Selective Expression of a Stable Cell Surface Molecule on Type 2 but Not Type 1 Helper T Cells

T cell clones used were as follows: Dorris (specific for hen egg lysozyme, H-2^k) and D10 (specific for conalbumin, H-2^k) were obtained from American Type Culture Collection (Atlanta, GA). X4, X9, X12, X17, D2.2, and D2.3 were cloned and maintained as described previously (15). The X and D series of clones are all H-2^d restricted and specific against group A streptococcal M protein. The cell lines were maintained by periodic antigen stimulation with the appropriate irradiated antigen-presenting cells followed by expansion in medium containing IL-2. For the derivation of naive T cell lines, CD4⁺ cells were purified from the spleen of OVA TCR-α/β transgenic mice (D011.10; reference 16) by negative selection as described previously (17). They were cultured with OVA peptide (OVA_323-339) and irradiated BALB/c spleen cells in the presence of IL-12 plus anti–IL-4 (for Th1 line) or IL-4 (for Th2 line) for 6 d (17).

This was performed (9) using RNAimage kit (GenHunter Corp., Brookline, MA). Total RNA was purified from Th1 or Th2 clones and treated with DNase I (GenHunter Corp.) to remove any trace of DNA. Purified RNA (1 μg) was transcribed into cDNA using MMLV reverse transcriptase. The cDNA was amplified by AmpliTaq DNA polymerase (Perkin-Elmer Corp., Norwalk, CT) with 50 pairs of primers (GenHunter). 40 cycles of PCR reaction were performed in the presence of [³⁵S]dATP (New England Nuclear, Boston, MA), and products separated in 6% sequencing gels under denaturing conditions. DNA fragments from Th1s and Th2s were then compared. Bands unique to Th1 or Th2 clones and longer than 150 bp were excised from the dried gel, extracted by boiling in 100 μl of double-distilled H₂0, and precipitated with ethanol in the presence of glycogen. DNA was further amplified with the original set of primers used to generate the particular band by the same thermal cycle conditions. PCR products were electrophoresed in 1.5% agarose ethidium bromide gel to confirm the band size before cloning into PCRII vector (Invitrogen, Carlsbad, CA) and sequenced using M13R and M13(-20)F primers.

This was used to screen and confirm the differentially expressed mRNA (18) using a ReversePrime Kit (GenHunter Corp.). Denatured PCR products (100 ng) were dot blotted onto duplicate nylon membranes with the slot-blot filtration manifolds (Hoefer, San Francisco, CA), using murine IL-5 and β-actin cDNA as controls for specificity and equal loading, respectively. Membranes were UV cross-linked and hybridized with equal amounts of Th1 or Th2 cDNA probes (10⁶ cpm). The probes were prepared from 10–50 μg of each of Th1 or Th2 RNA by reverse transcription using 1,000 units of MMLV reverse transcriptase in the presence of α-[³²P]dCTP and T20 primer. The duplicate membranes were hybridized with Th1 or Th2 probes in 6 × SSC, 5x Denhardt's solution, 0.5% SDS, and 100 μg/ml of denatured salmon sperm DNA at 60°C overnight. The membranes were then washed, exposed for 4–8 h to a x-ray film, and the hybridization patterns compared.

Total RNA was isolated from cells by RNAZolB and was separated (20 μg/lane) in 1% formaldehyde agarose gels, blotted to nylon membrane, and hybridized with the 3′ terminal cDNA fragment of ST2L, and labeled with α-[³²P]dATP. The cDNA fragment was cloned from D10 cells by differential display PCR (DD-PCR) using the primer pair: AAGCTTTTTTTTTTTA and AAGCTTCAGCAGC (designated H-T11A and H-AP34, respectively; GenHunter Corp.).

Primers for amplifying ST2L were: sense, ACTTTGTTCACCACACTCTGC; antisense, AACAGATGCCGTCTTGGAGGC (product of ∼450 bp). Primers for β-actin were: sense, GTGGGCCGCTCTAGGCACCAA; antisense, CTCTTTGATGTCACGCACGATTTC (product of ∼540 bp). PCR products were separated in 1% agarose gels. Membranes were hybridized with internal probes (ST2L, ATTGAAATGGAGCCTCTGGGTGAGGCAAGC; β-actin, CAGAGCAAGAGAGGTATCCTGACCC) labeled with γ-[³²P]dATP using T4 polynucleotide kinase.

