S1P-Dependent trafficking of intracellular yersinia pestis through lymph nodes establishes Buboes and systemic infection

Abstract

Pathologically swollen lymph nodes (LNs), or buboes, characterize Yersinia pestis infection, yet how they form and function is unknown. We report that colonization of the draining LN (dLN) occurred due to trafficking of infected dendritic cells and monocytes in temporally distinct waves in response to redundant chemotactic signals, including through CCR7, CCR2, and sphingosine-1-phospate (S1P) receptors. Retention of multiple subsets of phagocytes within peripheral LNs using the S1P receptor agonist FTY720 or S1P1-specific agonist SEW2871 increased survival, reduced colonization of downstream LNs, and limited progression to transmission-associated septicemic or pneumonic disease states. Conditional deletion of S1P1 in mononuclear phagocytes abolished node-to-node trafficking of infected cells. Thus, Y. pestis-orchestrated LN remodeling promoted its dissemination via host cells through the lymphatic system but can be blocked by prevention of leukocyte egress from DLNs. These findings define a novel trafficking route of mononuclear phagocytes and identify S1P as a therapeutic target during infection.

Figure 1. Buboes Contain Intracellular Y. pestis

(A) H&E-stained LN sections, 24 hr after injection of saline or 1 × 10⁵ CFU of the designated bacteria, illustrate the distinction between a normal hypertrophic LN (E. coli) and the early stages of a bubo (Y. pestis). Image was obtained at 10× magnification. (B) A LN sinus from the medullary region of an H&E-stained LN, 24 hr after infection of Y. pestis (40× magnification). (C) Intracellular rod-shaped bacilli are visible within a LN (left), imaged by TEM, 72 hr after Y. pestis injection in the footpad. The right panel is a lower magnification of this area demonstrating that the infected cell is morphologically consistent with a macrophage. Images are representative of ≥2 independent experiments.

Figure 2. Two Waves of Infected Cellular Compartments within dLNs

(A) The relative phenotype of infected cells within dLNs based on staining for surface markers CD11c and CD11b as a percentage of total infected cells, determined by flow cytometry using YP-OFP to detect infected cells (n ≥ 3). Populations of single-positive (SP) and double-positive (DP) cells for each marker are presented. Abbreviation: ND, not detected. For corresponding total cell numbers see Figure S2A. (B) Percentages of YP-OFP⁺ cells according to phenotype 24 hr after infection after footpad infection with 5 × 10⁵ YP-OFP. n = 5; data are representative of two independent experiments. (C) Subsets of infected DCs (OFP⁺CD11c⁺IA/IE⁺) defined by flow cytometry as a percentage of total infected DCs. (D) The total numbers of iMono, moDC, and PMNs in either YP-OFP-infected or resting LNs were determined by flow cytometry at 24 hr. Data are representative of two independent experiments with n = 5. The ratio of each cell type is presented above the corresponding bars: resting:infected. for (B)–(D), gating strategies are included as in Figures S2C and S2D. (E) LN sections at 24 and 72 hr show a shift in the phenotype of the predominately infected cell (containing YP-OFP, red) from CD11c⁺ (green) to CD11c⁻. In both images at least some infected cells appear to be CD11b⁺ (blue). All images are representative of three or more independent experiments with n ≥ 3. Colocalization images are provided in Figure S2E. (F) Real-time PCR was performed to quantify mRNA levels for the chemokines CCL2, CCL3, CCL4, and CXCL10 present in LNs infected with E. coli, S. Typhimurium, or Y. pestis. Data are representative of three independent experiments performed with RNA from individual LNs at each time point. For a time course of chemokine production in dLNs, see Figure S2E.

Figure 3. CCR7- and CCR2-Directed Migration Enhance Y. pestis Virulence In Vivo

(A and B) Graph represents the total numbers of Y. pestis-infected cells (OFP⁺) within popliteal DLNs at 24 hr after infection with (A) Ccr2^−/− mice and C57BL/6 controls or (B) Ccr7^−/− mice and BALB/c controls. The cell counts for OFP⁺ subsets that were single or double positive for surface markers CD11c and CD11b were determined by flow cytometry. p < 0.0001 by two-way ANOVA. (C) Relative CCL2 expression in dLNs, 24 hr after infection of C57BL/6 mice with Y. pestis. Mice were administered an anti-CCL21 or isotype control (IC) antibody treatment 12 hr prior to infection. Real-time PCR data are representative of three independent experiments using RNA isolated from single LNs; p < 0.05 by Student’s t test. (D) Total numbers of CD11c⁺ cells in the dLNs of mice were determined by flow cytometry, 24 hr after Y. pestis infection and with pretreatment with either IC or anti-CCL21 blocking antibody. Pretreatment with anti-CCL21 reduced the numbers of CD11c⁺ cells in dLNs; p < 0.05 by Student’s t test with n = 3. (E) Significantly reduced CFUs were recovered from dLNs after pretreatment with anti-CCL21 blocking antibody compared to IC pretreatment, determined by Student’s t test (p = 0.0006, n = 4). (F) CFU data obtained day 3 after infection by isolating, homogenizing, and plating Y. pestis from spleen (left) or lung (right) tissue from Ccr2^−/− or C57BL/6 control mice. Differences are not significant by ANOVA (p = 0.35); n = 6, pooled from two independent experiments. ND indicates not detected. (G) Survival of Ccr2^−/− and C57BL/6 mice after subcutaneous infection with Y. pestis (n = 9–10, pooled from two independent experiments). Ccr2^−/− mice survived significantly longer; p = 0.009. (H) CFU data for the spleen (left) and lung (right) for Ccr7^−/− mice and BALB/c controls. Differences are not significant by ANOVA (p = 0.06); n = 6–7, pooled from two independent experiments. (I) Survival of Ccr7^−/− and Balb/c mice after subcutaneous infection with Y. pestis (n = 5–10, pooled from 2 independent experiments). Ccr7^−/− mice survived significantly longer; p = 0.0002. For all infections in this Figure, 1 × 10⁵ CFU of Y. pestis were injected in footpads for Ccr2^−/− and C57BL/6 mice and 1 × 10⁷CFU of Y. pestis were injected for Ccr7^−/− and Balb/c mice.

