KLF2 expression in IgG plasma cells at their induction site regulates the migration program

Abstract

Newly generated plasma cells in secondary lymphoid organs migrate to niches in the bone marrow, wherein they survive as long-lived plasma cells (LLPCs). Although LLPCs have been extensively characterized, it is still unclear what the key determinant(s) are for plasma cell longevity. One model postulates that plasma cell heterogeneity is established at the induction site, thereby instructing their longevity. Here, we found that, among newly generated IgG plasma cells, integrin β7hi marks plasma cells predisposed to home to the bone marrow, whereas integrin β7lo cells remain in secondary lymphoid organs. Mechanistically, this egress-prone fraction had a higher expression of the KLF2 transcription factor, the loss of which resulted in defective egress by downregulating S1PR1 and CD11b. Disruption of plasma cell egress results in defective antibody durability, thereby making mice more susceptible to influenza reinfection. Thus, the migration program of plasma cells established at the induction site plays a critical role in determining antibody durability.

Figure 1.

Integrin β7 ^hi PC s migrate from secondary lymphoid organs to BM . (A) Experimental design to analyze GC-derived IgG1⁺ PCs in the SPL or BM. GC B cells or their descendants were labeled with tdTomato following tamoxifen treatment as indicated. (B) Representative FCM histograms showing expression of various surface molecules on tdTomato⁺ IgG1⁺ PCs from SPL or BM. See Fig. S1 for gating strategy. (C) PCA of digital RNA-seq libraries from SPL integrin β7^hi or integrin β7^lo PCs, or BM PCs. Each dot represents a single RNA-seq sample (n = 4). (D) Experimental design for antibody repertoire analysis of GC-derived polyclonal PCs. (E) The clonal similarity between NP-specific integrin β7^hi or integrin β7^lo SPL and BM PCs in two independent mice. (F) The frequency of high-affinity W33L⁺ variants in NP-specific integrin β7^hi, integrin β7^lo SPL PCs, and BM PCs. Number in the center indicates total clones analyzed. (G) Representative FCM plots showing the expression of integrin β7 and CD11b on B1–8^germ PCs from SPL, blood, or BM 7 days after immunization. (H) Experimental design for the analysis of GC-derived, HA-specific IgG⁺ PCs upon influenza virus infection. (I) Representative FCM histograms showing expression of integrin β7 or CD11b on HA-specific IgG⁺ PCs from mediastinal lymph node (MLN) or BM. Data in B, G, and I are representative of two independent experiments. PCs, plasma cells; spleen, SPL; bone marrow, BM.

Figure S1.

Identification of integrin β7–expressing PCs in the SPL and BM. (A) Gating strategy for analyzing GC-derived PCs in SPL or BM. Experimental design is shown in Fig. 1 A. CD138⁺ cells were pre-enriched before surface staining. Expression of integrin β7 on CD138^hiTACI^hi CD45.1⁺tdTomato⁺ IgG1⁺ cells was analyzed. Data are representative of two independent experiments. (B) Schematic illustration of the experimental protocol to prepare PCs from SPL or BM for LEGENDScreen. (C) Scatter plot showing the surface molecules that are differentially expressed in SPL versus BM PCs. Integrin β7 and CD11b are highly expressed in BM PCs compared with SPL PCs. Donor-derived SPL or BM PCs were pooled from 9 or 50 mice, respectively. PC, plasma cell; SPL, spleen; BM, bone marrow.

Figure S2.

Generation of integrin β7 ⁺ PC s through pre- or post-GC response. (A) Experimental design for time-course analysis of generation of GC-dependent PCs in SPL. (B) The frequency of integrin β7^hi in tdTomato⁺IgG1⁺ SPL PCs (left) or the number (right) of tdTomato⁺IgG1⁺integrin β7⁺ SPL PCs (right) at the indicated time points after immunization with NP-CGG (day 9: n = 3, day 10: n = 4, day 11: n = 3, and day 12: n = 2). (C) Experimental design to analyze PCs that are generated through pre-GC (upper) or GC reaction (lower). (D) Gating strategy to analyze IgG1⁺integrin β7⁺SPL or BM PCs that were derived from pre-GC (CD138^hiTACI^hiCD45.1⁺tdTomato⁻) or post-GC response (CD138^hiTACI^hiCD45.1⁺tdTomato⁺) (left) and the frequency of integrin β7^hi in pre-GC PCs (day 7 tdTomato⁻) or post-GC PCs (day 12 tdTomato⁺) in SPL or BM (right, n = 5). (E) Experimental design to analyze generation of integrin β7⁺ PCs in a non-BCR transgenic setting. (F) Representative FCM plots showing integrin β7 expression on CD138^hiTACI^hiNP⁺IgG1⁺tdTomato⁺ cells in SPL or BM. CD138⁺ cells were pre-enriched before surface staining. Data in D and F are representative of two independent experiments. PC, plasma cell; SPL, spleen; BM, bone marrow.

