Vα24-Invariant NKT Cells from Patients with Allergic Asthma Express CCR9 at High Frequency and Induce Th2 Bias of CD3+ T Cells upon CD226 Engagement1

We have demonstrated that Vα24+Vβ11+ invariant (Vα24+i) NKT cells from patients with allergic asthma express CCR9 at high frequency. CCR9 ligand CCL25 induces chemotaxis of asthmatic Vα24+i NKT cells but not the normal cells. A large number of CCR9-positive Vα24+i NKT cells are found in asthmatic bronchi mucosa, where high levels of Th2 cytokines are detected. Asthmatic Vα24+i NKT cells, themselves Th1 biased, induce CD3+ T cells into an expression of Th2 cytokines (IL-4 and IL-13) in cell-cell contact manner in vitro. CD226 are overexpressed on asthmatic Vα24+i NKT cells. CCL25/CCR9 ligation causes directly phosphorylation of CD226, indicating that CCL25/CCR9 signals can cross-talk with CD226 signals to activate Vα24+i NKT cells. Prestimulation with immobilized CD226 mAb does not change ability of asthmatic Vα24+i NKT cells to induce Th2-cytokine production, whereas soluble CD226 mAb or short hairpin RNA of CD226 inhibits Vα24+i NKT cells to induce Th2-cytokine production by CD3+ T cells, indicating that CD226 engagement is necessary for Vα24+i NKT cells to induce Th2 bias of CD3+ T cells. Our results are providing with direct evidence that aberration of CCR9 expression on asthmatic Vα24+i NKT cells. CCL25 is first time shown promoting the recruitment of CCR9-expressing Vα24+i NKT cells into the lung to promote other T cells to produce Th2 cytokines to establish and develop allergic asthma. Our findings provide evidence that abnormal asthmatic Vα24+i NKT cells induce systemically and locally a Th2 bias in T cells that is at least partially critical for the pathogenesis of allergic asthma.

N atural killer T cells, originally characterized in mice as cells that express both a TCR and NK1.1 (Refs. 1 and 2), have more recently been defined as cells that have an invariant V␣14-J␣18 (mouse) or V␣24-J␣18 (human) rearrange-ment and reactivity to ␣-galactosylceramide (␣-GalCer) 4 presented by the CD1d (3). NKT cells play a critical role in modulating the upcoming immune responses, involving in protection against bacterial or parasitic infections and prevention of autoimmune diseases (4 -6). The capacity of NKT cells to activate rapid cytokine expression has been exploited to manipulate the outcomes of autoimmunity and cancer (7). In an experimental asthma model in mouse, V␣14i NKT cells are required to participate in allergen-induced Th2 airway inflammation through a CD1d-dependent mechanism (8). Pulmonary V␣14i NKT cells regulate crucially development of airway eosinophilia, hyperresponsiveness, Th2-cytokine production, and elevated levels of IgE Abs (8,9). In humans, CD1d-restricted TCR V␣24 ϩ i NKT cells are important regulators of immune responses through their efficient secretion of Th1 and Th2 cytokines (10). However, it is not understood how NKT cells can selectively regulate Th1 vs Th2 responses in vivo when they can produce both IL-4 and IFN-␥. Some of these differences are thought to depend on the stimulatory conditions driving the response, including cell-cell interactions and cytokine balance in the microenvironment, as well as the nature of the Ag (11)(12)(13). Up to now, the mechanism that NKT cells influence the development of allergic inflammation through Th2 differentiation of T cells in human remains nebulous.
The recruitment of leukocytes into tissues is dependent on a series of adhesion and activation steps mediated by adhesion molecules and chemokine receptor interactions (10). Distinct subsets of T cells have been characterized that are targeted to skin (CLA ϩ CCR4 ϩ ), small intestine (␣4␤7 high CCR9 ϩ ), and mucosal sites (CCR10 ϩ ) (reviewed in Ref. 10). Several chemokine receptors, such as CCR1, CCR4, CCR6, and CXCR6, are differentially expressed by CD4 ϩ , CD8 ϩ , and double-negative (DN) NKT cell subsets (14 -16). CCR9, mainly on T cells in human small intestine (17), is expressed at a very low level on normal human NKT cell subsets (14,18). Currently, there are very few data on the role of chemokines and their receptors in the recruitment of NKT cells to different tissue sites in vivo, particularly in pathogenesis in humans (10). It will be important to determine the chemokine receptor interactions that mediate both homeostatic trafficking of NKT cells and recruitment into different sites of inflammation.
Adhesion molecule CD226 belongs to an Ig superfamily (19 -27) with ligands CD112 and CD155. This molecule has very important biological functions, such as mediating human NK cell cytotoxicity and signaling transduction of T cell activation and differentiation (19 -27). We have reported the high sensitivity of NKT cells from active systemic lupus erythematosus patients to apoptosis because of CD226 deficiency (28).
In the present study, we have documented that asthmatic V␣24 ϩ i NKT cells selectively express CCR9. Th1-biased asthmatic V␣24 ϩ i NKT cells are inducers of Th2 bias of CD3 ϩ T cells. CCL25/CCR9 signals can cross-talk with CD226 to activate V␣24 ϩ i NKT cells for induction of Th2 bias of CD3 ϩ T cells.  20) pAb and anti-NKG2D (1D11) mAb were purchased from Santa Cruz Biotechnology. ␣-GalCer was purchased from Sigma-Aldrich. We had generated a mouse anti-human mAb against CD226.

