Type II membrane protein CD69 regulates the formation of resting T-helper memory

Abstract

Memory T-helper (Th) lymphocytes are crucial for the maintenance of acquired immunity to eliminate infectious pathogens. We have previously demonstrated that most memory Th lymphocytes reside and rest on stromal niches of the bone marrow (BM). Little is known, however, regarding the molecular basis for the generation and maintenance of BM memory Th lymphocytes. Here we show that CD69-deficient effector CD4 T lymphocytes fail to relocate into and persist in the BM and therefore to differentiate into memory cells. Consequently, CD69-deficient CD4 T cells fail to facilitate the production of high-affinity antibodies and the generation of BM long-lived plasma cells in the late phase of immune responses. Thus, CD69 is critical for the generation and maintenance of professional memory Th lymphocytes, which can efficiently help humoral immunity in the late phase. The deficit of immunological memory in CD69-deficient mice also highlights the essential role of BM for the establishment of Th memory.

Immunological memory is a defining characteristic of the adaptive immune response and can be manifested by CD4 and CD8 T cells and B cells. The generation and maintenance of memory T-helper (Th) lymphocytes is crucial for acquired immunity to eliminate various infectious pathogens (1–3). In their absence, the generation of high-affinity memory B cells and long-lived plasma cells (4–6) and the maintenance and secondary expansion of memory CD8 T cells (7–10) are impaired. We have previously shown that Ly-6C is a specific marker of memory Th cells in bone marrow (BM) (11). Upon challenge with antigen, Ly-6C^hi memory Th cells rapidly express cytokines and CD154 (CD40L) and efficiently induce the production of high-affinity antibodies by B cells. Despite their eminent importance for the regulation of immune reactions and immunological memory, the molecular mechanisms regulating relocation of Th memory precursors to the BM have not been investigated.

CD69 is a type II membrane protein expressed as a homodimer composed of heavily glycosylated subunits (12). T cells express CD69 rapidly upon stimulation of the T-cell receptor (TCR) (13, 14), which is why CD69 has been mostly regarded as an activation marker (15, 16). The precise role of CD69 in immunity has not been determined because its ligand is unknown. Freshly prepared thymocytes undergoing selection events express CD69, and regulatory roles for CD69 expression in T-cell development in the thymus have been suggested (17, 18). However, phenotypical analysis in previous studies using CD69-deficient mice has revealed that CD69 does not appear to be required for the development of CD4 T cells (19, 20). Although multiple target processes of CD69 have been suggested in the effector phase of an immune response, such as in an anticollagen antibody-induced arthritis model (21) and allergic airway inflammation model (20), its function in immunological memory has not been elucidated.

Here we show the role of CD69 in the generation of memory Th cells and investigate the ability of CD69-deficient Th cells to facilitate humoral immunity. CD69-deficient CD4 T cells failed to form memory cells because of defective relocation into, and persistence in, the BM. Moreover, the Th memory-deficient mice could not produce high-affinity antibodies and generate long-lived plasma cells. Thus, we have clarified the role of CD69 in Th memory and further defined the role of Th memory in humoral immunity, thus providing strategies for modulating T-B interactions in immunological memory.

Results

CD69 Is Required for Generation of Memory Th Cells.

In steady state, CD69 was expressed on, at most, 2% of CD44^lo, naive CD4 T cells and 20% to 30% of CD44^hi CD4 T cells in the spleen and BM, whereas it was expressed on approximately 70% of BM Ly-6C^hiCD44^hi CD4 T cells, which have been identified as a population including professional resting memory Th cells (Fig. S1A). Despite the constitutive expression of CD69, BM CD69⁺CD44^hi CD4 T cells were not activated in terms of DNA synthesis (Fig. S1 B and C), nor overall transcription, as we have shown previously (11).