Figure S3.

Survival, repertoire, and localization of antigen-specific PCs . (A) Schematic of the experimental procedure for fate mapping of PCs. (B) The number of donor-derived tdTomato⁺ PCs in SPL (left) or BM (right) at the indicated time points after immunization. Data are representative of two independent experiments. (C) Representative FCM plots of adoptively transferred integrin β7^hi or β7^lo PCs. Upper plots (“Before transfer”): Purity of donor β7^hi or β7^lo splenic PCs. Lower plots (“24 h after transfer”): Donor PCs detected in the BM 24 h after the transfer of β7^hi or β7^lo PCs. (D) Frequency of integrin β7^hi donor PCs detected in the BM (left) and the relative number of donor PCs in the BM compared with the input number (n = 4 for each donor) (right). (E) Affinity of HA-specific antibodies that were expressed in integrin β7^hi or β7^lo SPL PCs. The affinity of mAbs generated from HA-binding β7^hi or β7^lo PCs. HA-binding β7^hi or β7^lo PCs were isolated. mAbs that are derived from two clones (HV1–50 HJ4 or HV1–76 HJ1) from mouse 1 and three clones (HV1–69 HJ2, HV1–69 HJ3, or HV14-2 HJ1) from mouse 2 were generated, and their affinity against HA was measured. (F) Immunofluorescence analysis of Klf2-sufficient (+/+) or Klf2-deficient (fl/fl) PCs (CD45.1⁺CD138⁺ cells) in SPL. Experimental design is shown in Fig. 2 A. (G) gMFI of CXCR4 or CXCR5 on Klf2-sufficient (n = 6) or Klf2-deficient PCs (n = 6) in spleen. Data were analyzed by two-tailed unpaired Student’s t test. ns., not significant. Data are representative of two independent experiments. PCs, plasma cells; SPL, spleen; BM, bone marrow; gMFI, geometric mean fluorescence intensity.

Figure 2.

Klf2 is required for PC egress from the SPL to blood. (A) GSEA showing the enrichment for KLF2 target genes (left) or Klf2 mRNA expression (right), in SPL integrin β7^hi (n = 4) or integrin β7^lo PCs (n = 5). (B) Experimental setup to analyze the development of PCs from B1–8^hi R26-creERT2 Klf2^fl/fl B cells (for C–E). Klf2 in donor B cells was deleted by tamoxifen treatment. (C)Klf2 mRNA expression in SPL PCs derived from R26-creERT2 Klf2^+/+ or Klf2^fl/fl B cells after tamoxifen treatment (n = 3). (D) Representative FCM plots showing donor-derived PCs (CD45.1⁺) among CD138⁺ PCs (left), or the number of CD45.1⁺ PCs (right), in SPL, blood, or BM (n = 4). (E) Representative FCM plots showing expression of integrin β7 and CD11b (upper) or frequency of integrin β7^hi and CD11b^hi (lower) in R26-creERT2 Klf2^+/+ or Klf2^fl/fl SPL PCs (n = 5). (F) Experimental design of adoptive co-transfer of B1–8^hi R26-creERT2 Klf2^+/+ and Klf2^fl/+ B cells. (G) Representative FCM plots showing Klf2^+/+ (CD45.1/1) or Klf2^fl/+ (CD45.1/2) (left), or the frequency of Klf2^fl/+ PCs (right) among CD45.1⁺ SPL, blood, or BM PCs (n = 4). Data in A, C–E, and G are representative of two independent experiments (n = 3 for A, n = 5 for C, n = 3 or 4 for D and E, and n = 3 for G, in the second experiment). Data were analyzed by two-tailed unpaired Student’s t test (A and C–E) or one-way ANOVA followed by Tukey’s multiple comparison test (G). ns., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. PCs, plasma cells; SPL, spleen; BM, bone marrow.

Figure 3.