A rabbit was immunized with the peptide corresponding to residues 267–287 of ST2L. The peptide was selected to have no homology with IL-1 receptor, spanning the hydrophilic region and within the extracellular domain of ST2L. The rabbit was immunized subcutaneously with 100 μg of the peptide conjugated to KLH emulsified in complete Freund's adjuvant, and boosted sequentially with the peptide-KLH conjugate (100 μg) in incomplete Freund's adjuvant and in saline. Specificity of the antibody was tested by Western blot against recombinant ST2L and by ELISA with peptide-bound 96-well plates. Total IgG was purified from the immune serum and preimmunized serum by protein A column or by ammonium sulphate precipitation. Purity of the IgG was >95% on SDS-PAGE Coomassie blue staining.

Cells were incubated with anti-ST2L antibody (40 μg/ml) followed by FITC-conjugated donkey anti–rabbit IgG (SAPU, Paisley, Scotland). All reactions and washing were carried out at 4°C in PBS containing FCS (5%) and NaN₃ (0.1%), and all antibodies were centrifuged (11,000 g for 5 min) immediately before use. For three-color staining, cells were suspended at 10⁵–10⁶/ml and stimulated with PMA (50 ng/ml, Calbiochem Corp., La Jolla, CA) plus ionomycin (500 ng/ml; Calbiochem Corp.) for 4 h. Brefeldin A (10 μg/ml, Sigma Chemical Co., St. Louis, MO) was added during the last 2 h. Cells were then incubated with anti-ST2L antibody or control IgG followed by biotinylated goat anti–rabbit IgG (Dako Corp.) and then developed with PerCP-streptavidin (Becton Dickinson, Mountain View, CA). The cells were then fixed with paraformaldehyde (2%), and, after washing, permeabilized at room temperature with 0.5% saponin (Sigma Chemical Co.) in PBS and FCS (5%), and stained with FITC-conjugated anti–murine IFN-γ (IgG2a; PharMingen, San Diego, CA) and PE-conjugated anti– murine IL-4 (IgG2a; PharMingen), or FITC- and PE-conjugated isotype control antibodies (Becton Dickinson). Samples were analyzed on a FACScan^® flow cytometer.

Cloned T cell lines (10⁶ cells in 0.2 ml of culture medium) were incubated for 30 min on ice with various concentrations of anti-ST2L antibody or normal rabbit IgG (clarified by centrifugation). The cells were washed once and then incubated with guinea pig complement (1/10 dilution; Harlan Sera-lab, Oxon, UK) for 45 min at 37°C. Cell viability was estimated microscopically (counting >500 cells) by staining with acridine orange and ethidium bromide.

BALB/c mice (female, 6–8-wk-old; Harlan Olac, Bicester, UK) were infected in the right hind footpad with 5 × 10⁵ stationary phase promastigotes of L. major (LV39) and injected intraperitoneally with 250 μg of rabbit anti-ST2L antibody or normal rabbit Ig. The antibody was administered for the next 2 d, and then every 3–4 d. Lesion development was followed by measuring the footpad swelling (difference between the infected right and uninfected left hind footpad). The maintenance of parasites, infection, and measurement of disease progression were as described previously (19). At the end of the experiment, mice were killed by cervical dislocation. Footpad was removed to assay for parasite load by limiting dilution (20). Draining lymph node cells were harvested and cultured (10⁶/ml) in vitro with frozen and thawed parasite antigen (10⁵–10⁷ parasite/ml equivalent). Culture supernatant was harvested at 48 h and assayed for cytokines by ELISA using paired antibodies (PharMingen).

