Targeting TLR4 during vaccination boosts MAdCAM-1+ lymphoid stromal cell activation and promotes the aged germinal center response

Abstract

The failure to generate enduring humoral immunity after vaccination is a hallmark of advancing age. This can be attributed to a reduction in the germinal center (GC) response, which generates long-lived antibody-secreting cells that protect against (re)infection. Despite intensive investigation, the primary cellular defect underlying impaired GCs in aging has not been identified. Here, we used heterochronic parabiosis to demonstrate that GC formation was dictated by the age of the lymph node (LN) microenvironment rather than the age of the immune cells. Lymphoid stromal cells are a key determinant of the LN microenvironment and are also an essential component underpinning GC structure and function. Using mouse models, we demonstrated that mucosal adressin cell adhesion molecule-1 (MAdCAM-1)-expressing lymphoid stromal cells were among the first cells to respond to NP-KLH + Alum immunization, proliferating and up-regulating cell surface proteins such as podoplanin and cell adhesion molecules. This response was essentially abrogated in aged mice. By targeting TLR4 using adjuvants, we improved the MAdCAM-1⁺ stromal cell response to immunization. This correlated with improved GC responses in both younger adult and aged mice, suggesting a link between stromal cell responses to immunization and GC initiation. Using bone marrow chimeras, we also found that MAdCAM-1⁺ stromal cells could respond directly to TLR4 ligands. Thus, the age-associated defect in GC and stromal cell responses to immunization can be targeted to improve vaccines in older people.

Figure 1. Age-associated defects in the GC response to immunization are driven by the microenvironment.

(A-C) Adult (8-12week old) and aged (>98 week old) Balb/c mice were immunized subcutaneously with 50μg NP-KLH emulsified in Alum and the GC response was determined 7-21d later. Shown are representative flow cytometry profiles (A), and the proportion (B) and absolute number (C) of ki67⁺Bcl6⁺ GC B cells. (D-F) Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were immunized subcutaneously with 50μg NP-KLH emulsified in Alum and the GC response was determined 7d later. Shown are representative flow cytometry profiles (D), and the proportion (E) and absolute number (F) of ki67⁺Bcl-6⁺ GC B cells. (G-J) Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were immunized subcutaneously with 50μg NP-KLH emulsified in Alum and the output of the GC response was determined 21d later. Shown are the number of low (G) and high (H) affinity IgG-producing antibody secreting cells (ASCs) in the bone marrow, and the ratio of high:low affinity ASCs (I). Serum low (NP20)- and high (NP2)-affinity antibody titres were determined seven, 14 and 21 days after immunisation in adult and aged mice (J). (K-Q) Heterochronic parabiosis was used to determine the role of the aged LN microenvironment in the generation of the GC response. Congenically distinct (CD45.1⁺ or CD45.2⁺) C57BL/6 mice were paired in isochronic (adult:adult, red; or aged:aged, blue) parabionts or heterochronic (adult:aged, purple) parabionts (K). Mouse pairs were cohoused for two weeks, then surgically joined; three weeks after surgery mice were immunized subcutaneously with 50μg NP-KLH emulsified in Alum on the outer flanks. Seven days after immunization the GC response was determined by flow cytometry (L). Blood B cell (M) and draining (d)LN non-GC B cell (N) chimerism were used to determine successful parabiosis. Mice with 40-80% self-derived B cells were used for subsequent analyses. The GC response was measured by ki67⁺Bcl-6⁺ staining (O), displayed as the proportion of B220⁺ B cells (P). Within heterochronic parabionts, B cells were gated based on donor age using congenic markers, and their capacity to generate a GC response within adult or aged hosts was determined as a proportion of B220⁺ B cells (Q). Data in (A-J) are representative of at least two independent experiments with 5-8 mice per group. Data in (K-Q) are combined from six separate experiments, using 4-20 mice per group (2-10 pairs). Data in (M, N) show individual mice with lines connecting paired blood samples. In (P, Q) each data point represents an individual mouse; individual mice from isochronic parabionts are represented as a single group. Responses to immunization are deemed independent; blood and Fo B cell samples are paired. Each symbol represents a biological replicate and lines indicate the median. Statistical significance was determined using a two-way ANOVA with Sidâk’s multiple comparison test (B,C, M, N), a Mann-Whitney rank test (E-I), a multiple unpaired t-test (J), or a one-way ANOVA with Holm-Sidâk’s multiple comparison test (P, Q). * p<0.05, **p<0.01 ***p<0.001 ****p<0.0001.

