Role of CD44 in the differentiation of Th1 and Th2 cells: CD44-deficiency enhances the development of Th2 effectors in response to SRBC and chicken ovalbumin

Abstract

CD4 T cells can be primarily polarized to differentiate into Th1 or Th2 cells. CD44 is a marker of T cell activation and a property of long-lived memory cells and implicated in cell migration, activation and differentiation. To date, whether CD44 has a role in regulating Th1-Th2 differentiation has not been determined. In this study, we compared Th1 and Th2 responses in wild-type and CD44-deficient mice in response to SRBC and chicken ovalbumin, as well as examined Th1-Th2 differentiation in vivo and in vitro from CD44-sufficient and CD44-deficient naïve CD4 T cells. We observed that deficiency of CD44 tended to inhibit Th1 while promoting Th2 differentiation. Furthermore, chimeric studies suggested that CD44 expression by CD4 T cells was essential for such Th2 bias. The regulation by CD44 occurred at the transcription level leading to up-regulated GATA3 and down-regulated T-bet expression in activated CD4 T cells. We also noted that CD44-deficiency could modify the state of dendritic cell subsets to induce a Th2-biased development. Results presented here demonstrate for the first time that CD44 participates in the regulation of Th1-Th2 differentiation.

Introduction

CD4 helper T cells can differentiate into functionally distinct effector subsets with different cytokine expression profile and immune regulatory function based on the antigen receptor- and cytokine-mediated signals. This family is expanding. From original Th1 and Th2 lineage, it now includes Th17 and transforming growth factor-induced regulatory T cells (iTreg) as well as follicular helper T cells (Tfh), and possibly, Th9 (1-5). As an original sole paradigm, the signals that drive Th1-Th2 differentiation are clearly elucidated. Th1- or Th2-polarizing cytokines initiate signaling via JAK/STAT complexes and there are clear differences noted in these subsets. For example, STAT1, STAT4 and T-bet facilitate Th1 signaling while STAT6 and GATA3 promote Th2 differentiation. Th1 subset produces interferon-γ (IFNγ); Th2 subset produces IL-4, IL-5 and IL-13. Th1 response is often accompanied by the production of IgG2a while Th2 response by the production of IgG1 and IgE antibodies. Also, Th1 and Th2 responses are often mutually exclusive. Both Th1-and Th2-specific cytokines can facilitate growth or differentiation of their own respective T-cell subset, but additionally might inhibit the development of the opposing subset. At the transcription level, GATA3 promotes Th2 responses through three different mechanisms: induction of Th2 cytokine production, selective growth of Th2 cells and inhibition of Th1 cell-specific factor (6).

Polarization of naïve CD4 T cells towards Th1 or Th2 mainly relies on the cytokine milieu, but it is also influenced by many other factors, such as antigen affinity for the T cell receptor, concentration of antigen, and co-stimulatory molecules (6). The sources of the cells that produce the cytokines in vivo and the molecules that are involved in the regulation of Th development are still under investigation. Although Th1 and Th2 can themselves provide IFN-γ or IL-4 for the recruitment of Th1 or Th2 differentiation, respectively, the cells that initiate effector T cell differentiation in primary versus secondary responses have not yet definitively been determined (5). Many cell types may be involved in these processes, such as dendritic cells (DCs), natural killer (NK), NKT cells (5, 7) or macrophages (Mø) (8). There is not an exclusive conclusion so far. It is especially complicated and controversial to evaluate the relationship between DC subsets and Th development; even though the adoptive transfer of antigen-pulsed CD8α⁻ DCs induces a Th2 response, transfer of CD8α⁺ DCs leads to a Th1 differentiation (9, 10).

CD44 is a widely distributed cell surface glycoprotein expressed by a variety of lymphoid and non-lymphoid cells. The CD44 family consists of a standard form and a group of isoforms resulting from the extensive alternative splicing that might attribute to its sophisticated implication in the immune responses and immune regulation. CD44 molecule participates in cell adhesion and migration, lymphocyte homing, activation and proliferation, lytic activity of T cells and NK cells, and tumor metastasis (11-14). CD44 is coupled to at least two tyrosine kinases, p185^HER2 and c-Src kinase, and has broad functions in cellular signaling cascades not only by establishing specific transmembrane complexes but also by organizing signaling cascades through association with its partner proteins, which monitor changes in the extracellular matrix that influence cell growth, survival and differentiation as well as induction of the cytotoxicity of CTL and NK cells (15-17). CD44 is recruited to the immunological synapse during DC and T cell interactions and affects the subsequent T cell activation, IL-2 and IFN-γ production, phosphotyrosine and protein kinase C-θ enrichment at the synapse (18). CD44 splice variant expression is obligatory for the migration and function of Langerhan cells (LCs) and DCs. Blockade of CD44 inhibits the emigration of LCs from the epidermis, prevents binding of activated LCs and DCs to the T cell zones of lymph nodes, and severely inhibits their capacity to induce a delayed type hypersensitivity (DTH) (19).