To investigate the relevance of CD69 expression and function in memory Th cells, we monitored antigen-specific CD4 T cells after immunization in an adoptive transfer model (22). Naive CD4 T cells are activated in the secondary lymphoid organs immediately after immunization and, within 8 wk, some of the activated cells migrate to the BM via blood flow and persist there as memory cells for a long time (11). Naive normal mice were transferred with CD69-deficient or WT DO11.10 transgenic (Tg) naive CD4 T cells, immunized with ovalbumin (OVA) plus lipopolysaccharide (LPS) and then analyzed for the transferred CD4 T cells. Four days after immunization, activated CD69-deficient or WT DO11.10 Tg CD4 T cells were similarly detectable in the spleen, whereas none were detected in the BM of either group (Fig. 1 A and B and Fig. S2A). In the memory phase, CD69-deficient DO11.10 Tg CD4 T cells did not accumulate in the BM compared with WT cells as detected by flow cytometry (Fig. 1A) and by histological analysis (Fig. 1B), although most WT and CD69-deficient DO11.10 Tg cells disappeared from the spleen (Fig. 1A). Most WT DO11.10 Tg CD4 T cells of the BM in the memory phase express CD69, as do BM Ly-6C^hiCD44^hi CD4 T cells in steady status (Fig. 1C and Fig. S2A). To confirm the defect of BM memory Th cells in CD69-deficient mice, we alternatively analyzed the immune response of WT and CD69-deficient mice to 4-(hydroxy-3-nitrophenyl) acetyl-coupled (NP)-OVA plus incomplete Freund adjuvant (IFA), by monitoring OVA-specific CD4 T cells based on their expression of IFN-γ. In the memory phase of the immune response, significantly fewer OVA-reactive IFN-γ–producing CD4 T cells were detected in BM of CD69-deficient mice compared with WT mice (Fig. 1D). Thus, CD69 is required for the establishment of professional resting Th memory in the BM.

Fig. 1.

CD69 is required for the generation of memory Th cells. (A–C) CD4 T cells (3 × 10⁵) from CD69^+/+ (open column) and CD69^−/− (filled column) DO11.10 Tg mice were transferred into BALB/c mice, and the mice were then injected with 100 μg of OVA plus 10 μg LPS. (A) Activated CD69^−/− OVA-specific CD4 T cells do not accumulate in BM. The number of OVA-TCR⁺ donor T cells was determined by flow cytometry. Representative OVA-TCR/CD44 staining profiles of B220⁻ CD4 T cells in spleen and BM of host mice at days 4 and 48 after immunization are shown (Left). The cell numbers (mean ± SD) of donor T cells was quantified as OVA-TCR⁺B220⁻CD44^hi CD4 T cells (Right). (B) Few activated CD69^−/− OVA-specific CD4 T cells are observed in BM. Representative histological section of OVA-TCR (green) vs. laminin (red) labeling in BM of the host mice at day 56 after transfer are shown (Left). Outlined area at the left is enlarged (Center). The absolute number of OVA-TCR⁺ CD4 T cells per square millimeter of BM was quantified (Right; n = 5). (C) Most antigen-specific memory Th cells of BM express CD69. OVA-TCR⁺ CD4 T cells in the spleen and BM of the host mice at day 48 after immunization were stained with anti-CD69 antibodies (background staining with an isotype control shown in gray) and analyzed by flow cytometry. The inserted numbers indicate the mean fluorescence intensity (n = 3). (D) Antigen-specific functional memory Th cells are impaired in CD69-deficient mice. At day 90 after immunization with 100 μg NP-OVA plus IFA, OVA-specific CD4 T cells of spleen and BM of CD69^+/+ (open column) and CD69^−/− (filled column) mice were quantified by flow cytometry. For detecting OVA-reactive CD4 T cells, IFN-γ⁺CD44^hi CD4 T cells were counted after ex vivo stimulation with NP-OVA for 5 h. The frequencies of IFN-γ⁺ cells among CD44^hi CD4 T cells of the spleen and BM are shown (mean ± SD; n = 5). All data are representative of two or more independent experiments (*P < 0.05, **P < 0.01, and ***P < 0.001).

CD69 inhibits sphingosine-1-phosphate (S1P)/S1P receptor 1 (S1P₁)-mediated cell egress from lymphoid organs after exposure to IFN-α/β (23). Accordingly, inhibition or deficiency of CD69 can be expected to enhance the movement of activated CD4 T cells to the periphery. This, however, was not observed in the present study; rather, the number of antigen-specific CD4 T cells in the peripheral blood was comparable on days 4, 15, and 48 after immunization between WT and CD69-deficient CD4 T-cell-transferred mice (Fig. S3).

CD69-Deficient CD4 T Cells Fail to Provide Efficient Help for B Cells.