Patients, normal subjects (NMLs), and cell purification
In-or outpatients with allergic asthma at Departments of Internal Medicine in the university hospitals of Anhui Medical University and Wuhan University were voluntary recruited. All asthma patients had had at least two typical consecutive onsets and specific IgE Ն CAP class 3 (CAP system; Pharmacia). The group of untreated asthmatic patients (AST) were ages 18 -56 years (14 males and 18 females); the group of asthmatic patients with corticosteroid treatment (AST-t) were ages 16 -51 years (8 males and 10 females); and the group of asthmatic patients in remission (AST-r) were ages 17-48 years (9 males and 9 females). The general statuses of all patients were stable without any severe cardiopulmonary and other vital organ complication. The NMLs were ages 19 -50 years (8 males and 10 females) without history of atopy (hayfever, asthma, and eczema) and with undetectable levels of specific IgE. In the experimental procedure, the asthmatic patients in each group were randomly selected for carrying out each assay with sufficient numbers for statistically analysis. All samples were collected with patients' consent and local research ethics committee approval. PBMCs were isolated from heparinized peripheral blood of patients and healthy volunteers. V␣24 ϩ i NKT cells were positively isolated using V␣24-biotin mAb together with MACS streptavidin-labeled magnetic microbead cell sorting (Miltenyi Biotec), according to the manufacturer's instructions. CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were obtained using first a nylon wood column assay to negatively select CD3 ϩ T cells to avoid preactivation, followed by a depletion of CD56 ϩ or CD161 ϩ cells by CD56 and CD161 MACS multisort microbeads (Miltenyi Biotec) (29). For all magnetic separations, FcRs were blocked with Fc Block (Miltenyi Biotec). The purities of cell populations were analyzed on a flow cytometer. The samples of normal bronchi mucosa from lung tissue were anonymously from the patients without atopy background undergoing partial lung resection for accidental chest trauma. The samples of bronchi mucosa from allergic asthma patients were obtained through fiber-optic bronchoscopy.

Flow cytometry
For detection of chemokine receptors (CCR6, CCR9, or CXCR3), purified CD3 ϩ T cells either isolated from patients with allergic asthma or from normal controls were incubated with FITC-labeled anti-human V␣24 mAb and PE-labeled anti-human V␤11 mAb and PerCP-labeled mouse antihuman chemokine receptor mAb or CD226 (secondary PerCP-labeled) mAb at 5 g/ml or 5 g/ml matched isotype mouse Ab (DakoCytomation). The PerCP-labeled third Ab in PBS containing 2% BSA and 0.1% sodium azide was incubated for 20 min, followed by washing twice in staining buffer as described previously (30). All procedures were conducted at 4°C. The analyses were performed with a flow cytometer (COULTER XL; Coulter). For intracellular cytokine immunofluorescence staining, as described elsewhere (31), the cells were washed twice in PBS and then fixed and permeabilized using IntraPrep (Beckman Coulter), according to manufacturer's instructions. The cells were then incubated with the primary mouse anti-human cytokine mAb for 15 min at room temperature. Cells were washed twice and stained with PerCP-labeled goat anti-mouse Abs for 15 min at room temperature. Cells were washed twice and stained with FITC-labeled anti-human V␣24 mAb and PE-labeled anti-human V␤11 mAb for 15 min at 4°C. Cells were washed and resuspended in PBS containing 0.5% formaldehyde for flow cytometry.
All unknown cDNAs were diluted to contain equal amounts of ␤-actin cDNA. The standards, "no template" controls, and unknown samples were added in a total volume of 50 l/reaction. PCR retain conditions were 2 min at 50°C, 10 min at 95°C, 40 cycles with 15 s at 95°C, and 60 s at 60°C for each amplification. Potential PCR product contamination was digested by uracil-N-glycosylase because dTTP is substituted by dUTP (32). Uracil-N-glycosylase and AmpliTaq Gold (Applied Biosystems) were applied according to the manufacturer's instructions (33,34).

Western and Northern blotting
For mRNA detection (Northern blot), as previously described (35), 5 g of total RNA from each sample were electrophoresed under denaturing conditions, followed by blotting onto Nytran membranes, and cross-linked by UV irradiation as described previously (35). The target cDNA probes, labeled by ␣-[ 32 P]dCTP, were obtained by PCR amplification of the sequences mentioned above from total RNA from PBMCs from normal adults (for CCR6, CXCR3, V␣24, and CD226) or thymocytes from the specimen of thymusectomy (for CCR9). The membranes were hybridized overnight with 1 ϫ 10 6 cpm/ml of 32 P-labeled probe, followed by intensively washing with 0.2ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl, 0.015 M sodium citrate (pH 7.0)) and 0.1% SDS before being autoradiographed. For protein detection (Western blot analysis), the cells were lysed in lysis buffer (0.4% Nonidet P-40, 150 mM NaCl, 50 mM Tris-HCl, 1 mM sodium vanadate, 0.1 mM PMSF, and 2 g/ml leupeptin and apoprotein (pH 8.0)) as described previously (36). Lysates were centrifuged at 10,000 rpm for 5 min at 4°C. Protein concentration was measured by Bio-Rad protein assay (Bio-Rad). Protein (ϳ40 g) was loaded onto 16% SDS-PAGE, transferred onto PVDF membranes after electrophoresis, and incubated with appropriate Abs (0.5 g/ml). Analyses were conducted using ECL detection (Amersham Biosciences). For coimmunoprecipitations and phosphorylated protein detection (19), cell lysates were precleared three times with 20 l of protein A-Sepharose beads and were mixed with specific Ab (anti-CD226) for 3 h at 4°C under constant agitation. Immune complexes were allowed to bind to 20 l of protein A-Sepharose beads overnight, beads were washed three times with lysis buffer, and immunoprecipitates were separated on 12% SDS polyacrylamide gels and transferred to nitrocellulose membranes. Filters were blocked with 5% nonfat milk in blocking buffer and incubated with specified rabbit antiphosphorylated proteins Ab (Zymed Laboratories) for 2 h, followed by HRP-labeled secondary Ab (Amersham Biosciences) visualizing procedure.