Th1 and Th2 lymphocytes express CD44 and use CD44 for their rolling on and adhesion to the intestinal endothelium (20, 21). However, whether CD44 regulates Th1-Th2 differentiation is not clear. In this study, we analyzed Th1- and Th2- responses in wild-type and CD44-deficient mice to particulate and soluble antigens such as sheep red blood cells (SRBC) and chicken ovalbumin (OVA), respectively. Here, we provide significant evidence that CD44 plays a role in regulation of Th1-Th2 differentiation and CD44-deficiency triggers a Th2-biased Th development.

Materials and Methods

Mice and reagents

Wild-type C57BL/6 and CD4 T cell-deficient (CD4^−/−) mice were purchased from Jackson Laboratory. OT II mice were purchased from Taconic. CD44 knockout (CD44^−/−) mice were generated at Amgen Institute (Toronto, Canada) and kindly provided to us by Dr. Tak Mak. These mice are on C57BL/6 background and have been extensively characterized in our previous studies (22, 23). Mice were housed in the University of South Carolina Animal Facility. Animal procedures were performed according to NIH guidelines under protocols approved by the Institute Animal Care and Use Committee of the University of South Carolina. Sheep red blood cells (SRBC), and NycoPrep were purchased from Cedarlane Laboratories. Thioglycollate, OVA, and complete Freund's adjuvants (CFA) were purchased from Sigma. Incomplete Freund's adjuvant (IFA) was purchased from DIFCO.

SRBC-induced DTH reaction

Groups of six mice each were injected subcutaneously (s.c.) with 1×10⁸ SRBC/mouse in the abdomen. Fourteen days later, the sensitized mice were challenged s.c. with 2 × 10⁶ SRBC/mouse in a volume of 20 μl in left ear. As a control, 20 μl of PBS was applied in right ear of the same mouse. The swelling of ear was measured at 0, 24, 48 and 72 h of the challenge. The magnitude of ear thickness is reported as mean thickness ± S.E. in each group of six mice.

Hemagglutination assay

Groups of four mice each were immunized intraperitoneally (i.p.) or intravenously (i.v.) with 1×10⁸ SRBC. Blood was collected and serum prepared from each mouse on d5 post immunization. Anti-SRBC antibody was detected by hemagglutination assay. To 20 μl of two-fold diluted serum samples in U-shape micro-titration plate, 20 μl of 2% SRBC suspension was added. The plate was incubated at 37°C for 1 h and at 4°C overnight. The hemagglutination was recorded. The highest dilution causing hemagglutination was considered as antibody titer and calculated as the mean log2. Sera prepared the day before immunization were used as negative control.

IgM, IgG1 and IgG2a antibody detection

Groups of four to five mice each were immunized i.p. with 4 μg/ per g body weight of OVA emulsified in CFA. After two weeks, mice received the secondary immunization with the same amount of OVA but emulsified in IFA. Sera were collected on d10 of the primary immunization or d7 of the secondary immunization.

Anti-OVA IgM, IgG1 and IgG2a antibodies in the sera were measured by ELISA (enzyme-linked immunosorbent assay). Briefly, ELISA plate was coated with 100 μl/well of OVA (10 μg/ml in 0.1 M carbonate buffer, pH 9.6) overnight at 4°C and blocked with Assay Diluent for 2 h at room temperature (RT). Serum samples diluted at least 50-fold with Assay Diluent were added and incubated for 2 h at 37°C. Biotin-conjugated antibodies against IgM, IgG1, or IgG2a were added and incubated for 1 h at 37°C. Next, Avidin-HRP was added and incubated for 30 min at RT. TMB substrate was added, developed, and the reaction was stopped by 2N H₂SO₄. Absorbance at 450 nm was read with an ELISA reader (Victor² 1420, PerkinElmer). Purified mouse IgM, IgG1 or IgG2a were used to formulate each standard curve. All reagents used in ELISA assay were from BD Pharmingen except for those mentioned elsewhere.