Figure 2. Lymphoid stromal cells fail to expand in response to immunization in aged mice.

(A-E) Adult (8-12week old) and aged (>98 week old) Balb/c mice were immunized as described in Fig 1A-C and stromal cell responses were determined on d7-21. MAdCAM-1⁺ lymphoid stromal cells (MAdCAM-1⁺CD21/35^-) and FDCs (CD21/35⁺) were gated on CD45^- CD31^-Pdpn⁺ stromal cells (A) and results are shown as the proportion of CD45^-CD31^-Pdpn⁺ stromal cells (B, D) and absolute number (C, E). (F-J) Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were immunized as described in Fig 1D-F and stromal cell responses were determined on d7. MAdCAM-1⁺ lymphoid stromal cells and FDCs were gated on CD45^-CD31^-Pdpn⁺ stromal cells (F) and data are shown as the proportion of CD31^-CD45^- Pdpn⁺ stromal cells (G, I) and absolute number (H, J). (K-M) Heterochronic parabiosis was performed as in Fig 1K-L; the proportion of MAdCAM-1⁺ lymphoid stromal cells and FDCs was determined as the proportion of CD45^-CD31^-Pdpn⁺ stromal cells. Each symbol represents a biological replicate and lines indicate the median. (N-U) The number of different lymphoid stromal cells was determined in LNs from naïve younger adult and aged C57BL/6 mice. Shown are the total LN cellularity (N), the number of FRCs (CD31^-CD45^-Pdpn⁺MAdCAM-1^- CD21/35^-) (O), MAdCAM-1⁺ lymphoid stromal cells (P) and FDCs (Q), alongside gating for FDCs and MAdCAM-1⁺ lymphoid stromal cells (R) and the proportion of MAdCAM-1⁺ lymphoid stromal cells (S) and FDCs (T) amongst CD31^-CD45^-Pdpn⁺ stromal cells. LNs from younger adult and aged C57BL/6 mice (U) were stained for MAdCAM-1 (magenta), Lyve-1 (green) and CD35 (yellow). Scale bars 100μm. Data in (A-J) and (N-U) are representative of at least two independent experiments with 4-8 mice per group. Data in (K-M) are combined from six separate experiments, using 4-20 mice per group (2-10 pairs). Individual mice from isochronic parabionts are represented as a single group and are regarded as independent. Statistical significance was determined using a two-way ANOVA with Sidâk’s multiple comparison test (A-C), a Mann-Whitney rank test (F-J, N-T) or a one-way ANOVA with Holm-Sidâk’s multiple comparison test (M, N). * p<0.05, **p<0.01 ***p<0.001 ****p<0.0001.

Figure 3. MAdCAM-1⁺ lymphoid stromal cells respond to NP-KLH+Alum immunization.