Chemotaxis assay
The chemotaxis assay was performed in a 48-well microchamber (Neuro-Probe) technique (37). Briefly, chemokines were diluted in RPMI 1640 with 0.5% pooled human serum and placed in the lower wells (25 l). Twenty-five microliters of cell suspension at 2 ϫ 10 6 cells/ml were added to the upper well of the chamber, which was separated from the lower well by a 5-m pore size, polycarbonate, polyvinylpyrolidone-free membrane (Nucleopore). The chamber was incubated for 60 min at 37°C in an atmosphere containing 5% CO 2 . The membrane was then carefully removed, fixed in 70% methanol, and stained for 5 min in 1% Coomassie brilliant blue. The cells that migrated and adhered to the lower surface of the membrane were counted by using a light microscopy. Approximately 6% of the cells will migrate spontaneously (known as migrating cells on negative control, MCNCs) (38). The results were expressed as chemotactic index (CI) that was the ratios between the numbers of migrating cells in the sample and in the medium control (37) and with SD.

Laser confocal scanning microscopy
As previously described (39 -41), mAbs were labeled with ZnS-coated CdSe quantum fluorescent dots (ZnS/CdSe QDs). The biopsy slides were incubated with 100 l of fluorescent QDs-Ab conjugate colloids for 3 h. All samples were then rinsed with PBS (containing 0.5% Tween 20) and dried under argon before scanning fluorescence imaging. The 532-nm laser line was used for excitation of immunolabeled samples. Confocal microscopy analysis of the samples was performed using a confocal laser scanning microscope system (LSMSIO; Zeiss).

Dendritic cell (DC) generation and T cell stimulation
The DCs were generated from PBMCs as described previously (42). Briefly, monocytes were purified by positive sorting using anti-CD14-conjugated magnetic Dynabeads M-450 (Dynal). The recovered cells (Ͼ95 and Ͻ99% CD14 ϩ cells) were cultured at 3 ϫ 10 5 /ml in RPMI 1640 with 10% FCS supplemented with 25 ng/ml GM-CSF (R&D Systems) and 10 ng/ml IL-4 (R&D Systems) for 7 days. Maturation was induced by addition of 1 g/ml LPS (Sigma-Aldrich) and 50 ng/ml TNF-␣ (R&D Systems) for the last 40 h of cultures. For an optimal stimulation of freshly isolated CD3 ϩ CD56 Ϫ CD161 Ϫ T cell populations, mature syngeneic DCs were applied for stimulation (43). Freshly isolated CD3 ϩ CD56 Ϫ CD161 Ϫ T cells (2 ϫ 10 5 cells/well) were used for primary stimulation in the presence of optimal numbers of syngeneic DCs in 24well plates (400 l/well) for 2 days. A secondary stimulation was performed in the addition of different numbers of purified syngeneic V␣24 ϩ i NKT cells as indicated in the figure legends for 4 days in the presence of tetanus toxoid (2.5 g/ml). A total 6 days after the onset of the primary culture, CD3 ϩ T cells were harvested using a positive selection procedure of anti-CD3 mAb-coated MACS beads assay (Miltenyi Biotec), followed by real-time quantitative RT-PCR or intracellular cytokine immunofluorescence staining assay. In some cases, Transwell experiments were done in 24-well plates as described previously (44). Briefly, primary stimulation systems of CD3 ϩ CD56 Ϫ CD161 Ϫ T cells in the presence of optimal numbers of syngeneic DCs (after 2 days) were placed in Transwell chambers (Millicell, 0.4 m; Millipore) in the presence of different numbers of purified syngeneic V␣24 ϩ i NKT cells as indicated in the figure legends for 4 days in the presence of tetanus toxoid (2.5 g/ml). After 4 days of culture, activated CD3 ϩ T cells were harvested using MACS beads before further investigation.

Gene silencing assay
Short hairpin RNAs (shRNAs) were produced in vitro as previously described (45) using chemically synthesized DNA oligonucleotide templates (Sigma-Aldrich). Transcription templates were designed such that they contained U6 promoter sequences at the 5Ј end. Short hairpin RNA transcripts subjected to in vitro Dicer processing were synthesized using a Riboprobe kit (Promega). Double-stranded DNA oligonucleotides encoding shRNAs with 20 bases homologous to the targeted CD226 gene were ligated into the EcoRV site to produce expression constructs. Sequences inserted immediately downstream of the U6 promoters were as follows: CD226 sense pU6hairpin20, AGAACCAGCCTTTCAAACAGtt. Cells were transfected with indicated amounts of shRNA and plasmid DNA using standard calcium phosphate procedures at 50 -70% confluence in 6-well plates. Cells were cultured in DMEM containing 10% of heat-inactivated FBS, penicillin, and streptomycin. Cells were harvested 2 days after the transfection.

Cell transient transfection
As described elsewhere (46,47), human V␣24 ϩ i NKT cells were transfected with CMV vector pcDNA3/CCR9 using the Amaxa Nucleofection Technology (Amaxa) with optimization. Briefly, cells were resuspended in solution from nucleofector kit V, following the Amaxa guidelines for cell transfection. One hundred microliters of 3 ϫ 10 6 cell suspension mixed with 0.25 or 2.5 g of cDNA were transferred to the provided cuvette and nucleofected with an Amaxa Nucleofector apparatus (Amaxa).