Multiplexed microsphere cytokine immunoassay (Bio-Plex assay)

Serum cytokines were measured by multiplexed microsphere cytokine immunoassay using Bio-Plex Cytokine Assay kit from Bio-Rad. The kit measures 23 cytokines including Th1-Th2 panel. The assay was performed as per the manufacture's instruction.

Enrichment and isolation of spleen DCs

Spleen DCs were enriched or purified as described in our previous report (24). Briefly, spleens were chopped, and then digested for 30 min at 37°C with collagenase and DNase. After filtering with a 70-μm pore size cell strainer, cells were subjected to a gradient centrifugation in Nycodenz medium (NycoPrep). The interface was collected and washed with PBS. The resulting cells were incubated with anti-CD8α-FITC, anti-CD11c-PE, anti-PDCA-1-Alexa Fluor 647, and anti-CD11b-PE-Cy7 for 30 min at 4°C (all antibodies were from eBioscience). The staining was analyzed by flow cytometry. For cell culture, DCs were further purified by positive magnetic selection of CD11c⁺ cells. The purity of isolated CD11c⁺ DCs was 95–97% as determined by flow cytometric analysis.

Th1-Th2 polarization in vitro

Naïve CD4 T cells were purified from spleens by negative selection, with minor modifications as described (25), with PE-conjugated mAb against CD8, CD19 or B220, I-A^b, CD16/32, CD24, NK1.1, and Gr-1 (all antibodies from eBioscience) using EasySep PE Selection Kit (StemCell Technology) following the manufacturer's instruction.

CD4 T cell polarization was induced by polyclonal stimulation or antigen stimulation. In the polyclonal stimulation, CD4 T cells (>95%) were primed with anti-CD3 + anti-CD28 (3μg/ml each, eBioscience) and irradiated (3000 rads) splenocytes (10×) from CD44^+/+ mice that were depleted of CD3 T cells by magnetic beads (StemCell Technology). For the antigen stimulation, CD4 T cells isolated from OT II mice were primed with OVA_323-339 peptide (40ug/ml, Sigma) and DCs (0.2×). For Th1 priming, IL-12 (10 ng/ml, R&D Systems) and anti-IL-4 Abs (10μg/ml, eBioscience) were added; for Th2 priming, IL-4 (2 ng/ml, R&D Systems), anti-IL-12 Abs (10 μg/ml, eBioscience), and anti-IFNγ Abs (10 μg/ml, eBioscience) were added. All cultures contained IL-2 (100 units/ml, NCI). On d4, live cells were harvested over NycoPrep medium and used for intracellular cytokine staining.

On d4, the cultures were supplemented with 50ng/ml PMA (Sigma), 1ug/ml ionomycin (Sigma), and 2 μM Monensin (BD Pharmingen) and incubated for another 4 h at 37°C. For the staining, cells were fixed, permeabilized, and stained intracellularly with APC-anti-IL-4 and FITC-anti-IFNγ Abs (eBioscience) using Cytofix/Cytoperm intracellular staining Kit (BD Pharmingen). The expression of IL-4 and IFNγ was examined by flow cytometry.

Intracellular staining of T-bet and GATA3

Splenocytes were prepared on d5 of SRBC immunization or d10 of primary OVA immunization. T-bet and GATA3 expression were examined by intracellular staining using FoxP3 staining buffer kit (eBioscience). Cells were re-stimulated with PMA and ionomycin in the presence of Monensin for 4 h as described above. Cells were first stained extracellularly with FITC-anti-CD4 Abs, then fixed and permeabilized and subjected to the intracellular staining with Alexa Fluor 647-anti-T-bet and PE-anti-GATA3 Abs (eBioscience). The antibody incubation was carried out for 30 min at 4°C. The staining was analyzed by flow cytometry.

IL-12 secretion from splenic DCs

DCs (2 ×10⁵) per well were cultured in a 96-well flat-bottomed plates in 200μl RPMI 1640 containing LPS (1μg/ml) or sonicated B16F10 tumor lysate (at indicated concentrations). Culture supernatants were collected at 24 h. The IL-12 production was examined by ELISA assay using OptEIA Mouse IL-12 Detection Kit (BD Pharmingen) following the manufacturer's instruction.

Development of Chimeras

To generate bone marrow (BM) chimeras, CD4^−/− mice were lethally irradiated (950 rads from a ¹³⁷Cs source) and reconstituted with a total of 10 × 10⁶ BM cells from appropriate donor mice (described in the Results section). Mice were allowed to reconstitute for at least 6 weeks prior to the immunization.