S1pr1 is a downstream effector molecule of KLF2 for PC egress. (A) PCA of bulk RNA-seq data from B1–8^hi integrin β7^hi or integrin β7^lo SPL PCs, or tamoxifen-treated B1–8^hi R26-creERT2 Klf2^+/+, or Klf2^fl/fl SPL PCs (n = 3). (B) Heatmap illustrating relative RNA expression of Itgb7, Itgam, Sell, Klf2, S1pr1, and Mki67 in the SPL PC populations indicated in A. (C) Representative FCM histograms showing surface S1pr1 expression (left) or its geometric mean fluorescence intensity (gMFI) (right) in SPL PC populations. n = 5 for Klf2^+/+ and n = 6 for Klf2^fl/fl. (D) Experimental design of adoptive co-transfer of B1–8^hi R26-creERT2 S1pr1^+/+ and S1pr1^fl/+ B cells. (E) Representative FCM histograms showing surface S1pr1 expression (left) or its gMFI (right) in SPL R26-creERT2 S1pr1^+/+ or S1pr1^fl/+ PCs after tamoxifen treatment (n = 4). (F) Representative FCM plots showing S1pr1^+/+ (CD45.1/1) or S1pr1^fl/+ (CD45.1/2) PCs (left), or the frequency of S1pr1^fl/+ PCs (right), in CD45.1⁺ SPL, blood, or BM PCs (n = 4). (G) Schematic illustration of the experimental protocol to induce Klf2 deletion and Klf2 or S1pr1 expression by FLEX-RV in adoptively transferred B1–8^hi R26-creERT2 Klf2^fl/fl B cells. (H) Representative FCM plots showing FLEX-RV-Klf2 or S1pr1-expressing PCs (CD45.1⁺ GFP⁺) in SPL, blood, or BM (left), or relative frequency of FLEX-RV–expressing PCs in blood or BM compared with SPL (right). n = 4 for Klf2^+/+ empty and Klf2^fl/fl S1pr1, n = 3 for Klf2^fl/fl empty and Klf2^fl/fl Klf2. Data in C, E, F, and H are representative of two independent experiments (n = 3 or 4 for C and n = 3 for E, F, and H, in the second experiment). Data were analyzed by two-tailed unpaired Student’s t test (C, left, E, and F), two-tailed paired Student’s t test (C, right), or one-way ANOVA followed by Tukey’s multiple comparisons test (F). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. PCs, plasma cells; SPL, spleen; BM, bone marrow.

Figure S4.

Heatmap analysis of RNA-seq data. Heatmap (top) illustrates 749 genes differentially expressed in Klf2^+/+ or KLf2^fl/fl SPL PCs, or β7^hi or β7^loKlf2^+/+ PCs (n = 3). Heatmap (lower left) illustrates 166 genes in cluster 1 that were upregulated in β7^hi SPL PCs. Heatmap (lower right) illustrates 61 genes in cluster 5 that were upregulated in Klf2^fl/fl or β7^lo SPL PCs. PC, plasma cell; SPL, spleen.

Figure S5.

The role of S1PR1 or CD11b in PC migration. (A) Schematic of the experimental procedure. (B) The number of donor-derived (CD45.1⁺) PCs in SPL, blood, or BM in mice treated with vehicle (n = 3) or FTY (n = 3). Data are representative of two independent experiments. (C) Experimental design of adoptive co-transfer of B1–8^hi R26-creERT2 S1pr1^+/+ and S1pr1^fl/fl B cells. (D) Representative FCM plots showing S1pr1^+/+ (CD45.1/1) or S1pr1^fl/fl (CD45.1/2) PCs (left), or the frequency of S1pr1^fl/fl PCs (right) in CD45.1⁺ SPL, blood, or BM PCs (n = 3). (E) S1PR1 expression on retrovirally transduced PCs. S1PR1 gMFI on Klf2^+/+ or Klf2^fl/fl SPL (left) or blood (right) PCs that were infected with the indicated retrovirus. n = 4 for Klf2^+/+ empty and Klf2^fl/fl S1pr1 and n = 3 for Klf2^fl/fl empty and Klf2^fl/fl Klf2. Experimental procedure is shown in Fig. 3 G. (F) Schematic illustration of the experimental protocol to induce Klf2 deletion and Klf2 or CD11b expression by FLEX-RV in adoptively transferred B1–8^hi R26-creERT2 Klf2^fl/fl B cells. (G) Representative FCM histograms showing CD11b expression in RV-transduced PCs. (H) Representative FCM plots showing FLEX-RV–expressing PCs (CD45.1⁺GFP⁺) in SPL, blood, or BM (left), or relative frequency of FLEX-RV–expressing PCs in blood or BM compared with SPL (right). n = 4 for each condition. Data in B, D, E, G, and H are representative of two independent experiments. Data were analyzed by two-tailed unpaired Student’s t test (B) or one-way ANOVA followed by Tukey’s multiple comparison test (D). ns., not significant; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. PCs, plasma cells; SPL, spleen; BM, bone marrow; gMFI, geometric mean fluorescence intensity.