Highly increased frequency of asthmatic CCR9-expressing V␣24 ϩ i NKT cells
We investigated the expression of all known chemokine receptors on V␣24 ϩ i NKT cells from patients with allergic asthma. The results were either in agreements with previous reports or not different in comparison with that from normal subjects (data not shown; Ref. 10,18,48). It was reported that normal V␣24 ϩ i NKT cells expressed CCR6 and CXCR3 at a high frequency (10,14,18,48). Our results by flow cytometry showed that normal and asthmatic V␣24 ϩ i NKT cells expressed CCR6 and CXCR3 at identical frequency (Fig. 1A). CCR9 was selectively expressed at high frequency on the asthmatic V␣24 ϩ i NKT cells (87%) (Fig. 1A). To confirm the observation, we conducted the real-time quantitative RT-PCR assay to detect the expressions of different chemokine receptors at mRNA level. Normal and asthmatic V␣24 ϩ i NKT cells expressed CCR6 and CXCR3 mRNAs at identical levels. Asthmatic V␣24 ϩ i NKT cells expressed high level CCR9 mRNA, in comparison with normal cells (Fig. 1B). The observation on CCR6, CCR9, and CXCR3 expressions in normal and asthmatic V␣24 ϩ i NKT cells was confirmed by Northern blot (Fig. 1C) and Western blot (Fig. 1D) analyses. Additionally, normal and asthmatic CD3 ϩ CD56 Ϫ CD161 Ϫ T cells (NK and NKT cells were depleted) expressed indistinguishably CCR9 mRNA and protein at very low levels (data not shown).
We next examined the abilities of CCL20 (a ligand for CCR6), CCL25 (a ligand for CCR9), and CXCL9 (a ligand for CXCR3) to induce chemotaxis of normal and asthmatic V␣24 ϩ i NKT cells. CCL20 and CXCL9 induced significant chemotactic migration of normal and asthmatic V␣24 ϩ i NKT cells (Fig. 2). CCL25 induced significant chemotaxis of asthmatic V␣24 ϩ i NKT cells but not normal cells (Fig. 2B). To confirm CCL25 via CCR9 to induce chemotaxis, we used anti-CCR9 mAb to block the chemotactic activity of asthmatic V␣24 ϩ i NKT cells toward CCL25. The anti-CCR9 mAb could completely block the chemotaxis of the cells toward CCL25 (data not shown), whereas it had no any effect on the chemotaxis of the cells toward CCL20 and CXCL9 (data not shown). The isotype Ab had no blocking effect at all (data not shown).
By knowing that pulmonary V␣14i NKT cells were crucially regulating development of asthma in the mouse model (8,9), we next examined NKT cells infiltrating into bronchi mucosa from allergic asthma patients. Morphology of immunohistochemistry stained with fluorescent QD-conjugated Ab colloids revealed that a significantly increased amount of CCR9 ϩ cells was found in submucosa of patients with allergic asthma in comparison with NMLs (upper panels in Fig. 3A). Meanwhile, a large number of V␣24 ϩ i cells were also found in the submucosa of ASTs in comparison with NMLs (lower panels in Fig. 3A). CCR9 mRNA and V␣24 mRNA were highly increased in the mucosa samples from ASTs compared with that from the NMLs (Fig. 3B). The patterns of CCR9 and V␣24 expression in the NMLs and ASTs were confirmed by Northern blot (Fig. 3C) and Western blot (Fig. 3D) analyses. Normal and asthmatic CD3 ϩ CD56 Ϫ CD161 Ϫ T cells expressed indistinguishably CCR9 at low mRNA and protein levels (mentioned above). The specific marker V␣24 chain was also highly expressed in asthmatic mucosa. Therefore, we could reasonably conclude that CCR9-expressing V␣24 ϩ i NKT cells were infiltrating into bronchi mucosa in large scale during asthma pathogenesis.

Asthmatic CCR9-expressing V␣24 ϩ i NKT cells induced Th2 bias in CD3 ϩ T cells
Through their efficient secretion of Th1 and Th2 cytokines, V␣24 ϩ i NKT cells were important regulators of immune responses (10). To explore how V␣24 ϩ i NKT cells could selectively regulate Th1 vs Th2 responses in vivo when they could produce both IL-4 and IFN-␥, we further investigated the expression of cytokines in asthmatic and normal V␣24 ϩ i NKT cells by real-time quantitative RT-PCR assay. Normal V␣24 ϩ i NKT cells expressed low level of Th1 (IFN-␥ and IL-2), Th2 (IL-4 and IL-13), and regulatory T cell (Tr)-derived (IL-10 and TGF-1␤) cytokines (Fig.  4A). Purified and unstimulated asthmatic V␣24 ϩ i NKT cells selectively and highly expressed Th1 (IFN-␥ and IL-2) cytokines as well as Tr-derived (IL-10 and TGF-1␤) cytokines (Fig. 4A). Interestingly, both normal and asthmatic V␣24 ϩ i NKT cells