 (green), (iii) quantification of percent CD69 area across whole image, unpaired T test, ****:p<0.00005. Each dot is a chip. Donor information: D76M (a-e), D24F (f).

Figure 5.. IgM secretion was influenced by the presence and proximity of pre-Tfh cells when co-cultured with activated B cells.

(a) Schematic of experimental setup for comparing naïve T versus pre-Tfh cells and presence/absence of SEB on IgM secretion. (b) ELISA data showing IgM secretion, dots represent pooled supernatant from four chips per donor. (c) Schematic of experimental setup for analyzing IgM secretion dependency on proximity between pre-Tfh and activated B cells. (d) Brightfield images showing patterned lymphocytes on chip, day 1. (e) ELISA data showing IgM secretion. Results from four pooled chips from one donor, D68F. Red dotted lines are ELISA limit of detection: 0.140 ng/mL IgM.

Figure 6.. Cell expansion, B cell maturation, and IgM production was dependent on the T:B ratio.

(a) Schematic of experimental setup. (b) Representative composite images (fluorescence overlayed over brightfield) from D69M, cells stained with AF546-anti-CD38 (cyan) on day 7. (c) Change in %CD38 positive area (3 donors). (d) IgM secretion from co-culture, quantified by ELISA on day 6. N=2–3 chips/donor (5 donors). Analyzed using two-way ANOVA with Tukey’s multiple comparisons test.

Figure 7.. B cells were skewable to a plasmablast state (CD20^lo CD38^hi) when seeded in the LN-MPS from a naïve state.

(a) Schematic showing experimental setup. (b) Brightfield images showing on-chip cell density after 16 days of co-culture, post skewing. (c) Flow cytometry data for the plasmablast markers CD20 and CD38 from cells recovered from chips. Plots were gated on CD19+ B cells. (d) Total live cell count as determined by flow cytometry. (e) Percent CD20^lo CD38^hi plasmablast cells recovered from on-chip cultures. One donor, D39M, 5–6 pooled chips/condition.