Male DBA/1 mice (6–8-wk-old, Harlan Olac) were injected intradermally with 200 μg of bovine type II collagen (Sigma Chemical Co.) in Freund's complete adjuvant (day −21). Collagen (200 μg in PBS) was given again on day 0 by intraperitoneal injection. Daily injections (intraperitoneal, 200 μg/mouse) of anti-ST2L antibody or normal rabbit IgG were administered from day 1 for 5 d. Mice were monitored daily for signs of arthritis for which severity scores were derived as follows: 0, normal; 1, erythema; 2, erythema and swelling; 3, extension/loss of function; total score = sum of four limbs. Paw thickness was measured with a dial-calliper (Kroeplin, Munich, Germany). Quantification of arthritis was performed as previously described (21, 22). In some experiments, mice were killed 48 h after antibody treatment and spleen cells collected and cultured with graded concentrations of collagen for 48 h in vitro. Supernatants were collected for cytokine assays by ELISA using paired antibodies. T cell proliferation was analyzed by [³H]thymidine uptake.

Total RNA from a panel of four different cloned Th1 and Th2 lines were compared by DD-PCR. Using 50 primer pairs, >50 distinct bands were detected (Fig. 1,a), of which 10 fragments have so far been confirmed to be differentially expressed between Th1s and Th2s using reverse Northern blot (data not shown) and Northern blot analysis. Nucleotide sequence of one band (Fig. 1,b) has complete homology with a gene originally designated as ST2L (10). Northern blot analysis, using the cloned ST2L fragment as cDNA probe, demonstrated the constitutive expression of ST2L in all four of the Th2, but not in Th1, lines (Fig. 1 c). Using the more sensitive reverse transcription PCR Southern blot, we again failed to detect ST2L message in a panel of Th1 lines. In contrast, the message was readily demonstrated in four Th2 lines.

We next investigated the regulation and expression of ST2L in Ths by various immunological stimuli. In contrast to Th2s that expressed ST2L constitutively and stably, Th1s did not express ST2L even when cultured with Con A, IL-1β, IL-2, IL-4, IFN-γ, or with anti–IFN-γ or anti– IL-4 antibodies. After incubation with IL-5 or anti-CD3 antibody, Th1s expressed a weak ST2L message transiently at 3 and 6 h after activation, respectively (Fig. 2,a); but this disappeared completely by 12 h (data not shown). The expression of ST2L in Th2s was not affected by these stimuli (Fig. 2,a). ST2L is a delayed, early proliferative response gene in BALB/c-3T3 cells induced by serum factors (10– 13). We therefore examined the regulation of ST2L expression in T cells by serum. Cells were rested for 48 h in serum-free medium and then cultured in medium containing 20% FCS. The expression of ST2L in Th2s was diminished when cultured in serum-free medium, but restored soon after the addition of serum. In contrast, no ST2L message was detectable in Th1s before or 6 h after addition of serum (Fig. 2 b).

To study the location and function of the molecule, a polyclonal rabbit antibody corresponding to the hydrophilic region of the extracellular domain of ST2L was raised. Th2 clones, but not Th1 clones, were positively labeled with this antibody. The staining was clearly detected on the cell surface (Fig. 3,a). Single cell three-color flow cytometric analysis demonstrated that ST2L coexpressed with IL-4, but not with IFN-γ in cloned T cell lines (Fig. 3,b). We then determined whether this staining pattern also occurs in cells derived from naive T cells. Splenic CD4⁺ T cells from the OVA TCR-α/β transgenic mice (D011.10) were driven to the Th1 or the Th2 lineage in vitro in a 6-d culture and stimulated with PMA and ionomycin for 4 h (17). Staining for ST2L and intracellular IFN-γ and IL-4 demonstrated that the Th1 line expressed no ST2L or IL-4, but showed intracellular staining for IFN-γ. In contrast, the Th2 line expressed strong ST2L and IL-4 staining, but produced no detectable IFN-γ (Fig. 3 c).

We next examined the effect of anti-ST2L antibody on Th1 and Th2 clones in vitro. A panel of Th1 and Th2 clones were incubated in vitro with graded concentrations of anti-ST2L or normal rabbit IgG, and then with or without complement. In the presence of complement, anti-ST2L antibody strongly lysed the Th2 clones, but had only a modest effect on Th1 clones (Table 1).