expressed Th2 (IL-4 and IL-13) cytokine at identical level. Asthmatic V␣24 ϩ i NKT cells expressed significantly higher levels of CXCL9, identical levels of CCL20 and CCL25, in comparison with normal cells (Fig. 4A). The observation of cytokine and chemokine expression was confirmed at mRNA levels by Northern blot analysis (Fig. 4B). IL-2, TGF-1␤, and CCL20 data were not shown. We further investigated the expression of cytokines and chemokines in bronchi mucosa from patients with allergic asthma and NMLs by real-time quantitative RT-PCR and Northern blot assays. The expression levels of Th1 (IFN-␥ and IL-2) cytokines in normal and asthmatic bronchi mucosa were identical (Fig. 4C). Interestingly, Th2 (IL-4 and IL-13) and Tr-derived (IL-10 and TGF-1␤) cytokines were selectively and highly up-regulated in asthmatic bronchi mucosa, in comparison with normal subjects (Fig. 4C). Moreover, both CCL25 and CXCL9 expressions in asthmatic bronchi mucosa were significantly increased, whereas CCL20 expressions in asthmatic and normal bronchi mucosa were identical (Fig. 4C). The observation was confirmed at mRNA levels by Northern blot analysis (Fig. 4D). Data on IL-2, TGF-1␤, and CCL20 were not shown. We investigated cytokine expression on subsets of CD4 Ϫ CD8 Ϫ , CD4 ϩ , or CD8 ϩ V␣24 ϩ i NKT cells from different subjects. Normal CD4 Ϫ CD8 Ϫ V␣24 ϩ i NKT cells expressed higher Th1 cytokines, whereas normal CD4 ϩ V␣24 ϩ i NKT cells had both Th1 and Th2 cytokines. Asthmatic CD4 Ϫ CD8 Ϫ and CD4 ϩ V␣24 ϩ i NKT cells both expressed significantly higher IL-2 and IFN-␥ (Table I). It was well established that unactivated NKT cells constitutively expressed IL-4 and IFN-␥ mRNA without production of the relevant cytokine (49). Therefore, we investigated the cytokine production profile of V␣24 ϩ i NKT cells upon stimulation. Considering the possibility that V␣24 ϩ i NKT cells in ASTs could be preactivated and potentially had down-regulated their TCRs, we chose ␣-GalCer as the stimulus to optimally activate the cells. Asthmatic V␣24 ϩ i NKT cells expressed significant higher IL-2, IFN-␥, IL-10, and TGF-␤ at mRNA and protein levels than that in normal cells upon stimulation with ␣-GalCer (Table II). Thus, the results were suggesting that asthmatic V␣24 ϩ i NKT cells were Th1 biased and expressed high levels of Tr-derived (IL-10 and TGF-1␤) cytokines and CXCL9, which is a ligand for CXCR3 that was highly up-regulated on activated CD3 ϩ T cells (50). Upon arrival to allergic inflammatory site, V␣24 ϩ i NKT cells could induce the bias of other immune cells such as CD3 ϩ T cells.
To further confirm the function of different V␣24 ϩ i NKT cells, we had cocultured the V␣24 ϩ i NKT cells with syngeneic CD3 ϩ CD56 Ϫ CD161 Ϫ T cells in the presence of mature syngeneic DCs and subsequently investigated some cytokine expressions in CD3 ϩ T cells. The results from flow cytometric and real-time quantitative RT-PCR assays documented that normal V␣24 ϩ i  NKT cells could significantly drive CD3 ϩ T cells into an expression of Th1 cytokine (IFN-␥) in a concentration-dependent manner, whereas the asthmatic cells did not show such ability (Fig.  5A). We observed similar phenomena of IL-2 expression (data not shown). In contrast, asthmatic V␣24 ϩ i NKT cells could significantly drive CD3 ϩ T cells into expressions of Th2 cytokines (IL-4 and IL-13), whereas the normal cells did not show such ability (Fig. 5, B and C). Interestingly, both normal and asthmatic V␣24 ϩ i NKT cells could significantly drive CD3 ϩ T cells into an expression of Tr1 cytokine IL-10 in a concentration-dependent manner ( Fig. 5D; data on TGF-1␤ not shown).
To elucidate the mechanism of functions of Th1-biased V␣24 ϩ i NKT cells to regulate the expression of Th2-cytokine bias of syngeneic CD3 ϩ CD56 Ϫ CD161 Ϫ T cells, we analyzed the properties of the cells in Transwell stimulations in comparison with the original condition. The semipermeable Transwell membrane prevents from direct cell-cell contact between the responsive CD3 ϩ T cells and the V␣24 ϩ i NKT cells. In contrast to the findings mentioned above, asthmatic V␣24 ϩ i NKT cells lost the functions to induce Th2 bias of syngeneic CD3 ϩ T cells (Fig. 5, E and F). These data strongly suggested that the cytokine-expression-directing function of V␣24 ϩ i NKT cells was cell-cell contact dependent.