Statistical analysis

The differences between experimental groups were analyzed using the Student t test with p < 0.05 being considered statistically significant.

Results

SRBC-induced DTH and antibody responses are decreased in CD44^−/− mice

To measure SRBC-triggered Th1 responses, we first induced DTH reaction that is considered to be Th1-driven (26, 27). The SRBC-sensitized mice were challenged in the ear and the ear-swelling reflecting the activated T cell reaction was measured. Mice challenged with PBS served as negative controls and showed a small, non-specific increase in ear thickness. As shown in Figure 1A, the magnitude of DTH, as determined by ear swelling, in CD44^−/− mice was dramatically reduced compared to the CD44^+/+ mice. However, the time course of DTH, namely onset and decrease in ear swelling, remained the same in both groups of mice (Figure 1A).

Figure 1. Decreased response against SRBC as measured by DTH reaction and anti-SRBC antibody production in CD44^−/− mice.

A. Mice were sensitized with SRBC or PBS (control) and challenged on d14. DTH responses were evaluated at different time points after SRBC challenge. Each data point represents the mean ± SE from a group of six mice. B and C. Anti-SRBC responses were measured in mice immunized with SRBC either i.p (B) or i.v (C) 5 d after immunization by examining hemagglutination.

We next measured SRBC-induced antibody production by hemagglutination assay. On d5 of the primary immunization, we found that the SRBC-specific antibodies were produced in both CD44^−/− and CD44^+/+ mice, but CD44^−/− mice showed a significantly decreased level in comparison to CD44^+/+ mice (Figure 1B and C). Immunization of i.p. (Figure 1B) or i.v. (Figure 1C) produced similar results. We also measured anti-SRBC Ig level on d8 of the primary immunization, which also showed a decreased response in CD44^−/− mice when compared to CD44^+/+ mice (data not shown).

OVA- induced antibody in CD44^−/− mice: enhanced IgG1 and decreased IgG2a

To measure OVA-triggered response, we immunized mice with OVA and measured OVA-specific IgM on d10 of the primary immunization and OVA-specific IgG1 and IgG2a on d7 of the secondary immunization. As shown in Fig 2, all three Ig subtypes could be detected in both CD44^−/− and CD44^+/+ mice. Interestingly, however, IgM and IgG1 responses were dramatically increased while IgG2a response was significantly decreased in CD44^−/− mice in comparison to CD44^+/+ mice (Figure 2).

Figure 2. Evaluation of anti-OVA antibody levels in CD44^−/− mice.

Mice were immunized with OVA. Sera were collected on d10 after primary immunization or d7 following secondary immunization. OVA-specific IgM, IgG2a, and IgG1 were assessed by ELISA. Panel A: IgM; Panel B: IgG1; Panel C: IgG2a.

Serum cytokine profile: Predominant Th2-cytokines in CD44^−/− mice

Serum cytokine productions were detected following SRBC- and OVA-immunization by the Bio-Plex assay. Immunizations were performed as described above. For SRBC-induced cytokines, mice were immunized either by i.p. or i.v route and sera were prepared on d5 of the primary immunization. For OVA-induced cytokines, sera were prepared on d10 of the primary immunization or d7 of the secondary immunization. As shown in Fig 3, IL-4, IL-5, IL-12, IL-13, IL-17 and IFN-γ were produced in both CD44^−/− and CD44^+/+ mice. The level of IL-4, IL-12 and IL-17 were much lower following SRBC immunization when compared to OVA immunization. Furthermore, Th1 cytokines, IFN-γ and IL-12, were down-regulated in CD44^−/− mice either with SRBC or OVA immunization. In contrast, Th2 cytokines, IL-4, IL-5 and IL-13 were up-regulated in CD44^−/− mice. Overall, these results suggested a Th2-predominant cytokine phenotype in CD44^−/− mice (Figure 3).

Figure 3. Serum cytokines production.

Groups of four to five mice were immunized with SRBC or OVA and serum cytokines in immunized mice were screened by Bio-plex assay. SRBCiv: i.v. immunization with SRBC; SRBCip: i.p. immunization with SRBC; OVAprimary: primary immunization with OVA; OVAboost: secondary immunization with OVA. * p < 0.01.