To investigate the function of CD69 in Th cells, we tested whether CD69-deficient CD4 T cells can help B cells for antibody production in vivo. In immune responses to NP-coupled chicken γ-globulin (NP-CGG) plus LPS or NP-coupled keyhole limpet hemocyanin (NP-KLH) plus alum, CD69-deficient mice produced significantly less NP-specific high-affinity, but not total, antibodies compared with WT mice, on days 28 and 90 after immunization of NP-CGG (Fig. S4A) or on day 28 after immunization of NP-KLH (Fig. S4B), whereas, on day 14, the levels of anti–NP-CGG (Fig. S4A) and total serum IgG and IgG₁ (Fig. S4C) were comparable. Next, CD4 T cells were sorted from spleens of CD69-deficient or WT DO11.10 Tg mice and transferred into normal mice, and then the transferred mice were immunized with NP-OVA and analyzed for NP-specific antibodies (Fig. 2A and Fig. S2A). CD69-deficient CD4 T cells failed to induce the production of high-affinity antibodies, especially in the late phase (Fig. 2A). In contrast, CD69-deficient mice transferred with WT DO11.10 Tg CD4 T cells could normally produce NP-specific high-affinity antibodies (Fig. S4D). Thus, the defective production of high-affinity antibodies in CD69-deficient mice appeared to be caused by the lack of CD69 expression on CD4 T cells.

Fig. 2.

Impaired help for B cells by CD69-deficient CD4 T cells in the late phase. (A) CD69-deficient CD4 T cells fail to induce the production of high-affinity antibodies in vivo. CD4 T cells (3 × 10⁵) from CD69^+/+ and CD69^−/− DO11.10 Tg mice were transferred into BALB/c mice, and the mice were then immunized i.p. with 100 μg NP-OVA plus 10 μg LPS. Blood taken at each time point was analyzed for anti–NP₃₆-IgG₁ and anti–NP₄-IgG₁ by ELISA (n = 5–7). (B) CD69^−/− CD4 T cells fail to provide efficient help for the generation of long-lived plasma cells in vivo. CD4 T cells (3 × 10⁵) from CD69^+/+ and CD69^−/− DO11.10 Tg mice were transferred into BALB/c mice, and the mice were then immunized i.p. with 100 μg NP-OVA plus 10 μg LPS. NP-specific IgG₁ ASCs in spleen and BM were detected by enzyme-linked immunosorbent spot assay at days 14 and 28 after immunization (n = 7). (C) Splenic plasmablasts fail to home to the BM of CD69-deficient mice. Thy1.2⁻ splenocytes from BALB/c mice at day 14 after immunization with 100 μg NP-KLH in alum were transferred into KLH-immunized (14 d before) or unimmunized CD69^+/+ or CD69^−/− mice. One week after the cell transfer, NP-specific IgG ASCs in BM were detected by enzyme-linked immunosorbent spot assay (n = 4; *P < 0.05 and **P < 0.01).

Affinity maturation of antibodies in B cells occurs on antigen-bearing follicular dendritic cells of germinal centers (GCs) of the secondary lymphoid organs and is assisted by T follicular helper (T_FH) cells (24–27). To evaluate the generation of T_FH cells and GC formation in CD69-deficient mice, CXCR5⁺PD-1⁺ T_FH cells and GL-7⁺PNA^hiIgD^lo GC B cells were counted by flow cytometry, and gene expression of transcriptional repressor Bcl6, which is a critical regulator of T_FH cell differentiation, on CXCR5⁺PD-1⁺ T_FH cells were evaluated (Fig. S5). T_FH and GC B cells of the spleen were normally generated in CD69-deficient mice (Fig. S5).