(A) MAdCAM-1⁺ lymphoid stromal cells and FDCs were sort-purified from adult (8-12week old) naïve C57BL/6 mice and adult (8-12week old) C57BL/6 mice immunized two days prior with 50μg NP-KLH emulsified in Alum and subjected to bulk RNA sequencing. Principal component analysis (PCA) was conducted using the 1000 most variable genes. (B-D) Data from bulk sequencing of MAdCAM-1⁺ lymphoid stromal cells derived from naïve and immunized adult mice were re-analysed separately from that of FDCs. PCA analysis was conducted using the 1000 most variable genes (B). Differential gene expression between naïve and immunized MAdCAM-1⁺ lymphoid stromal cells was determined by DESeq2 and genes with expression differences with an absolute shrunken log₂(fold change)>1 and an adjusted p value <0.05 were identified (C); orange = genes up-regulated after immunization, teal = genes up-regulated in naïve MAdCAM-1⁺ lymphoid stromal cells. Gene set enrichment analysis was conducted using the Hallmark gene sets on all expressed genes (D). Gene sets were graphed based on z-score and the adjusted p-value is indicated; orange = gene sets up-regulated after immunization, teal = gene sets up-regulated in naïve MAdCAM-1⁺ lymphoid stromal cells. (E-O) Adult (8-12week old) C57BL/6 mice were immunized subcutaneously with 50μg NP-KLH+Alum and the MAdCAM-1⁺ lymphoid stromal cell response was determined four days later in the dLN compared to ndLN controls. MAdCAM-1⁺ lymphoid stromal cells were measured as the proportion amongst CD45^-CD31^- CD21/35^-Pdpn⁺ cells (E, F) and in absolute number (G). Cell surface phenotype was determined by measuring the MFI of antibody staining for Pdpn (H), ICAM-1 (I), VCAM-1 (J), MAdCAM-1 (K) and CD44 (L) on MAdCAM-1⁺ lymphoid stromal cells. (M-O) MAdCAM-1⁺ lymphoid stromal cell proliferation was determined four days after immunization in younger adult (8-12 week old) C57BL/6 mice. The proportion of proliferating MAdCAM-1⁺ lymphoid stromal cells was determined using ki67 staining (M, N). MRC proliferation was confirmed with confocal microscopy (O), where cryosections of the dLN obtained four days after immunization were stained for ki67 (green), CD21/35 (yellow), RANKL (magenta) and DAPI (grey); scale bar 100μm. Arrow indicates ki67⁺ MRC. Data in (E-O) are representative of at least three independent experiments with 4-6 mice per group. Statistical significance was determined using a Mann-Whitney rank test; *p<0.05 **p<0.01.

Figure 4. Aging impairs the MAdCAM-1⁺ lymphoid stromal cell response to Alum-adjuvanted immunization.

(A-G) MAdCAM-1⁺ lymphoid stromal cells were sort-purified from aged (>98 week old) naïve C57BL/6 mice and aged (>98 week old) C57BL/6 mice immunized two days prior with 50μg NP-KLH emulsified in Alum and subjected to bulk RNA sequencing. PCA was conducted using the 1000 most variable genes (A). Differential gene expression between unimmunized and immunized aged MAdCAM-1⁺ lymphoid stromal cells was determined by DESeq2 and genes with significant expression differences (p<0.05) with an absolute shrunken log₂(fold change)>1 were identified (B); orange = genes up-regulated after immunization, teal = genes up-regulated in naïve MAdCAM-1⁺ lymphoid stromal cells. The transcriptional changes induced after immunization were compared for adult (Fig 3 I-K, 2881 genes) and aged (289 genes) MAdCAM-1⁺ lymphoid stromal cells (C). PCA of all naïve and immunized adult and aged MAdCAM-1⁺ lymphoid stromal cell samples was conducted using the 1000 most variable genes (D). Differential gene expression between unimmunized adult and aged MAdCAM-1⁺ lymphoid stromal cells was determined by DESeq2; genes with significant expression differences (p<0.05) and an absolute shrunken log₂(fold change)>1 were identified; teal = genes up-regulated in aged, orange = genes up-regulated in adult. Differential gene expression between day two immunized adult and aged MAdCAM-1⁺ lymphoid stromal cells was determined by DESeq2; genes with significant expression differences and an absolute shrunken log₂(fold change)>1 were identified; teal = genes up-regulated in aged, orange = genes up-regulated in adult (F). Gene set enrichment analysis was conducted on MAdCAM-1⁺ lymphoid stromal cells isolated from adult and aged mice two days after immunization using the Hallmark gene sets (G). Gene sets were graphed based on z-score and the adjusted p-value is indicated; teal = gene sets up-regulated in aged, orange = gene sets up-regulated in adult. (H-M) Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were immunized with NP-KLH+Alum and the stromal cell response determined four days later. MAdCAM-1⁺ lymphoid stromal cells were measured as the proportion of CD45^-CD31^-CD21/35^-Pdpn⁺ stromal cells (H, I) and in absolute number (J). Proliferation was measured by ki67⁺ staining (K, L). Pdpn expression by MAdCAM-1⁺ lymphoid stromal cells was measured by flow cytometry (M-N). Data in (H-N) are representative of three independent experiments with 5-6 mice per group. Statistical significance was determined using a Mann-Whitney rank test, *p<0.05 **p<0.01.

