IL-17A and IL-17F, produced by the Th17 CD4+ T cell lineage, have been linked to a variety of inflammatory and autoimmune conditions. We recently reported that activated human CD4+ T cells produce not only IL-17A and IL-17F homodimers but also an IL-17F/IL-17A heterodimeric cytokine. All three cytokines can induce chemokine secretion from bronchial epithelial cells, albeit with different potencies. In this study, we used small interfering RNA and Abs to IL-17RA and IL-17RC to demonstrate that heterodimeric IL-17F/IL-17A cytokine activity is dependent on the IL-17RA/IL-17RC receptor complex. Interestingly, surface plasmon resonance studies indicate that the three cytokines bind to IL-17RC with comparable affinities, whereas they bind to IL-17RA with different affinities. Thus, we evaluated the effect of the soluble receptors on cytokine activity and we find that soluble receptors exhibit preferential cytokine blockade. IL-17A activity is inhibited by IL-17RA, IL-17F is inhibited by IL-17RC, and a combination of soluble IL-17RA/IL-17RC receptors is required for inhibition of the IL-17F/IL-17A activity. Altogether, these results indicate that human IL-17F/IL-17A cytokine can bind and signal through the same receptor complex as human IL-17F and IL-17A. However, the distinct affinities of the receptor components for IL-17A, IL-17F, and IL-17F/IL-17A heterodimer can be exploited to differentially affect the activity of these cytokines.
Interleukin 17A (IL-17A) is the founding member of the proinflammatory IL-17 cytokine family. This cytokine family consists of six members (A–F), with IL-17A and IL-17F sharing the closest amino acid sequence identity of ∼50%. IL-17A and IL-17F are homodimeric cytokines produced by the Th17 T cell lineage and share similar biological activities, including induction of cytokines and chemokines involved in inflammatory responses (1, 2, 3, 4, 5, 6, 7). IL-17A and IL-17F have been implicated in a variety of autoimmune diseases, such as rheumatoid arthritis, multiple sclerosis, inflammatory bowel disease, asthma, and psoriasis (5, 6, 8, 9, 10, 11, 12, 13, 14, 15, 16).
The crystal structure of IL-17F has been reported (17) and shows that IL-17F forms a disulfide-linked dimer that contains a cysteine knot motif similar to that reported for members of the nerve growth factor and TGF-β superfamilies (18, 19). Given the high degree of amino acid homology between IL-17A and IL-17F and the conservation of the four cysteines that form the knot, it is likely that IL-17A and IL-17F adopt a similar structure. Recently, we have shown that activated human CD4+ T cells not only express IL-17A and IL-17F homodimers, but also an IL-17F/IL-17A heterodimer, and that the conserved cysteines used in the knot formation of IL-17A and IL-17F homodimers are the cysteines involved in the disulfide linkage of the IL-17F/IL-17A heterodimer (20). The murine IL-17F/IL-17A heterodimer has also been shown to be expressed by differentiated Th17 cells (21, 22).
The IL-17 receptor family consists of five members: RA, RB, RC, RD, and RE. IL-17A has been shown to bind IL-17RA (also referred to as IL-17R) with high affinity, and IL-17RA is required for the biological activity of IL-17A (17, 23, 24). Binding of IL-17F to IL-17RA was not detected by surface plasmon resonance; however, weak binding of IL-17F to cells expressing IL-17RA was detected, suggesting that both IL-17A and IL-17F utilize IL-17RA as part of their receptor complex (17). The IL-17RA receptor is a preformed multimeric complex and is thought to undergo a conformational change upon binding to IL-17A or IL-17F (25).
It has been shown that IL-17A cannot bind to T cells, B cells, and myeloid cells that are deficient in IL-17RA (24). Toy et al. showed that human IL-17A or IL-17F could not elicit a response on mouse IL-17RA−/− cells and the activity could only be regained by the cotransfection of human IL-17RA and IL-17RC, demonstrating that both IL-17RA and IL-17RC are necessary for the biological activity of IL-17A and IL-17F (26). More recently, it has been shown that a soluble form of IL-17RC could neutralize the activity of IL-17A and IL-17F on human bronchial epithelial cells stimulated with IL-17 cytokine plus TNF-α (27).