Asthmatic CCR9-expressing V␣24 ϩ i NKT cells functioned upon CD226 activation
Recently, we described that expression and activation of CD226 on NKT cells were involved into survival and other intracellular signaling of the cells (28). To further search the mechanism of Th1biased asthmatic CCR9-expressing V␣24 ϩ i NKT cells to induce Th2 bias of syngeneic CD3 ϩ T cells, we examined the expression of CD226 on the normal and asthmatic V␣24 ϩ i NKT cells. The results from flow cytometric analyses documented that there were Ͻ1% V␣24 ϩ i NKT cells in total T cell population in normal subjects and 68% of CD226 positive thereof. There were 15% V␣24 ϩ i NK T cells in total T cell population in patients with allergic asthma and 97% of CD226 positive thereof (Fig. 6A). Other T cell populations in NMLs and patients with allergic asthma were expressed low levels of CD226 (17 and 19%, respectively). To confirm the observation mentioned above, we conducted the real-time quantitative RT-PCR assay to detect the different expressions of CD226 at mRNA level. Asthmatic V␣24 ϩ i NKT cells expressed significantly higher level of CD226 mRNA than normal cells (Fig.  6B). The observation of CD226 mRNA and protein expressions in asthmatic and normal V␣24 ϩ i NKT cells was confirmed by Northern and Western blot analyses (Fig. 6B).  To further investigate the correlation between the high expression of CCR9, Th2 bias of CD3 ϩ CD56 Ϫ CD161 Ϫ T cells and CD226 expression, we applied a series of chemokines, immobilized TCR mAbs, specific NKT surface molecule mAb, and substance to stimulate normal and asthmatic V␣24 ϩ i NKT cells. The results by immunoprecipitation and phosphorylated protein detection revealed that only CCL25/CCR9 and immobilized anti-CD226 mAb caused significant phosphorylation of CD226 protein in asthmatic V␣24 ϩ i NKT cells but not in normal cells (Fig. 6C), indicating that CCL25 by means of high-expressed CCR9 activated asthmatic V␣24 ϩ i NKT cells in association with CD226 activation. Immobilized TCR mAbs (anti-CD3 and V␣24 mAbs) induced slight phosphorylation of CD226 protein in both normal and asthmatic V␣24 ϩ i NKT cells (Fig. 6C). The expression of The CD3 ϩ T cells were purified from NMLs and ASTs. The CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were then cocultured with optimal numbers of mature syngeneic DCs and with indicated different numbers of purified syngeneic V␣24 ϩ i NK T cells in the presence of tetanus toxoid as described in Materials and Methods. CD3 ϩ T cells were then harvested using MACS beads. The cytokine protein and mRNA detection by flow cytometry and real-time quantitative RT-PCR were conducted as described in Materials and Methods. The numbers listed in the figure were in percentage detected with intracellular cytokine assay and cytokine mRNA copies. The illustrated data were mean values (Ϯ SD) of eight performed experiments. Statistically significant differences as compared with normal controls where indicated ‫,ء(‬ p Ͻ 0.001). E and F, The purified CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were then cocultured with optimal numbers of mature syngeneic DCs and with indicated different numbers of purified syngeneic V␣24 ϩ i NK T cells in the Transwell manner (]), in the presence of tetanus toxoid, followed by a procedure of CD3 ϩ cell purification, protein, and mRNA detections as described above. Purified V␣24 ϩ i NKT cells from NMLs and ASTs were stimulated with ␣-GalCer (5 g/ml) for 2 days before cytokine measurements.
b,c The cytokines were determined by real-time quantitative RT-PCR (mRNA copies per 25 ng of cDNA) and flow cytometry (FL, percentage of positive cells) as described in Materials and Methods. ‫,ء‬ p Ͻ 0.001, n ϭ 6, ASTs versus NMLs. CD155 (PVR) and CD112 (nectin-2) (21) on asthmatic V␣24 ϩ i NKT cells was under detectable level (data not shown), excluding the possibility of CD226 ligands (CD155 and CD112) on neighboring V␣24 ϩ i NKT cells to activate the cells in culture.
To further characterize the function of CD226 on different V␣24 ϩ i NKT cells, we had cocultured the pretreated V␣24 ϩ i NKT cells with syngeneic CD3 ϩ CD56 Ϫ CD161 Ϫ T cells in the presence of mature syngeneic DCs, subsequently investigated some cytokine expression levels in CD3 ϩ T cells. CCL25-prestimulated asthmatic V␣24 ϩ i NKT cells could significantly increase IL-4 and IL-13 expressions in CD3 ϩ T cells, whereas prestimulation with immobilized CD226 mAb in asthmatic V␣24 ϩ i NKT cells did not significantly change the ability of Th2-cytokine production (Fig. 7A). In contrast, preblockage with soluble CD226 mAb in asthmatic V␣24 ϩ i NKT cells could inhibit Th2-cytokine production in syngeneic CD3 ϩ CD56 Ϫ CD161 Ϫ T cells (Fig. 7A).
Furthermore, prestimulation with immobilized CD226 mAb did not significantly change the chemotactic migration toward CCL25 in asthmatic V␣24 ϩ i NKT cells, whereas preblockage with soluble CD226 mAb could inhibit chemotaxis against CCL25 in asthmatic V␣24 ϩ i NKT cells (Fig. 7B).
To be sure of the results on correlation between CD226 activation and V␣24 ϩ i NKT cell-induced bias of CD3 ϩ CD56 Ϫ CD161 Ϫ T cells, we applied shRNA of CD226 (shRNACD226) to knockdown CD226 expression and activation. The shRNACD226 at high concentration completely abolished CD226 expression in V␣24 ϩ i NKT cells at both mRNA and protein levels. Neither lowconcentration shRNACD226 nor DNACD226 nor vector had such effects (Fig. 8A). Only high-concentration shRNACD226 significantly blocked the effects of CCL25-pretreated asthmatic V␣24 ϩ i NKT cells on Th2 bias of CD3 ϩ CD56 Ϫ CD161 Ϫ T cells (Fig. 8, B and C). The illustrated data were mean values (Ϯ SD) of eight performed experiments. Northern and Western blots of CD226 mRNA and protein in freshly isolated V␣24 ϩ i NK T cells from NMLs and ASTs. The illustrated data were from a single representative experiment of eight performed. C, The peripheral V␣24 ϩ i NKT cells from NMLs and ASTs were freshly isolated and followed stimulation with chemokines at optimal concentration (100 ng/ml), or with immobilized mAbs, or substance (␣-GalCer at 100 ng/ml in the presence of syngeneic monocyte-derived DCs; Ref. 58) for 60 min as indicated before cell lysis. The stimulating procedures were described in Materials and Methods. CD226 proteins were immunoprecipitated from the cell lysates for detection of levels of phosphorylation using a rabbit antiphosphorylated protein Ab. The levels of pan-CD226 proteins in each identical sample were detected by a specific anti-CD226 mAb. The immobilized anti-CD226 mAb was used as a positive control (28).
Thus, CCL25 ligating with CCR9 directly activated CD226 and enhanced functions of asthmatic V␣24 ϩ i NKT cells. CD226 engagement is necessary for asthmatic V␣24 ϩ i NKT cells to induce Th2 bias of CD3 ϩ T cells.