To examine where and when CD69-deficient CD4 T cells exhibit the defects in the ability to help production of high-affinity antibodies, we enumerated antibody-secreting cells (ASCs) in spleen and BM of mice transferred with CD69-deficient or WT DO11.10 Tg CD4 T cells (Fig. 2B and Fig. S2A). At days 14 and 28 after immunization, CD69-deficient CD4 T cells could induce splenic ASCs but not BM ASCs, which include long-lived plasma cells (Fig. 2B). To analyze whether homing of plasmablasts to the BM indeed requires the help of BM memory CD4 T cells, splenic plasmablasts of WT mice, on day 14 after immunization of NP-KLH, were transferred into KLH-immunized or intact WT or CD69-deficient mice, and analyzed for the homing of NP-specific ASCs to the BM (Fig. 2C). In immunized hosts, the WT plasmablasts failed to efficiently home to the BM in the absence of CD69, suggesting that antigen-specific CD4 T cells of BM control the establishment of BM plasma cells (Fig. 2C). Interestingly, also in nonimmunized mice, CD69 significantly enhanced the efficient homing of plasma blasts to the BM (Fig. 2C), suggesting that BM CD4 T cells support the establishment of BM plasma cells independent of antigen specificity. Taken together, our results indicate that CD69-deficient CD4 T cells fail to help B cells in the late phase of an immune response because of the defective control of homing of long-lived plasma cell precursors, although they can normally form GCs and generate splenic plasmablasts.

BM Functional Th Cells Were Markedly Decreased in CD69-Deficient Mice.

To identify the functional defect of CD4 T cells in CD69-deficient mice, we first analyzed their proliferation, apoptosis, and ability to produce cytokines. Splenic naive CD4 T cells from CD69-deficient or WT mice were stimulated in vitro with anti-CD3 and anti-CD28. Clonal expansion (Fig. S6A) and production of the cytokines IFN-γ and IL-2 (Fig. S6B) in CD69-deficient and CD69-sufficient CD4 T cells were comparable. To assess the regulation of survival and apoptosis in CD4 T cells, effector CD4 T cells from CD69-deficient or WT mice were cultured with the survival factor IL-7. The rates of survival and apoptosis in WT and CD69-deficient effector CD4 T cells were indistinguishable (Fig. S6C). CD154, which is essential for help to B cells (28–30), was expressed equivalently in WT and CD69-deficient splenic naive, CD44^hi, and CXCR5⁺PD-1⁺ T_FH cells (Fig. 3A). BM CD44^hi CD4 T cells of CD69-deficient mice, however, did express less CD154 and also IFN-γ and TNF-α (Fig. 3 A and B), whereas they normally expressed the other immunoregulatory molecules such as inducible T-cell costimulator, PD-1, and OX40 (Fig. S6D). This corresponded to a drastic reduction of Ly-6C^hiCD44^hi CD4 T cells, which were absent as a distinct population (Fig. 3C), and functional CD154⁺ Ly-6C^hiCD44^hi CD4 T cells were decreased in the BM of CD69-deficient mice (Fig. 3D). Deficiency of CD69 itself does not affect the expression of Ly-6C, because the absolute numbers and the frequency of Ly-6C^hiCD44^hi CD8 T cells in the BM of CD69-deficient and WT mice were equivalent (Fig. S1A). The expression levels of CD49b, CD119 (IFN-γR1), CD127 (IL-7Rα), CD186 (CXCR6), and CD197 (CCR7) on BM CD44^hi CD69-deficient CD4 T cells were normal (Fig. S6E). As BM CD44^hi CD4 T cells are potent in giving help for antibody production of B cells (11), we examined the function of CD69-deficient BM CD44^hi CD4 T cells. BM CD44^hi CD4 T cells of CD69-deficient mice provide less help for the production of high-affinity antibodies in vivo (Fig. 3E). Thus, CD69-deficient mice have few functional memory Th cells with Ly-6C^hiCD44^hi phenotype.

Fig. 3.