In this report, we investigate the roles of IL-17RA and IL-17RC receptors in mediating the biological activity of the IL-17F/IL-17A heterodimer. We find that the heterodimer, along with IL-17A and IL-17F homodimers, requires IL-17RA and IL-17RC receptors for signaling. Additionally, our results strongly suggest that different biological activities among homodimers and heterodimer of IL-17A and IL-17F cytokines may be attributed to affinity differences among these cytokines for IL-17RA and IL-17RC receptors.
Materials and Methods
Cloning of the human IL-17 receptors
Full-length human IL-17RA and IL-17RC were PCR amplified from unstimulated MG63 cDNA and sequence confirmed. The extracellular portions of human IL-17RA (residues 1–317) and human IL-17RC (residues 1–452) were fused in frame with a GSGSGSG linker and the human IgG1 Fc portion (28). PCR-derived fusion receptors were subcloned into a CMV promoter-driven mammalian expression construct and sequence verified.
Expression and purification of the human IL-17RA.Fc and IL-17RC.Fc receptors
HEK 293 cells were transiently transfected with IL-17RA.Fc or IL-17RC.Fc expressing vectors using TransIT-LT1 (Mirus Bio). Twenty-four hours post transfection, media was removed and replaced with serum-free media. Conditioned media was harvested at 48 h and protein production evaluated by Western blot analysis. Media containing IL-17RA.Fc or IL-17RC.Fc was flowed over a protein A column (Amersham Biosciences), washed with PBS, and protein eluted with 20 mM citric acid, 200 mM NaCl (pH 3). Protein aggregates were removed by size exclusion chromatography (PBS (pH 7.2)) and the resulting preparations dialyzed (PBS (pH 7.2)) and characterized by SDS-PAGE, Western blot analysis, and analytical size-exclusion chromatography.
BJ human foreskin fibroblast cells (American Type Culture Collection) were maintained in DME + 10% FCS, 2 mM glutamine, 1 mM sodium pyruvate, 0.1 mM MEM nonessential amino acids, and 100 U/ml penicillin + 100 μg/ml streptomycin. BJ cells were seeded at 5 × 103
Binding of human IL-17A, IL-17F, and IL-17F/IL-17A to human IL-17RA.Fc and IL-17RC.Fc was determined by indirect sandwich ELISA. ELISA plates (Costar) were coated with goat anti-human IgG-Fc (10 μg/ml, Bethyl Laboratories). Human IL-17RA.Fc (6 ng/ml) or IL-17RC.Fc (30 ng/ml) was added to plates and incubated for 3 h, followed by serial dilutions of biotinylated IL-17A, IL-17F, and IL-17F/IL-17A for 2 h at room temperature. ELISA was developed with poly-HRP strepavidin (Pierce Biotechnology) and tetramethylbenzidine substrate (Kirkegaard & Perry Laboratories).
Twenty-four hours before transfection, BJ cells were seeded at 104 cells/well in 96-well plates. Cells were transfected with siRNA (20 nM) using DharmaFECT 1 transfection reagent (Dharmacon) according to the manufacturer’s instructions. Each transfection contained one of eight siRNAs listed in Table I⇓
2-microglobulin gene expression.
Western blot analysis
HEK 293 cells (1.2 × 105 cells/ml) stably transfected with human IL-17RA or IL-17RC plasmid were seeded in 96-well plates. After 16–24 h the cells were transfected with 20 nM of siRNA and DharmaFECT transfection reagent (Dharmacon) according to the manufacturer’s protocol. At 48 h, cells were washed once with PBS and lysed on ice using M-PER, mammalian protein extraction reagent (Pierce Biotechnology). Extracted protein was then loaded on an SDS-PAGE gel, transferred to a nylon membrane, and the membrane blocked with 5% nonfat dried milk in PBS containing 0.1% Tween 20. IL-17RA or IL-17RC Ab was added at 1/4000 and incubated overnight at 4°C. Membranes were washed three times (PBS with 0.1% Tween 20) and incubated with donkey anti-goat IgG-HRP (1/2000 dilution, Santa Cruz Biotechnology). Proteins were visualized using Western Lightning Western Blot Chemiluminescence Reagent Plus (PerkinElmer).