Up-regulated GATA3 and down-regulated T-bet in CD44^−/− mice

The above results clearly demonstrated that Th1 immune response was down-regulated and Th2 immune response was up-regulated in CD44^−/− mice. To determine whether this phenotype could occur at the transcriptional level, we examined the expression of two transcription factors, T-bet and GATA3. Splenocytes were isolated on d5 of the primary SRBC immunization and d10 of the primary OVA immunization. After re-stimulation with PMA and ionomycin, the expression of T-bet or GATA3 was evaluated in CD4⁺ T cells by intracellular staining. Figure 4 shows cells that were gated for CD4+ T cells and analyzed for T-bet and GATA3. Our results showed that percentage of CD4⁺T-bet⁺ cells was decreased whereas the percentage of CD4⁺GATA3⁺ cells was increased in CD44^−/− mice when compared to CD44^+/+ mice using both SRBC and OVA immunization protocols (Figure 4). These results suggested that CD44 regulation may occur at the transcriptional level.

Figure 4. Enhanced GATA3 and decreased T-bet expression in CD4+T cells from CD44^−/− mice.

Splenocytes were isolated on d5 of SRBC immunization or d10 of primary OVA immunization. The cells were triple stained for CD4, T-bet and GATA3. The cells gated for CD4 were analyzed using flow cytometry for the expression of T-bet and GATA3. Panel A: Shows a representative experiment using isotype controls for Alexa Fluor 647 and PE and gating for CD4+ T cells; Panel B: CD4⁺ T cells analyzed for T-bet and GATA3 with PBS treatment as a negative control; Panel C: CD4⁺ T cells analyzed for T-bet and GATA3 following OVA immunization; Panel D: CD4⁺ T cells analyzed for T-bet and GATA3 after SRBC immunization.

CD44 expression on CD4 T cells affects Th development

To further verify that the CD44 molecule expression on CD4 T cells regulates Th1-Th2 polarization, we investigated the SRBC-triggered response using bone marrow (BM) chimeras. The chimeras were created as described by others (28). CD4^−/− mice were used as recipients that lack CD4 T cells but are CD44-sufficient. The chimeras were made by re-constituting the irradiated recipients with mixture of BM that consists of 3 parts of BM from recipient⁻mice and 1 part BM from CD44^−/− mice. Hence, the chimeras contained CD4 T cells from CD44^−/− BM and would be CD44 deficient (marked as CD44−/− CD4 in Fig 5), while the other hematopoietic cell types would be derived from the recipient BM itself and therefore CD44 sufficient (28). Another group of chimeras was prepared by using 3 parts of BM from recipient mice and 1 part BM from WT mice, in which all hematopoietic cells, including CD4 T cells, were CD44^+/+ (marked as CD44+/+CD4 in Fig 5). The phenotype of these chimeras was confirmed by flow cytometry. For this, the cells were gated for CD3, CD11c or B220 and such gated cells were analyzed for combined expression of CD4 and CD44, CD8 and CD44, CD11c and CD44 or B220 and CD44 As seen from Figure 5A, chimeras designated CD44+/+ CD4 had ∼20.2% CD4+CD44+ whereas the CD44−/− CD4 chimeras had only 1.6% CD4+CD44+T cells. Thus, almost all of the CD4+ T cells in CD44−/− CD4 chimeras were CD44-deficient. The phenotyping also showed that the generation of chimeras did not influence the expression of CD44 in other cell types; the proportion of CD44 cells that also expressed CD8, B220 or CD11c were almost identical in the two types of chimeras (Figure 5A). Next, the chimeras, CD4^−/− and WT mice were immunized i.p. with SRBC. The anti-SRBC antibody and serum cytokines were examined on d5 of immunization by hemagglutination assay and Bio-Plex assay as described in this text. As expected, we found that the chimeras with CD44^−/−CD4 T cell reconstitution showed minimal level of antibody production which was similar to the level seen in CD4^−/− mice. Dramatically, the chimeras with CD44^+/+CD4 T cell reconstitution exhibited full restoration of the antibody production that reached similar levels as that seen in the WT mice (Figure 5B). Again, the serum cytokine profile correlated with these data (Figure 5C). Together, these results indicated that CD44 expression on CD4 T cells contributes to the Th1-Th2 development.

Figure 5. Anti-SRBC antibody and cytokines production in chimeras.