Figure 4.

KLF2 is dispensable for PC entry into the BM . (A) Schematic illustration of adoptive transfer of PCs into naïve mice. SPL PCs derived from tamoxifen-treated donor B1–8^hi R26-creERT2 Klf2^+/+ or KLf2^fl/fl B cells were purified and transferred. (B) Representative FCM plots of transferred CD45.1⁺ R26-creERT2 Klf2^+/+ or Klf2^fl/fl PCs detected in BM parenchyma (CD45 i.v. negative) (left), or relative number of Klf2^+/+ or Klf2^fl/fl PCs in BM parenchyma compared with the number of transferred PCs (n = 3 or 4) at 24 h or 30 days after transfer. (C) Representative FCM histograms showing surface S1pr1 expression (left) or its gMFI in SPL, blood, or BM PCs (n = 4). Data in B and C are representative of two independent experiments (n = 3 for B and n = 4 for C, in the second experiments). Data were analyzed by two-tailed unpaired Student’s t test (C). ns, not significant. PCs, plasma cells; SPL, spleen; BM, bone marrow; gMFI, geometric mean fluorescence intensity.

Figure 5.

CD11b, but not integrin β7, contributes to PC egress. (A) Experimental setup to analyze the development of PCs from B1–8^hi R26-creERT2 Itgb7^fl/fl B cells (for B and C). Itgb7 in donor B cells was deleted by tamoxifen treatment. (B) Representative FCM histogram showing expression of integrin β7 (left) or frequency of integrin β7^hi in SPL Itgb7^+/+ (n = 4) or Itgb7^fl/fl PCs (n = 5). (C) Representative FCM plots showing donor-derived PCs (CD45.1⁺) in CD138⁺ PCs in SPL, blood, or BM (left) and number of CD45.1⁺ donor-derived PCs in SPL, blood, or BM (right). n = 4 for Itgb7^+/+ and n = 5 for Itgb7^fl/fl. (D) Experimental setup to analyze the development of PCs from B1–8^germ B cells expressing CD11b-targeting shRNA (for E and F). (E) Representative FCM plots showing expression of CD11b (left), or the frequency of CD11b^hi (right) in SPL PCs transduced with shControl (n = 4) or shCD11b (n = 3). (F) Representative FCM plots showing shRNA-transduced PCs (CD45.1⁺GFP⁺) in SPL, blood, or BM (left) and the ratio of shRNA-transduced PCs in blood to SPL or BM to SPL (right). n = 4 for shControl and n = 3 for shCD11b. (G) FLAG-KLF2 binding to the Itgb7, S1pr1, Itgam, or Cd4 locus determined by ChIP-qPCR using Klf2-deficient PCs transduced with empty (open circles, n = 3) or 3xFLAG-KLF2 RV (closed red circles, n = 3) during their in vitro generation. Data in B, C, E, and F are representative of two independent experiments (n = 3 or 4 for B, C, E, and F, in the second experiment). Data were analyzed by two-tailed unpaired Student’s t test (B, C, and E–G). ns., not significant; **P < 0.01, ***P < 0.001, and ****P < 0.0001. PCs, plasma cells; SPL, spleen; BM, bone marrow.

Figure 6.

Klf2 in PC s is required for antibody-mediated protection from influenza infection. (A) Schematic illustration of the experiments to examine the protective ability of serum antibodies against influenza infection. (B and C) Mixed BM chimeras were generated with BM cells from μMT and S1pr2-creERT2 Klf2^+/+ or Klf2^fl/fl mice, vaccinated with inactivated whole PR8 virus, followed by tamoxifen treatment and collection of serum for measurement of anti-HA titers (B) or for transfer into naïve mice (C). (B) High avidity anti-HA IgG responses at the indicated time points after influenza vaccination in S1pr2-creERT2 Klf2^+/+ chimeras (closed circles, n = 10) or S1pr2-creERT2 Klf2^fl/fl chimeras (closed squares, n = 13). (C) Survival of mice that received sera from naïve mice (open circles, n = 6), vaccinated S1pr2-creERT2 Klf2^+/+ chimeras (closed circles, n = 8), or vaccinated S1pr2-creERT2 Klf2^fl/fl chimeras (closed squares, n = 8) before challenge with a high dose of influenza virus. Data in B and C are representative of two independent experiments (n = 10 for each group in B and n = 5 for each group in C, in the second experiment). Data were analyzed by two-tailed unpaired Student’s t test (B) or log-rank (Mantel–Cox) test (C). ns, not significant; PCs, plasma cells; BM, bone marrow. **P < 0.01, ***P < 0.001, and ****P < 0.0001.