Binding kinetics of the IL-17 cytokines with receptors
Kinetic analysis was performed using a Biacore 2000 (Biacore). All experiments were performed at 22°C. IL-17RA.Fc or IL-17RC.Fc was immobilized between 1000 and 2000 resonance units (RU) using amine coupling as directed by the manufacturer. Human IL-17F, IL-17A, and IL-17F/IL-17A heterodimer were each diluted into HBST buffer (10 mM HEPES with 0.15 M NaCl, 3.4 mM EDTA, and 0.005% surfactant P20) at concentrations of 400, 100, 50, 25, 12.5, 6.25, 3.12, and 1.56 nM. Each sample was injected at 50 μl/min for 3 min followed by 10-min dissociation. The surface was regenerated by a 30-s injection of a solution of 0.549 M MgCl2, 0.138 M potassium thiocyanate, 0.276 M urea, and 0.549 M guanidine-HCl followed by two consecutive 60-s HBST buffer injections. Data were double referenced (29) using Scrubber2 software (BioLogic Software), then fit to a 1:1 binding model using Biacore evaluation software version 3.2.
Functional activity of IL-17A, IL-17F, and IL-17F/IL-17A on BJ cells
We have previously reported that IL-17F, IL-17A, and IL-17F/IL-17A induce the secretion of GRO-α from the BEAS-2B human bronchial epithelial cell line (20). We have extended this observation to BJ, a primary human foreskin fibroblast cell. ELISA analysis of conditioned medium obtained from BJ cells cultured with purified IL-17F, IL-17A, or the IL-17F/IL-17A heterodimer showed that all three cytokines can induce GRO-α, IL-6, and IL-8 secretion, albeit with different potencies (Fig. 1⇓A). We have chosen to evaluate functional activity of IL-17 cytokines using GRO-α production as a readout due to low levels of GRO-α secretion in the absence of cytokine stimulation and the increased signal-to-background ratio (8- to 12-fold) upon cytokine stimulation. IL-17F was less potent than IL-17A, and the IL-17F/IL-17A heterodimer was found to be more potent than IL-17F but less potent than IL-17A (Fig. 1⇓B). The potency differences are in agreement with data reported previously on human bronchial epithelial cells (20) and on mouse embryonic fibroblast cells (21).
IL-17F, IL-17A, and IL-17F/IL-17A bind to the IL-17RA/IL-17RC receptor complex with different kinetics
Indirect sandwich ELISAs were done to evaluate the binding of IL-17F, IL-17A, and IL-17F/IL-17A to IL-17RC.Fc and IL-17RA.Fc (Fig. 2⇓). We confirmed, as previously reported (30), that IL-17A and IL-17F homodimers can bind independently to IL-17RA or IL-17RC. IL-17F/IL-17A heterodimer can also bind independently to either IL-17RC or IL-17RA (Fig. 2⇓). All three of the IL-17 cytokines bind to IL-17RC.Fc with approximately the same EC50 of 12–17 ng/ml (Fig. 2⇓A). However, the EC50 for the IL-17 cytokines are quite different for binding to IL-17RA.Fc. The lowest EC50 occurs between IL-17A and IL-17RA.Fc with a value of 19 ng/ml, whereas IL-17F binds weakly to IL-17RA.Fc with an EC50 > 2000 ng/ml. The IL-17F/IL-17A heterodimer has an EC50 of 329 ng/ml, which is 10-fold higher than the binding of IL-17A but ∼10-fold lower than the IL-17F, as shown in Fig. 2⇓B.
To better characterize the binding kinetics of human IL-17F, IL-17A, and IL-17F/IL-17A to each receptor component, the association and dissociation rate constants were measured by surface plasmon resonance. The calculated dissociation constants are shown in Table II⇓. Each IL-17 cytokine tested binds IL-17RA.Fc with a different Kd, which is reflected by differences in the on and off rates. IL-17A binds with the fastest on rate (1.4 × 105 M/s) and slowest off rate (2.9 × 10−4/s), which result in the lowest Kd of 2.2 nM. IL-17F has about a 10-fold slower on rate (9.4 × 103 M/s) and 10-fold slower off rate (1.6 × 10−3/s), which results in about a 100-fold higher Kd for IL-17RA.Fc (170 nM) compared with IL-17A. Interestingly, the IL-17F/IL-17A heterodimer has an on rate (4.3 × 104 M/s) that is intermediate between IL-17A and IL-17F whereas the off rate (1.0 × 10−3/s) is the same as IL-17F. Overall, the Kd of the IL-17F/IL-17A heterodimer (26 nM) is intermediate between IL-17A and IL-17F. In contrast to their different binding kinetics for IL-17RA.Fc, IL-17F, IL-17A, and IL-17F/IL-17A, each bind IL-17RC.Fc with similar on and off rates and hence also have similar Kd values of 10–20 nM. Altogether, these results indicate that IL-17A binds better to IL-17RA than IL-17RC, IL-17F binds to IL-17RC with ∼10-fold higher affinity than to IL-17RA, while IL-17F/IL-17A binds with similar affinity to both receptors.