CCR9 transfection, asthma remission, and medication affected CCR9 and CD226 expression in V␣24 ϩ i NKT cells
To be sure of the results of correlation among CCR9 expression, CD226 activation in asthmatic V␣24 ϩ i NKT cells, and induction of Th2 bias in CD3 ϩ T cells, we applied Amaxa nucleofection technology to transfect cDNA encoding CCR9 into normal V␣24 ϩ i NKT cells to observe the effect of CCR9 expression on induction of Th2 bias in CD3 ϩ T cells. The asymptomatically remised or effective corticosteroid-treated patients were also recruited to examine V␣24 ϩ i NKT cell functions. The data demonstrated that successful transfection of CCR9 took place (Fig. 9A,  upper panels). The remised and successful corticosteroid-treated patients were asymptomatic. CCR9 and CD226 expressions were significantly decreased on the V␣24 ϩ i NKT cells from these asthma patients detected by flow cytometry (Fig. 9A), real-time quantitative RT-PCR assay (Fig. 9B), and Northern and Western blot analyses (Fig. 9C), in comparison with untreated asthma patients. The results from flow cytometric (data not shown) and realtime quantitative RT-PCR assays documented that CCR9-transfected normal V␣24 ϩ i NKT cells could significantly drive CD3 ϩ T cells into expressions of IL-4 and IL-13 in a concentration-dependent manner, whereas the cells from allergic asthma patients with remission or successful corticosteroid treatment had lost such ability in comparison with untreated asthma patients (Fig. 9D).
Together, asthmatic V␣24 ϩ i NKT cells selectively expressed CCR9 at a high frequency. Asthmatic V␣24 ϩ i NKT cells, themselves Th1 biased, significantly induced Th2-cytokine expressions in CD3 ϩ T cells in cell-cell contact manner in vitro. Asthmatic V␣24 ϩ i NKT cells overexpressed CD226. CCL25/CCR9 caused direct phosphorylation of CD226, indicating that CCL25/CCR9 signals cross-talked with CD226 signals to activate the V␣24 ϩ i NKT cells for induction of Th2 bias in CD3 ϩ T cells. The CD3 ϩ T cells were purified from NMLs and ASTs. The CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were then cocultured with optimal numbers of mature syngeneic DCs and with indicated different numbers of purified syngeneic V␣24 ϩ i NK T cells in the presence of tetanus toxoid as described in Materials and Methods. The cells V␣24 ϩ i NKT cells were prestimulated or preblocked (soluble CD226 mAb at 5 g/ml for 8 h at room temperature) as indicated. CD3 ϩ T cells were then harvested using MACS beads. The cytokine protein and mRNA detections by flow cytometry and real-time quantitative RT-PCR were conducted as described in Materials and Methods. The numbers listed in the figure were in percentage detected with intracellular cytokine assay and cytokine mRNA copies. The illustrated data were mean values (Ϯ SD) of seven performed experiments. Statistically significant differences as compared with normal controls where indicated ‫,ء(‬ p Ͻ 0.001). B, The migration of V␣24 ϩ i NK T cells from ASTs toward CCL20, CCL25, and CXCL9. V␣24 ϩ i NK T cells were preblocked with soluble CD226 mAb (5 g/ml for 8 h at room temperature) (Ⅺ) or preactivated with immobilized CD226 mAb (f) before chemotaxis assay. All results were determined as described in Materials and Methods and expressed as chemotactic index (CI Ϯ SD) and based on triplicate determination of chemotaxis on each concentration of chemokine applied. The applied chemokine concentrations were indicated as ng/ml. The spontaneous migration toward negative (MCNC) control was known as basic line in each experiment. Statistically significant differences as compared with between unstimulation and stimulation where indicated ‫,ء(‬ p Ͻ 0.001). The illustrated data were mean values (Ϯ SD) of seven performed experiments.