Figure 5. TLR4 stimulation promotes MAdCAM-1⁺ lymphoid stromal cells responses.

(A) RNA sequencing data from naïve younger adult (8-12week old) MAdCAM-1⁺ lymphoid stromal cells and FDCs were mined for the expression of toll-like receptor genes and downstream signalling elements associated with Tlr4, represented as log₂(RPM) counts normalized per gene (n=5), as indicted by the color scale. (B) Protein staining for TLR4 was determined by flow cytometry, showing TLR4 (filled histogram) and isotype control (empty histogram) staining on MAdCAM-1⁺ lymphoid stromal cells isolated from C57BL/6 and Tlr4^-/- mice. (C) Protein staining for CD14 was determined by flow cytometry, showing MAdCAM-1⁺ lymphoid stromal cells (red) overlaid with total live cells (grey). (D-K) Adult (8-12week old) C57BL/6 mice were administered PBS, Alum or 5μg LPS emulsified in Alum subcutaneously and the MAdCAM-1⁺ lymphoid stromal cell response was determined four days later. The MAdCAM-1⁺ lymphoid stromal cell response was measured as the proportion of MAdCAM-1⁺ cells within the CD45^-CD31^-CD21/35^-Pdpn⁺ population (D) and in absolute number (E). The proliferation (F) and upregulation of Pdpn (G), ICAM-1 (H), VCAM-1 (I), CD44 (J) and MAdCAM-1 (K), were also determined. (L-T) Adult (8-12 week old) C57BL/6 mice were administered 50μg NP-KLH in AL007 or GLA-SE and the MAdCAM-1⁺ lymphoid stromal cell response was determined four days later. MAdCAM-1⁺ lymphoid stromal cells were measured as the proportion of MAdCAM-1⁺ cells amongst CD45^-CD31^- CD21/35^-Pdpn⁺ stromal cells (L, M) and in absolute number (N). The proliferation of MAdCAM-1⁺ lymphoid stromal cells was determined by ki67 expression (O) as was the upregulation of Pdpn (P), ICAM-1 (Q), VCAM-1 (R), CD44 (W) and MAdCAM-1 (T). (U-Z) Adult (8-12week old) C57BL/6 mice were administered 50μg NP-KLH in AL007 or GLA-SE and the GC response and output was determined 21 days later. Shown are gating of GC B cells (U), and both the proportion (V) and total number (W) of GC B cells in the dLN, alongside the number of high (X) and low (Y) affinity IgG-producing ASCs in bone marrow, and the ratio of high:low affinity ASCs (Z). Data in (B, C) are representative of at least three independent experiments. Data in (D-K) are representative of two independent experiments with four mice per group. Statistical significance was determined using a one-way ANOVA using Dunnett's multiple comparison test, with Alum and Alum+LPS groups compared to PBS. * p<0.05, **p<0.01, ***p<0.001. Data in (M-Z) are representative of at least two independent experiments with five mice per group. Statistical significance was determined using a Mann-Whitney rank test, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 6. TLR4 expression by MAdCAM-1⁺ lymphoid stromal cells contributes to their activation in GLA-SE-adjuvanted immunization.