Defects of functional memory Th cells in the BM of CD69-deficient mice. (A) Expression of CD154 is impaired selectively in CD4 T cells from BM of CD69^−/− mice. Naive (CD44^loCD62L⁺), splenic CD44^hi CD4 T cells, T_FH cells, and BM CD44^hi CD4 T cells were isolated and stimulated with immobilized anti-CD3 antibodies for 4 h in the presence of brefeldin A. T_FH cells were sorted as splenic CXCR5⁺PD-1⁺ CD4 T cells from C57BL/6 mice at day 7 after immunization with NP-CGG plus IFA. Cells were fixed and stained with anti-CD154 antibodies (n = 3–6). (B) Impaired cytokine production of BM CD4 T cells from CD69-deficient mice. BM CD44^hi CD4 T cells were isolated and stimulated with immobilized anti-CD3 antibodies for 4 h in the presence of brefeldin A. Cells were fixed and stained with antibodies against IFN-γ and TNF-α (n = 6). (C) Ly-6C^hiCD44^hi CD4 T cells are significantly reduced in the BM of CD69^−/− mice. The histograms show Ly-6C expression on CD44^hiCD62L⁻CD25⁻B220⁻ CD4 T cells of spleen and BM in CD69^+/+ and CD69^−/− mice. The isotype control is shown in gray. Right: Frequencies of Ly-6C^hi cells among CD44^hiCD62L⁻CD25⁻B220⁻ CD4 T cells of spleen and BM (n = 7). (D) CD69^+/+ Ly-6C^hiCD44^hi CD4 T cells of BM rapidly express CD154 after stimulation with anti-CD3 antibodies ex vivo. CD44^hiCD62L⁻B220⁻ CD4 T cells from BM of CD69^+/+ and CD69^−/− mice were isolated, stained with anti–Ly-6C antibodies, and then stimulated for 4 h with immobilized anti-CD3 antibodies in the presence of brefeldin A. Harvested cells were fixed and stained with anti-CD154 antibodies (n = 5). (E) Antigen-specific memory CD4 T cells from BM induce efficient production of high-affinity antibodies upon stimulation. A total of 10⁵ CD44^hiCD62⁻CD25⁻ CD4 T cells from BM of NP-OVA–immunized mice at day 25 were transferred together with 2 × 10⁶ splenocytes depleting Thy1.2⁺, CD11c⁺, CD11b⁺, Gr-1⁺, CD49b (DX5)⁺, and CD138⁺ cells by MACS (including approximately 92% of CD19⁺ cells) from the immunized mice into naive C.B-17/scid mice, followed by injection of NP-OVA. At day 7 after transfer and secondary immunization, the blood of the recipient mice was analyzed by ELISA for the presence of high-affinity antibodies (n = 3). All data (means ± SD) are representative of two or more independent experiments (*P < 0.05, **P < 0.01, and ***P < 0.001).

CD69 Regulates Relocation of Effector Th Cells to BM.

How does CD69 regulate the generation of memory Th cells in BM? In an immune response to OVA, antigen-experienced DO11.10 Tg CD4 T cells of the spleen and peripheral blood but not BM were normally present at days 4 and 48 after immunization (Fig. 1A and Fig. S3). The biased distribution indicated that CD69 works in the relocation of activated CD4 T cells from blood to BM. To analyze the migration ability of CD69-deficient CD4 T cells to the BM, CD4 T cells from spleen of WT or CD69-deficient DO11.10 Tg mice at day 4 after immunization were labeled with different fluorescent dyes and transferred into one normal mouse, and, 2 h later, the transferred cells in the spleen and BM were counted (Fig. 4A and Fig. S2B). The transferred CD69-deficient effector CD4 T cells could migrate to the spleen efficiently compared with WT, but not into the BM (Fig. 4A). The genetic loss of CD69 did not affect the differentiation of naive CD4 T cells into effector cells, nor did it indirectly affect cell trafficking, because the relocation of WT effector CD4 T cells into the BM was significantly inhibited by treatment with Fab fragments of anti-CD69 antibodies 2 h (Fig. 4B) or 24 h (Fig. S7) after cell transfer. Injection of CD69-specific antibodies into immunized mice also inhibited the relocation of antigen-specific effector CD4 T cells into the BM (Fig. 4C). In addition, anti-CD69–treated BM CD44^hi CD4 T cells did not rehome to the BM efficiently (Fig. 4D). These results indicate that CD69 regulates the homing of effector and memory Th cells to BM.

Fig. 4.