Discussion
For the definition of NKT cells, it is apparently inadequate that the simple classification of NKT cells as T cells also express NK receptors. This issue has been discussed in an excellent review in detail (51). The present study has been focusing on the group of cells that are CD1d dependent and express an invariant V␣24-J␣18 TCR rearrangement (7,51). In C57BL/6 mice, the majority of NK1.1 (CD161)-expressing T cells are in fact a CD1d-reactive subgroup (52). However, in humans, a much larger set of T cells expresses CD161, and the V␣24/V␣11-expressing subset is only a minority (6). NKT cells have now been shown to control various immune responses, including autoimmune, allergic, antitumor, and antimicrobial immune responses. We have shown in this study that there is elevated frequency of V␣24 ϩ i NKT cells from peripheral blood in patients with allergic asthma in comparison with that in normal subjects (Fig. 1A). Th1 (IFN-␥ and IL-2) and Trderived (IL-10 and TGF-1␤) cytokines are selectively and highly expressed in purified and unstimulated asthmatic V␣24 ϩ i NKT cells (Fig. 4A). There are no differences on Th2 (IL-4 and IL-13) cytokine expression between asthmatic and normal V␣24 ϩ i NKT cells. Asthmatic V␣24 ϩ i NKT cells are infiltrating into bronchi mucosa. The infiltrating cells cause selective up-regulation of Th2 (IL-4 and IL-13) and Tr-derived (IL-10 and TGF-1␤) cytokines in bronchi mucosa in comparison with that in NMLs (Fig. 4C). Asthmatic V␣24 ϩ i NKT cells can significantly drive CD3 ϩ T cells into an expression of Th2 cytokines (IL-4 and IL-13) (Fig. 5, B and C). To our knowledge, this is the first report on high frequency and Th1-biased cytokine pattern of asthmatic V␣24 ϩ i NKT cells. The results suggest that V␣24 ϩ i NKT cells infiltrate into local inflammatory sites in allergic asthma and induce other T cells to produce Th2 cytokines, which are crucial factors for development of airway eosinophilia, hyperresponsiveness, and elevated levels of IgE Abs. The knowledge is providing with direct evidence that the abnormality of V␣24 ϩ i NKT cells to induce systemically and locally Th2 bias of T cells is responsible for, at least partially, pathogenesis of allergic asthma.
V␣24 ϩ i NKT cells include at least two subsets, distinguishable as CD4 ϩ or CD4 Ϫ , although in humans some of these cells express CD8 (53). The relative frequency of both the CD4 ϩ and CD4 Ϫ subsets of NKT cells varies in a tissue-specific manner (54). Under normal condition, V␣24 ϩ i NKT cells resemble effector T cells because CD4 ϩ , CD8 ϩ , and DN subsets uniformly express nonlymphoid tissue-homing chemokine receptors, such as CCR2, CCR5, and CXCR3 (14 -16). In contrast, only few of V␣24 ϩ i NKT cells express CCR7, and almost no peripheral V␣24 ϩ i NKT cells express CXCR5 (14 -16). Many more CD4 ϩ NKT cells express CCR4, whereas CCR1, CCR6, and CXCR6 are preferentially expressed by DN and CD8 ϩ NKT cells (14 -16). CD1d-unrestricted NKT cells (CD3 ϩ CD56 ϩ ) share a similar pattern of chemokine receptor expression with V␣24 ϩ i NKT cells (55), suggesting that these TCR-diverse cells can migrate to similar sites. CCR4 is only detected at significant levels on CD4 ϩ V␣24 ϩ i NKT cells, suggesting that these cells can migrate to additional sites. In the present study, we have found that CCR9 is selectively expressed at a high frequency on asthmatic V␣24 ϩ i NKT cells, systematically and locally (Figs. 1 and 3), whereas it is present at very a low level on the normal cells. CCR9 is coexpressed with ␣4␤7 on gut-homing lymphocytes and promotes adhesion to MAdCAM-1 on mucosal vessels. The most compelling evidence comes from studies showing that only DCs from mesenteric lymph nodes and Peyer's patches are able to imprint a CCR9 ϩ␣ 4␤7 ϩ phenotype on lymphocytes that directs homing to the gut after adoptive transfer in vivo (17). The mucosal CCR9 ϩ T cells are infiltrating to liver in primary sclerosing cholangitis by aberrant expression of the gutspecific chemokine CCL25 (56). A nice recent study provides strong evidence that the memory subset of circulating CCR9 ϩ CD4 ϩ T cells in NMLs has characteristics of mucosa T FIGURE 8. Effects of shRNACD226 on cytokine expression in V␣24 ϩ i NKT cells. V␣24 ϩ i NKT cells from NMLs and ASTs were cultured for 2 days in the presence or in the absence of shRNA as indicated. Vector (2 g); DNACD226, DNA CD226 sequence (2 g); shRNACD226 (l), low concentration (0.02 g); and shRNACD226 (h), high concentration (2 g). Cells were then treated at the presence of CCL25 (100 ng/ml). A, The CD226 was examined using Northern (upper panel) and Western blot (lower panel) analyses as described in legend for Fig. 6. B and C, Intracellular Th2-cytokine (IL-4 and IL-13) protein and mRNA detection by flow cytometry (left panels) and real-time quantitative RT-PCR (right panels) as indicated. The CD3 ϩ T cells were purified from NMLs and ASTs. The CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were then cocultured with optimal numbers of mature syngeneic DCs and with indicated different numbers of purified syngeneic CCL25-pretreated V␣24 ϩ i NKT cells in the presence of tetanus toxoid as described in Materials and Methods. The V␣24 ϩ i NKT cells were precultured with shRNA as indicated. CD3 ϩ T cells were then harvested using MACS beads. The illustrated data were in percentage detected with intracellular cytokine assay and copies of cytokine mRNA. The data were mean values (Ϯ SD) of six performed experiments. Statistically significant differences as compared between vector and siRNA where indicated ‫,ء(‬ p Ͻ 0.001). cells and contains cells with either Th1 or Tr1 cytokine profiles (57). This subset of CCR9 ϩ T cells has a phenotype of activated cells and constitutively express the costimulatory molecules CD40L and OX-40. Peripheral CCR9 ϩ CD4 ϩ T cells proliferate to anti-CD3 or anti-CD2 stimulation, and produce high levels of IFN-␥ and IL-10 (57). Our results are providing with direct evidence on the abnormality of CCR9 expression on asthmatic V␣24 ϩ i NKT cells. CCL25 is shown for the first time promoting the recruitment of CCR9 ϩ V␣24 ϩ i NKT cells into the lung and the establishment and development of allergic inflammation.
Adhesion molecule CD226 mediates NK cell cytotoxicity and signaling transduction of T cell activation and differentiation. Most recent evidence show that CD226 activation can rescue NKT cells from apoptosis (28). The results that CCL25/CCR9 causes directly phosphorylation of CD226 protein in V␣24 ϩ i NKT cells (Fig. 6C) are indicating that CCL25/CCR9 ligation can cross-talk with CD226 signals to activate V␣24 ϩ i NKT cells. These cellular signaling pathways may be important to result in induction of Th2 bias of CD3 ϩ T cells from patients with allergic asthma. The essential correlation between the CCL25/CCR9 ligation and CD226 signals, the abnormality of CD226, e.g., the increase in CD226 expression levels, and expressions of Th2-biased cytokine in allergic asthma should be subjected to further investigations. mission, and corticosteroid treatment on CCR9 and CD226 expression on V␣24 ϩ i NK T cells and their functions. The V␣24 ϩ i NK T cells were purified from NMLs and ASTs (untreated, remission, or corticoid treatment). Untreated patients (AST), onset of the patients with allergic asthma with typical symptoms. Remission (AST-r), the patients were asymptomatic because they were out of season for at least 3 mo. Corticosteroid treatment (AST-t), the patients were administrated Prednison as a major therapy (30 Ϯ 5 mg/ day) and asymptomatic. CCR9 transfection (CCR9 ϩ ), normal V␣24 ϩ i NK T cells were purified and transfected with vectors encoding CCR9 (600 ng) as described in Materials and Methods before further assays. A, The NKT cells were freshly isolated or transfected and followed V␣24 ϩ V␤11 ϩ staining for V␣24 ϩ i NK T cells and consequent CCR9 or CD226 staining as described in Materials and Methods. Isotype Ab controls were expressed as dished curves. The numbers represent percentages of CCR9-or CD226-positive cells as shown in the figure. The data were from a single experiment, which was representative of eight similar experiments performed. B and C, The real-time quantitative detection of RT-PCR, Northern blot, and Western blot for mRNA and protein of CCR9 or CD226 in V␣24 ϩ i NK T cells from NMLs (CCR9-transfected) and AST patients (untreated, remission, and corticosteroid treatment). The procedure for quantitative RT-PCR amplification was described in Materials and Methods. The illustrated data were mean values (Ϯ SD) of six performed experiments. ‫,ء‬ p Ͻ 0.001. D, Th2 (IL-4 and IL-13)-cytokine mRNA detection by real-time quantitative RT-PCR. The CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were purified from NMLs and ASTs as indicated. The CD3 ϩ CD56 Ϫ CD161 Ϫ T cells were then cocultured with optimal numbers of mature syngeneic DCs and with indicated different numbers of purified syngeneic V␣24 ϩ i NK T cells in the presence of tetanus toxoid as described in Materials and Methods. CD3 ϩ T cells were then harvested using MACS beads. The cytokine by real-time quantitative RT-PCR mRNA detection were conducted as described in Materials and Methods. The numbers listed in the figure were cytokine mRNA copies. The illustrated data were mean values (Ϯ SD) of six performed experiments. ‫,ء‬ p Ͻ 0.001.