(A-G) Adult (8-12 week old) C57BL/6 and Tlr4^-/- mice were administered 50μg NP-KLH in AL007 or GLA-SE on opposing flanks and the MAdCAM-1⁺ lymphoid stromal cell response was determined four days later. MAdCAM-1⁺ lymphoid stromal cells, gated as MAdCAM-1⁺ amongst CD45^-CD31^-CD21/35^- Pdpn⁺ cells (A), were measured as both by proportion B) and in absolute number (C). MAdCAM-1⁺ lymphoid stromal cell responsiveness was measured by proliferation (D) and upregulation of cell surface markers Pdpn (E), VCAM-1 (F) and CD44 (G). (H-S) Adult (8-12 week old) C57BL/6 or Tlr4^-/- bone marrow donors were used to reconstitute irradiated CD45.1⁺ C57BL/6.SJL hosts to generate chimeric animals in which TLR4 expression is restricted to radioresistant lymphoid stromal cells. Eight weeks after reconstitution, mice were administered 50μg NP-KLH in AL007 or GLA-SE on opposing flanks and the MAdCAM-1⁺ lymphoid stromal cell and GC responses were determined four days later (H). MAdCAM-1⁺ lymphoid stromal cell responses were measured by proportion of CD31^-CD45^-Pdpn⁺ lymphoid fibroblasts (I) and in absolute number (J). MAdCAM-1⁺ lymphoid stromal cell proliferation was measured by ki67 staining (K) and their phenotype was determined by upregulation of cell surface markers: Pdpn (L), VCAM-1 (M), ICAM-1 (N) and CD44 (O). The GC response was measured by the total LN cellularity (P), the total number of B cells (Q) and the proportion (R) and total number (S) of GC B cells. Data in (A-G) are representative of at least two independent experiments with five mice per group. Paired immunizations in individual mice are linked by a line. Data in (H-S) are combined of two independent experiments, with 7-8 mice per group in each experiment. Statistical significance was determined using a paired two-way ANOVA (B-G) or a one-way ANOVA (M-T) with Sidak's multiple comparison test, comparing AL007 to GLA-SE within each chimera group, * p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 7. TLR4-adjuvanted vaccination boosts MAdCAM-1⁺ lymphoid stromal cell and GC responses in aging.

(A) RNA sequencing data from MAdCAM-1⁺ lymphoid stromal cells and FDCs isolated from naïve y Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were mined for the expression of Tlr4, Cd14 and Myd88. Data are represented as log₂(RPM) counts normalized per gene (n=5). (B-K) Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were administered 50μg NP-KLH in AL007 or GLA-SE on opposing flanks (B) and the MAdCAM-1⁺ lymphoid stromal cell response determined four days later. MAdCAM-1⁺ lymphoid stromal cells were measured as the proportion of MAdCAM-1⁺ cells amongst CD45^-CD31^-CD21/35^-Pdpn⁺ stromal cells (C, D) and in absolute number (E). The proliferation of MAdCAM-1⁺ lymphoid stromal cells was measured by ki67 expression (F), and the cell surface expression of Pdpn (G), VCAM-1 (H), ICAM-1 (I), CD44 (J) and MAdCAM-1 (K) were also determined. (L-T) Adult (8-12week old) and aged (>98 week old) C57BL/6 mice were administered 50μg NP-KLH in AL007 or GLA-SE and the GC response and output determined 21d later. Shown are flow cytometry gating (L), the proportion (M) and total number (N) of GC B cells in the dLN, the number of low (O) and high (P) affinity IgG-producing ASCs in the bone marrow, the ratio of high:low affinity ASCs (Q), as well serum antibody levels, showing low (R) and high (S) affinity antibody titres, and the ratio of high:low antibodies (T). Data in (C-T) are representative of at least two independent experiments with 5-7 mice per group. Paired immunizations in individual mice are linked by a line (D-K). Statistical significance was determined using a paired two-way ANOVA (D-K) or a one-way ANOVA (M-T) with Sidak's multiple comparison test, comparing AL007 to GLA-SE within each age group, * p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.