Figure 4. Intracellular Y. pestis Travel within Lymphatics between LNs

(A) Merged and channel series staining of mouse thigh tissue to show post-popliteal nodal lymphatic vessels (Lyve-1, green), monocytes (CD11b, blue), and OFP⁺ Y. pestis (red). Colocalization reveals that some, but not all, CD11b⁺ cells traveling in the postnodal lymphatics appear to contain Y. pestis. Images are representative of observations from two independent experiments with two mice per group. (B) Graph represents the percentage of Y. pestis within the iliac node that are intracellular after footpad infection with 1 × 10⁵ CFU. Values were obtained using a gentamicin protection assay, where a portion of the LN homogenate was treated with gentamicin to kill only the extracellular bacteria followed by washing and treatment to lyse cells. The surviving intracellular fraction was plated to determine the CFU, which is represented as a percentage of the total CFU for the organ. (C) Graph of total and intracellular Y. pestis, corresponding to (B). Blue shading illustrates the growth in numbers of extracellular bacteria over time by highlighting the difference between total and intracellular bacterial counts. Data are representative of two independent experiments with n = 3–5 animals per group.

Figure 5. S1P Promotes Y. pestis Dissemination through Lymphoid Tissue

Animals were pretreated 24 hr before infection and then administered a daily regimen of S1P-agonist FTY720 or vehicle. (A and B) Graphs represent the total numbers of cells staining single- (SP) and double-positive (DP) for markers CD11b and CD11c in the dLN, the popliteal LN(A) or in iliac LNs distal to dLNs (B) 24 hr after infection with Y. pestis. (C and D) Graphs represent the number of cells infected by Y. pestis in the popliteal (C) and iliac (D) LNs, after gating on the OFP⁺ population. Infected CD11b⁻ CD11c⁺ were not detected in iliac nodes at 24 hr. (E and F) The graphs represent the numbers of footpad-derived cells in popliteal (E) and iliac (F) LNs that were infected with Y. pestis, by gating on cells CFSE⁺ and OFP⁺. CFSE labeling was performed 4 hr prior to infection with Y. pestis expressing OFP. (G and H) Graphs depict numbers of CD11b⁺CD11c^+/− cells within popliteal (G) and iliac (H) LNs of mice depleted of T cells prior to infection and treatment with FTY720 or vehicle control. (I) The CFU counts from the spleens and lungs, harvested day 3 of infection, from animals treated with FTY720 or vehicle are presented. (J and K) Graphs represent the total number of cells with phenotypes SP or DP for CD11b and CD11c in popliteal (J) or iliac (K) LNs, 24 hr after infection and with treatment with S1P₁ receptor agonist SEW2871. (L and M) Numbers of footpad-derived cells (CFSE⁺OFP⁺) in the popliteal (L) and iliac LNs (M) after SEW2871 treatment or vehicle control. (N and O) CFUs were determined in the popliteal LN (N) and iliac LN (O), 24 hr after infection, as above. For (A)–(H) and (J)–(M), significance was determined by Student’s unpaired t test; *p ≤ 0.05. n = 3–6 per group. For (I), significance was determined by 2-way ANOVA; p < 0.0001. For (N) and (O), p < 0.05 by 1-way ANOVA with Bonferroni’s post-test. (P) Mice were monitored for survival after infection with Y. pestis and treatment with FTY720, SEW2871, or vehicle control (n = 5–10 per group). Survival curves of infected mice treated with FTY720 and SEW2871 both differed significantly from the survival curve of mice treated with vehicle alone (p = 0.04 and 0.01, respectively, by log-rank test). Surviving proportions also differed significantly; p = 0.04 and 0.01, respectively, by Fisher’s exact test.

Figure 6. Conditional Deletion of S1P₁ in Mononuclear Phagocytes Limits Intranodal Y. pestis Spread

(A) S1P1 expression on CX₃CR1⁺ and CX3CR1⁻ cells in CX₃CR1-Cre S1P₁^fl/fl mice compared to CX₃CR1-Cre S1P₁^+/+ littermate controls, determined by flow cytometry. (B) The graphs represent the numbers of cells in popliteal and iliac LNs that were infected with Y. pestis, by gating on OFP⁺ cells. Statistics were performed by 2-way ANOVA, where n = 4 mice per group (p < 0.01). Not significant = ns. Data are representative of two independent experiments. (C) Survival of CX₃CR1-Cre S1P₁^fl/fl and CX₃CR1-Cre S1P₁^+/+ littermate controls (n = 7–9) after footpad infection with 1 × 10⁵ CFU of Y. pestis; p = 0.01 by log-rank test.