Panel A: phenotype of chimeras. On 8^th wk of the reconstitution, splenocytes from chimeras were triple-stained with CD44PE, CD4-FITC or CD8-FITC, and CD3 PE-Cy7. To study CD11c+ cells, they were enriched from the spleen and double-stained with CD44-PE and CD11c-FITC. To analyze B220+ cells, splenocytes were double-stained with CD44-PE and B220-FITC. Panel B: Mixed bone marrow chimeras in which the CD4 T cells were either CD44 sufficient or deficient were generated as described in Materials and Methods. Reconstituted mice were i.p. immunized with SRBC and anti-SRBC antibody level was measured d5 post immunization by hemagglutination assay. CD4^−/− and CD44 wild-type mice (WT) were included as controls. The data represent mean ± SE from groups of four mice. Panel C: measurement of serum cytokines in the same groups of mice. The differences in each of the cytokine production between the two groups was statistically significant (p<0.01 for each cytokine).

CD44^−/− CD4 T cells polarize towards Th2 in vitro

The in vivo Th1 to Th2 shift in immune response seen in CD44^−/− mice suggested that CD44 modulates CD4 T cell differentiation. To address this further, we performed in vitro CD4 T cell polarization experiment. CD4 T cells from CD44^−/− or CD44^+/+ mice were stimulated with anti-CD3 and anti-CD28 under Th1- or Th2-polarizing conditions. Intracellular expression of IL-4 and IFN-γ were determined on d4. As shown in Fig 6, when compared to CD44^+/+ CD4 T cells, CD44^−/− CD4 T cells showed decreased levels of polarization to Th1; whereas they readily differentiated into Th2 cells (Fig 6).

Figure 6. Polyclonal stimulation leading to Th1-Th2 polarization in vitro.

CD62L^highCD4 T cells from CD44+/+ and CD44−/− mice and T cell-depleted splenocytes from CD44^+/+ mice were co-cultured in Th1- or Th2- polarization condition with addition of anti-CD3 and anti-CD28 antibodies as described in Materials and Methods. On d4 of the culture, cells were re-stimulated with PMA and ionomycin for 4 h. IFNγ and IL-4 production were then evaluated by intracellular staining. The staining was analyzed by flow cytometry. The percentage represents positive cells gated on live cells.

Involvement of DCs

While our studies demonstrated that CD44 expression on CD4 T cells played a role in Th1-Th2 differentiation, we wondered if CD44-deficiency on DCs could also impact the Th1-Th2 polarization. Recent studies from our laboratory demonstrated that CD44 expression on DC plays a critical role in T cell activation. Specifically, we noted that deficiency of CD44 on DC affected the functional immune synapse, resulting in decreased phosphotyrosine and protein kinase C-θ enrichment at the synapse (18). To test this hypothesis further, we examined DC functions related to Th differentiation. First, we were interested to know if CD44-deficiency would induce a DC phenotype that favors Th2 differentiation. To this end, mice were immunized with SRBC or OVA and splenocytes were enriched for DCs. We found that percentage of plasmacytoid dendritic cells (pDC, CD11c^intermediatePDCA-1⁺) in wild-type mice significantly increased after the immunization (Fig 7 panel A). However, it did not show significant change in CD44-deficient mice. On the other hand, the non-pDCs were increased in a similar pattern in both wild-type and CD44-deficient mice after the immunization (Fig 7 panel A). Further, we examined lymphoid dendritic cells (LDC, CD11c⁺CD8α⁺CD11b⁻PDCA-1⁻) and myeloid dendritic cells (mDC, CD11c⁺CD8α⁻CD11b⁺PDCA-1⁻) from cells that were identified as CD11c+ non-pDCs. We found a higher percentage of mDC and a lower proportion of LDC in CD44-deficient mice when compared to wild-type mice (Fig 7 panel B).

Figure 7. Splenic DC subsets in CD44^−/− and CD44^+/+ mice.

DCs from spleen on d10 of primary OVA immunization and d5 ofSRBC immunization were enriched and examined for their subsets by flow cytometry. Panel A: The dot plots show cells double stained for CD11c and PDCA-1. Vertical bars represent data (mean ± SEM) from three mice. Panel B: CD11c⁺PDCA-1⁻ cells from Panel A in each group were gated and further analyzed for expression of CD11b and CD8α. Vertical bars represent data (mean ± SEM) from three mice. The asterisk notations in panel B show statistically significant results in which each CD44^−/− group was compared to its corresponding CD44^+/+ group.