CD69 regulates the relocation of Th cells to the BM. (A) Relocation of CD69^−/− antigen-specific effector CD4 T cells to BM is impaired. A total of 1 × 10⁷ splenic CD4 T cells from CD69^+/+ (open column) and CD69^−/− (filled column) DO11.10 Tg mice 4 d after immunization with 100 μg OVA plus 10 μg LPS were labeled with CMFDA or CMTMR, respectively, mixed, and transferred into C.B-17/scid mice. Two hours later, CMFDA⁺ or CMTMR⁺ OVA-TCR⁺CD44^hi CD4 T cells in spleen and BM of the host mice were assessed by flow cytometry (n = 5). (B) Relocation of antigen-specific CD4 T cells to the BM can be blocked with the Fab fragment of anti-CD69 antibody. Splenic CD4 T cells from DO11.10 Tg mice, 4 d after immunization with 100 μg OVA plus 10 μg LPS, were pretreated with the Fab fragment of anti-CD69 antibody (filled column) or control (open column) before labeling with CMFDA or CMTMR. These cells were mixed and transferred into C.B-17/scid mice. Two hours later, CMFDA⁺ or CMTMR⁺ OVA-TCR⁺CD44^hi CD4 T cells in spleen and BM of the host mice were assessed by flow cytometry (n = 6). (C) Relocation of antigen-specific memory Th cells to the BM is dependent on CD69. CD4 T cells (3 × 10⁵) from DO11.10 Tg mice were transferred into BALB/c mice, and the mice were then injected with 100 μg of OVA plus 10 μg LPS. At days 4, 6, and 8 after immunization, mice were injected with 100 μg of Fab fragments of anti-CD69 antibodies (filled column) or isotype control (open column), and analyzed for OVA-TCR⁺CD44^hi CD4 T cells in spleen and BM at day 12 after immunization. (D) CD69-dependent rebound of BM-derived CD44^hi CD4 T cells to the BM. A total of 2 × 10⁶ B220⁻MHC class II⁻ cells from BM of BALB/c mice were treated with the Fab fragment of anti-CD69 antibodies (filled column) or isotype control (open column), labeled, transferred, and quantified as shown in B (n = 5). All data (mean ± SD) are representative of two or more independent experiments (*P < 0.05 and **P < 0.01).

CD69 Regulates Persistence of Th Cells on BM Stromal Cells.

Although anti-CD69–treated effector cells could not efficiently migrate to the BM (Fig. 4B), the few BM-resident cells of mice transferred with anti-CD69–treated cells could not also efficiently contact laminin⁺ cells, i.e., stromal cells and endothelial cells (Fig. 5A). Thus, the absolute numbers of laminin⁺-interacting effector cells in the BM were markedly reduced by inhibition of CD69 (Fig. 5B). These data indicate that memory Th cells persist on BM stromal cells by using CD69. To confirm the residence of memory Th cells on the stroma, BM Ly-6C^hi or CD69⁺ CD44^hi CD4 T cells in steady status could also be analyzed histologically. More than 80% of CD69⁺CD4⁺ cells in the BM expressed Ly-6C highly as detected by flow cytometry (Fig. 5C). Therefore, we generated CD69:GFP knock-in mice (Fig. S8), and investigated the localization of BM CD69⁺CD4⁺ cells, which include most memory Th cells in vivo. More than 80% of GFP⁺CD4⁺ cells compared with approximately 30% of GFP⁻CD4⁺ cells contacted laminin⁺ stromal cells (Fig. 5D). Taken together, these data indicate that CD69 regulates the persistence of memory Th cells on survival niches organized by stromal cells.

Fig. 5.

CD69 regulates the persistence of Th cells on BM stromal cells. (A) Anti-CD69–treated antigen-specific effector CD4 T cells of BM do not contact laminin⁺ stromal cells. As described in Fig. 4B, antigen-specific effector CD4 T cells were treated with the Fab fragment of anti-CD69 antibodies or isotype control, labeled with CMFDA or CMTMR, respectively, and transferred into C.B-17/scid mice. Representative histological section of CMFDA (green), CMTMR (red), and laminin (gray) is shown (Left). A percentage of CMFDA⁺ or CMTMR⁺ cells contacting laminin⁺ cells in the BM of the host mice were enumerated by immunohistological analysis (Right). (B) CD69-blocked effector Th cells significantly failed to persist on the stromal cells. Absolute numbers of CMFDA⁺ or CMTMR⁺ OVA-TCR⁺CD44^hi CD4 T cells contacting laminin⁺ cells in the BM were calculated from data by flow cytometry in Fig. 4B and immunohistological analysis in Fig. 5A (n = 3). (C) CD69⁺CD4⁺ cells in BM expressed Ly-6C highly. A representative CD69/CD4 profile of BM cells and a histogram of Ly-6C among CD69⁺CD4⁺ cells are shown. The isotype control is shown in gray (n = 3). (D) CD69-expressing memory Th cells contact with laminin⁺ cells. Representative histological section of GFP (green), CD4 (red), and laminin (gray) labeling in BM of CD69^gfp/+ mice is shown (Upper). A percentage of GFP⁺CD4⁺ or GFP⁻CD4⁺ cells contacting laminin⁺ cells in the BM of CD69^gfp/+ mice were enumerated by immunohistological analysis (Lower). Thirty-six of 43 GFP⁺CD4⁺ cells were in contact with laminin⁺ cells, compared with 24 in 82 GFP⁻CD4⁺ cells. All data (mean ± SD) are representative of two or more independent experiments (**P < 0.01).