We next investigated the effect of anti-ST2L antibody in vivo. BALB/c mice are genetically highly susceptible to infection by the protozoan parasite, L. major. The disease progression can be modified and resistance induced by a variety of immunological manipulations (for review see reference 23). Disease progression and resistance are associated with the preferential development of Th2s or Th1s, respectively (for review see reference 24). Fig. 4 demonstrates that BALB/c mice infected with L. major and treated with anti-ST2L antibody developed significantly smaller lesions (Fig. 4,a) and three orders of magnitude less parasite load (Fig. 4,b) compared with mice treated with control rabbit IgG. Draining lymph node cells from the anti-ST2L– treated mice produced significantly more IFN-γ, but less IL-4 and IL-5 compared with cells from control IgG-treated mice (Fig. 4 c). IL-12 levels were not affected.

Collagen-induced arthritis (CIA) in susceptible DBA/1 mice is a predominantly Th1-mediated disease (25–27), the immunopathogenesis of which closely resembles clinical rheumatoid arthritis (28, 29). We therefore examined whether administration of anti-ST2L antibody could modify CIA. Mice injected intradermally with type II collagen in Freund's complete adjuvant developed arthritis when challenged intraperitoneally 21 d later with collagen. Incidence and severity of disease development were markedly enhanced in mice that received daily intraperitoneal injections of anti-ST2L antibody from day 22, compared with controls that received normal IgG (Fig. 5, a–c). On stimulation with collagen in vitro, spleen cells from antibody-treated mice also produced significantly stronger proliferative response and more IFN-γ and IL-6 compared with cells from control mice (Fig. 5 d).

Data presented here demonstrate that ST2L is a stable cell surface marker expressed strongly on Th2s, but not Th1s. Furthermore, antibody against ST2L selectively lysed Th2s in vitro and enhanced Th1 response in vivo. Thus, anti-ST2L antibody increased resistance to L. major infection in the highly susceptible BALB/c mice. It also exacerbated CIA in DBA/1 mice. These effects were accompanied by enhanced IFN-γ production and diminished IL-4/ IL-5 synthesis. It is likely that the antibody exerted its effect through direct depletion of Th2s in vivo and hence the regulation of Th1s by Th2s (1–3), such that the balance is shifted towards Th1s. This is consistent with the observation that treatment with anti-ST2L antibody in the L. major–infected mice did not significantly alter the production of IL-12, a cytokine principally secreted by macrophages and a major mediator of Th1 differentiation (30). Such a mechanism would also explain the immediate effect of the antibody seen in the CIA model (Fig. 5).

ST2L is not detectable on B cells, a macrophage cell line (J774), or a CD8⁺ cell line (CTLL) (data not shown). The developmental stage of Th2s at which ST2L expression occurs is currently unclear. However, the expression of ST2L is clearly associated with the production of IL-4 and IL-5. It has been reported that the transcription factor GATA-3 is necessary and sufficient for Th2 cytokine gene expression in CD4⁺ T cells (31). Furthermore, the promoter region of ST2L contains the GATA consensus binding motif (32). It is therefore possible that GATA-3 influences Th2 functions by directing the transcription of ST2L. The ontogeny of Th2s in relation to ST2L and the sequence of events leading to the induction of ST2L and Th2 cytokine production remain to be explored.

It is important to note that a soluble form (ST2, mol wt: 37-kD) without the trans-membrane domain (10–13, 33) is also transcribed from the same gene by differential splicing, determined by the use of alternate promoters (13, 32). We have found that all Th2s expressing ST2L gene also express ST2 message (data not shown). The receptor-like nature of ST2L suggests the existence of a ligand that is yet to be identified. It may be that the membrane bound form (ST2L) and the soluble form (ST2) act as regulators of Th2 functions. Furthermore, they should serve as important biologically relevant tools for investigating gene regulation for the expression of a trans-membrane versus a secreted form of a protein.

The in vivo effect of anti-ST2L antibody in mice suggests that selective modulation of ST2L expression could have considerable therapeutic potential. The close association of ST2L expression with IL-4/IL-5 production suggests that the membrane-bound ST2L molecule could be a potentially important therapeutic target for allergic responses such as asthma. ST2L is present in humans (34) and its relevance in human T cell functions is currently being addressed.

In conclusion, we have identified a stable cell surface marker distinguishing Th2s from Th1s. Furthermore, an antibody against this molecule has a range of biological effects in vivo including infection and inflammation. Thus, this marker will not only facilitate functional analysis of Th1 and Th2 interaction, it may also be a potential target for therapeutic intervention for some of the important diseases.