Next, CD4 T cells from OT-II mice were stimulated in vitro with OVA323-339 peptide presented by CD44^−/− or CD44^+/+ DCs. Th1-Th2 differentiation was determined by the intracellular expression of IL-4 and IFN-γ on d4. Similar to the results under the polyclonal stimulation, CD44^−/− DCs induced a decreased Th1 but an enhanced Th2 differentiation (Fig 8A). These results suggested that CD44 expression on DC modified the DC instruction on the CD4 T cell differentiation. IL-12 produced by activated DCs is a key factor to induce Th1 development (29, 30). Therefore, in addition to the measurement of IL-12 production in the serum (Fig 3 and 5B), we examined the capacity of the IL-12 secretion from splenic DCs. After stimulation with LPS or melanoma antigen for 24 h, DCs were activated and produced IL-12. We found that the production of IL-12 was down-regulated in CD44^−/− DC. However, the endogenous production of IL-12 (without LPS or melanoma stimulation) was not affected by the CD44-deficiency (Fig 8B). These results suggested that CD44 deficiency affects DC activation and the consequent capacity to produce IL-12. Together, the above results indicated that the specific state of DCs that is induced by the CD44-deficiency could be an additional factor accountable for the CD44 modulation in Th differentiation.

Figure 8. OVA-specific Th1-Th2 polarization in vitro and IL-12 production.

Panel A: OVA-specific Th1-Th2 polarization in vitro. CD62L^highCD4 T cells from OT-II mice and DCs from the spleens of CD44^−/− or CD44^+/+ mice were co-cultured in Th1- or Th2-development condition with addition of OVA_323-339 peptides as described in Materials and Methods. On d4 of the culture, cells were re-stimulated and stained as described in Fig 6. Panel B: IL-12 production in the culture supernatant. DCs were stimulated with LPS or B16F10 lysate in the indicated concentrations. IL-12 production in the culture supernatant was determined at 24 h by ELISA. * p < 0.01. # p > 0.05.

Discussion

Consistent with the paradigm of the two types of immmune responses (Th1 versus Th2), we observed a Th2-biased immune response in CD44 knockout mice. Deficiency of CD44 inhibited Th1 development and promoted Th2 differentiation. This conclusion was supported by several findings in CD44^−/− mice : 1) the cytokine profile in the serum from imunized mice was dominated by Th2 cytokines, with diminished Th1 cytokines; 2) decreased IgG2a production along with the enhanced IgG1 production; 3) up-regulated GATA3 but down-regulated T-bet expression; 4) decreased response to SRBC as seen using DTH reaction and hemagglutination; 5) In the in vitro culture system, CD44^−/− CD4 T cells were less polarized to Th1, but more readily polarized to Th2. Together, these results suggested that CD44 participates in the regulation of CD4 T cell differentiation into Th1 or Th2 lineage. To the best of our knowledge, this is the first report that demonstrates the modulatory effects of CD44 on Th differentiation both in vivo and in vitro.

CD44-deficient mice possesses normal T cell development as compared to their wild-type counterparts. Our previous studies have showed that there is statistically no significant difference in the proportion of T cells (CD4+, CD8+, CD3+), B cells, and macrophages in the spleen and other peripheral organs between wild-type and CD44-deficient mice (22, 23). These results can rule out the possibility that biased Th development in our study was caused by differential development of T cell subsets in thymus.

When we reconstituted CD4^−/− mice with CD44-deficient CD4 T cells, the hemagglutination antibody production against SRBC was diminshed; however, this diminished response could be completely restored with CD44-sufficient CD4 T cells (Figure 5B). In the in vitro Th1-Th2 polarization experiments, CD44-deficient CD4 T cells showed a less degree of Th1 differentiation than the wild-type counterparts (Figure 6). These results suggested that in situ expression of CD44 on CD4 T cells is indispensable to maintain normal Th1 differentiation and response.

The differential expression of GATA3 and T-bet in this study suggests that the regulatory effect of CD44 could occur at the transcription level (Figure 4). The intracellular domain of CD44 isoforms selectively interacts with different kinanses or transducer proteins and regulates specific signaling that are related to the miscellaneuous functions of the CD4 molecule (15, 16). Specifically, CD44 is tightly coupled with Src kinases, such as Lck and Fyn, and such signaling cascades induce tumor cell migration (16, 31). T-bet and GATA3 are activated through JAK/STAT signaling pathway (4, 5). Accumlating evidence indicates that STAT activation can be mediated by members of both JAK and Src family. The integration of these diverse signaling cues from active Src, JAK and STAT (Src-JAK-STAT model) leads to cell proliferation, survival and differentiation (32, 33). We reckon that CD44-deficiency reforms its original signaling cascades and modulates Src-JAK-STAT interaction, which leads to the differential activation and expression of T-bet and GATA3, and thereafter, Th1-Th2 differentiation.