Discussion

We herein demonstrate that CD69 functions as a homing receptor on CD4 T cells and is required for relocation to, and persistence in, the BM, and that its function is essential for the generation of memory Th cells. In fact, CD69-deficient mice have few professional resting memory Ly-6C^hiCD44^hi, functional CD154⁺, and also antigen-specific memory Th cells in the BM. We have previously reported that memory Th cells are maintained on their stromal niches in the BM (31) and now find that antigen-specific and CD69(GFP)⁺ CD4 T cells of the BM contact stromal cells in a CD69-dependent fashion. Moreover, we show a role for memory Th cells in the BM: the defect of BM memory Th cells impairs the production of high-affinity antibodies because of the defective generation of long-lived plasma cells in the BM.

The molecular mechanisms regulating Th cell trafficking to the BM were unclear. CD69 is required for the trafficking of effector CD4 T cells to the BM, especially for their relocation and persistence on the BM stromal cells. Although most splenic antigen-specific CD4 T cells express CD69 at days 1 and 2 after immunization, 80% to 90% have lost the expression by day 4 (Fig. S9). This suggests that splenic CD69-expressing effector CD4 T cells detected at day 4 may preferentially migrate to the BM and transit to memory cells in the BM. The S1P/S1P₁ pathway is another regulator of cell trafficking. CD69 is physically associated with S1P₁ and directly inhibits S1P/S1P₁-mediated cell egress from lymphoid organs after exposure to IFN-α/β (23). Both CD69-overexpressing and S1P₁-deficient thymocytes accumulate in the thymus and do not enter the periphery (17, 32). From these studies, CD69 is presumed to inhibit the egress of mature thymocytes through blockage of S1P₁. However, thymocyte development and the generation of peripheral CD4 and CD8 T cells were normal in CD69-deficient mice (21). Also in an immune response, antigen-experienced CD69-deficient splenic CD4 T cells normally egress from the spleen to the peripheral blood (20) (Fig. S3). Therefore, the expression of CD69 itself on thymocytes and splenocytes may not be physiologically involved in their egress from thymus and spleen, respectively. It had been reported that the egress of CD4 T cells and B cells from BM to the blood is dependent on the S1P/S1P₁ pathway (33, 34). The constitutive expression of CD69 on resting memory Th cells may inhibit their egress from the BM.

CD69-deficient effector CD4 T cells have a defect in the ability to enter into the BM (Fig. 4A). In addition, most CD69⁺CD4⁺ memory Th cells of the BM contact stromal cells in vivo, whereas most CD69⁻CD4⁺ cells do not (Fig. 5D). The treatment of effector or memory cells with Fab fragments of anti-CD69 antibodies markedly inhibited not only relocation to the BM but also the interaction with BM stromal cells after adoptive transfer (Figs. 4B and 5 A and B). These data suggest that CD69 of Th cells works as an adhesion molecule in a two-step process: in the transmigration via BM sinusoidal endothelial cells and in the adhesion to the stromal niches of the BM. Although the existence of a ligand for CD69 is expected based on crystal structure analysis (35), it is still unknown but tempting to speculate that the ligand is expressed on BM stromal cells. Some of the C-type lectin family molecules, e.g., dendritic cell-specific ICAM-3 grabbing nonintegrin and macrophage mannose receptor act as adhesion molecules that regulate lymphocyte functions and are involved in establishing adaptive immunity (36). Thus, CD69 may also function as an adhesion molecule to BM stromal cells as well as other members of the C-type lectin family.