In vitro differentiation using CD44-sufficient and -deficient DCs produced a similar pattern of Th1-Th2 differentiation (Figure 8A). It suggested that CD44 also influences DC effect on CD4 T cell differentiation. CD44-deficient DCs showed a decreased ability to produce IL-12 in response to the stimulation of LPS and melanoma antigen (Figure 8B). This could be one factor that contributes to the DC instruction toward Th2 differentiation. However, there are other cell sources in vivo for the IL-12 production. As cited, macrophages are another main source of IL-12 and modulators of Th1-Th2 differentiation (8, 29). Macrophage subsets, M1 and M2, initiate and regulate Th1 or Th2 response respectively (8). Blockade of CD44 with antibodies can inhibit IL-12 production from the thioglycollate-activated macrophages (34). Therefore, it is plausible that CD44-deficient macrophages may contribute to the decreased IL-12 production in the serum in our study. Nevertheless, this environment could modify functionalities of DCs that may render CD44-deficient DC a Th2-biased polarizing feature.

The pattern of DC subsets showed a higher percentage of mDC, a decreased proportion of pDC and LDC in CD44-deficient mice when compared to CD44-sufficient mice (Figure 7). These data may be useful for further dissection of differential role of CD44-deficiency on each DC subset in Th development. Other studies have demonstrated that OVA or KLH-loaded-LDC promote Th1 differentiation; and mDC promote Th2 differentiation (9, 10). We assume that SRBC-loaded-mDC and -LDC may also follow the same pattern. Activation of pDC by respiratory syncytial virus and measle virus profoundly promote Th1 and suppress Th2 immune response (35). Also, toxoplasma gondii-activated pDC enhance Th1 immune response (36). pDC is also indispensable for conventional DCs to produce IL-12 in response to L. monocytogens infection (37). It is known that pDC play a more flexible role in directing either Th1 or Th2 development that is dependent on nature and dose of antigen, immunization route, differential toll-like receptor ligation and other factors (38-40). We assume that pDC, in our system, may promote Th1 development. It seems that the DC subsets present in CD44-deficient mice, after immunization, favored an up-regulated Th2 and down-regulated Th1 differentiation. However, further functional studies using isolated CD44-deficient DC subsets are needed in extending the results and determining which subset plays critical role in instruction of Th1-Th2 differentiation.

In addition to the effect of CD44 on CD4⁺ T cells and DCs, the possibility that CD44 influences the B cells also remains. It should be noted that in our previous studies, CD44 expression on B cells affect their differentiation and Ig production while certain Abs against CD44 blocked B cell activation induced by agents such as LPS (41, 42).

Interestingly, in the current study, we noticed a decreased IL-17 production in CD44-deficient mice (Figure 3 and 5B). Recent studies have suggested that IL-17 deficiency can result in both compromised DTH reaction and T-dependent humoral immune response (5, 43). On the other hand, T-bet is important for continued IL-17 production in the presence of IL-23 and regulates the fate of Th1 and Th17 cells in autoimmunity (44, 45). We also found that the number of Th17 cells in SRBC-immunized CD44^−/− mice was decreased, and the differentiation to Th17 lineage was also inhibited in CD44^−/−CD4 T

cells (data not shown). We speculate that CD44 modulation on T-bet/IL-17/Th17 interaction may form a unique mechanism in the regulation of Th development. We are currently testing this hypothesis.

Our study highlights CD44 as a player in Th development. These data may contribute towards better understanding of the pathogenesis of some immune diseases and development of a CD44-targeted modality in treatment of T-cell-elicited immune diseases. For example, recently, we found that CD44-deficient mice revealed attenuated multiple sclerosis in the autoimmune encephalomyelitis (EAE) model (manuscript under preparation). These findings correlated with decreased myelin-specific Th1 and Th17 cells and increased, myelin-specific Th2 cell response. Such a shift is known to suppress EAE, thereby further corroborating the data presented in the current study.

Abbreviations used in this paper

DC

dendritic cell

pDC

plasmacytoid dendritic cell

mDC

myeloid dendritic cell

LDC

lymphoid dendritic cell

DTH

delayed type hypersensitivity

CD44+/+

CD44-sufficient or CD44 wild-type

CD44−/−

CD44 deficient or CD44-knockout.