Our previous data indicates that Ly-6C^hiCD44^hi CD4 T cells in the BM reflect memory Th cells which are reactive, resting, and long-lived. CD69-deficient CD4 T cells are normally developed, proliferate, and differentiate into effector cells. Although the numbers of CD44^hi memory-phenotype CD4 T cells in spleen, lymph nodes, BM, and blood of CD69-deficient mice are also normal, only BM Ly-6C^hiCD44^hi CD4 T cells are selectively lacking (Fig. 3 C and D). In an immune response to OVA, splenic OVA-specific CD69-deficient CD4 T cells at day 48 after immunization were detectable and similar in number to WT cells (Fig. 1 A and B). The selective loss of BM memory Th cells in CD69-deficient mice provides the opportunity to analyze BM memory cells for their unique roles in humoral immunity. The lack of CD69 in CD4 T cells induced the defective production of high-affinity antibodies (Fig. 2A and Fig. S4A). To clarify the rationale of the defect, GC B cells and splenic and BM plasma cells were enumerated, and CD154 expression and cytokine production, which are essential for help to B cells, were measured in T_FH cells and splenic and BM CD44^hi CD4 T cells. Splenic B/plasma and T cells were numerically and functionally normal. In contrast, BM plasma cells and T-cell numbers were decreased. These data indicate that the defect of BM plasma cells was caused by the absence of BM antigen-specific Th cells. BM Th cells may play a unique and essential role in the transmigration of plasma cells to the BM and/or their persistence in the BM.

Injection of Fab fragments of anti-CD69 antibodies blocked the relocation of effector CD4 T cells to the BM (Fig. 4C). This may be used as a possible therapeutic for prevention of the persistence of harmful memory cells, i.e., those responsible for allergy and autoimmune diseases. To date the precise roles of memory Th cells in acquired immunity have remained unclear. However, this study and the use of Th memory-deficient mice will contribute to the clarification of their roles in the primary and secondary immune responses.

Materials and Methods

Detailed descriptions of all materials and methods are provided in SI Materials and Methods.

Mice.

CD69-deficient mice were backcrossed more than 13 times onto C57BL/6 or BALB/c backgrounds (21). DO11.10 Tg mice were provided by D. Loh (Washington University School of Medicine, St. Louis, MO) (37). For all experiments, mice were used at 6 to 16 wk of age and were maintained under specific pathogen-free conditions. BALB/c, C57BL/6, and C.B-17/scid mice were purchased from Clea. CD69:GFP knock-in mice (CD69^gfp/+) were generated by homologous recombination in ES cells. Transfection of ES cells and selection of clones were performed essentially as described for CD69 KO mice (21). To introduce cDNA-encoding egfp into the cd69 locus, we generated a replacement vector to remove the first exon of the cd69 gene encompassing the initiation codon (Fig. S8A). CD69^gfp/+ mice were backcrossed 11 times to the C57BL/6 background. All animal experiments were approved by the Chiba University Review Board for Animal Care. For immunization, mice were injected with OVA (Sigma), NP₂₉-KLH, NP₂₉-OVA, or NP₃₆-CGG (Biosearch Technologies) with LPS (Invivogen), alum (Imject Alum; Pierce), or IFA (Sigma).

Cell Labeling and Adoptive Transfer.

For adoptive transfer, CD4 T cells from BALB/c or DO11.10 Tg mice were sorted by magnetic-activated cell sorting (MACS) and transferred i.v. into BALB/c or C.B-17/scid mice. For positive selection and neutralization by antibodies, we used the Fab fragment of anti-CD4 or anti-CD69 antibodies and streptavidin-MACS microbeads (Miltenyi Biotec). For induction of OVA-TCR⁺ T_FH cells, mice were immunized i.p. with 100 μg NP₂₉-OVA plus LPS after adoptively transferred CD4 T cells from DO11.10 Tg mice. OVA-TCR⁺ cells were phenotyped by staining with antibodies against PD-1 and CXCR5. To monitor donor cells in host mice, cells were labeled with the cytoplasmic probes CellTracker Green 5-chloromethylfluorescein diacetate (CMFDA) and CellTracker Orange (5-(and-6)-(((4-chloromethyl)benzoyl)amino)tetramethylrhodamine) (CMTMR; Invitrogen) before transfer. Briefly, cells (1 × 10⁷ cells/mL) were incubated with 0.1 μM of CMFDA or 5 μM of CMTMR in PBS solution for 15 min at 37 °C, washed, and incubated for another 30 min at 37 °C, according to the manufacturer's instruction. Flow cytometric data were analyzed with FlowJo software (Tree Star).