IL-17RA and IL-17RC siRNA decrease IL-17F, IL-17A, and IL-17F/IL-17A activity
We used siRNA to knock down the expression of the IL-17RA and IL-17RC receptors to investigate the requirement of these two receptors for the functional activity of IL-17F, IL-17A, and IL-17F/IL-17A. BJ cells were transfected with various IL-17RA (RA-1 to RA-4) or IL-17RC (RC-1 to RC-4) siRNAs, and the level of IL-17RA and IL-17RC transcript expression was determined by quantitative PCR (Fig. 3⇓A). Additionally, the effect of siRNA on IL-17RA and IL-17RC protein expression in HEK 293 cells overexpressing IL-17RA or IL-17RC was determined by Western blot analysis (Fig. 3⇓B). Based on the specific inhibition of both mRNA transcripts and protein, two siRNAs per receptor were selected to evaluate IL-17RA (RA-3 and RA-4) and IL-17RC (RC-2 and RC-4) activity in functional assays. IL-17F, IL-17A, and IL-17F/IL-17A were added at three different concentrations to BJ cells transfected with IL-17RA, IL-17RC, or control (NTC1) siRNAs (Fig. 4⇓). IL-17RA and IL-17RC siRNAs decreased GRO-α production in response to IL-17A, IL-17F, or IL-17F/IL-17A at all three concentrations tested relative to the siRNA control. These results suggest that the IL-17F/IL-17A heterodimer activity depends on both IL-17RA and IL-17RC and thus, this cytokine shares a common receptor complex with IL-17A and IL-17F.
Anti-IL-17RA and anti-IL-17RC decrease the functional activity of IL-17F, IL-17A, and IL-17F/IL-17A
To further delineate the role of IL-17RA and IL-17RC, we evaluated whether Abs to IL-17RA and IL-17RC could block IL-17F/IL-17A functional activity. It has been previously shown that Abs to IL-17RA and IL-17RC can block the activity of IL-17F and IL-17A (30). BJ cells were stimulated with suboptimal concentrations of cytokines in the presence or absence of Abs to IL-17RA and IL-17RC. As shown in Fig. 5⇓A, IL-17A-, IL-17F/IL-17A-, and IL-17F-dependent induction of GRO-α was significantly decreased when the cells were treated with anti-human IL-17RA or anti-human IL-17RC Abs. However, anti-IL-17RC Ab had a more profound effect on the activity of IL-17F than on the activities of IL-17A and IL-17F/IL-17A heterodimer. These results, in conjunction with the siRNA knockdown experiments, strongly support the notion of a common receptor complex for IL-17A, IL-17F, and IL-17F/IL-17A.
Preferential inhibition of homodimers and heterodimer of IL-17A and IL-17F cytokines by soluble forms of IL-17RA and IL-17RC
Our surface plasmon resonance results demonstrated different binding affinities between IL-17 cytokines and IL-17RA and IL-17RC receptor components (Table II⇑), so we investigated whether the soluble form of the receptors would affect differentially the functional activity of IL-17A, IL-17F, and IL-17F/IL-17A. Soluble receptor forms were generated by in-frame fusing of the extracellular portions of IL-17RA or IL-17RC to human IgG1 Fc portions, and protein production was done in HEK 293 cells. The ability of the soluble receptors to inhibit cytokine activity was determined at a 50-fold molar excess of receptor to cytokine. As shown in Fig. 5⇑B, IL-17RA.Fc and IL-17RC.Fc significantly neutralize the activity of IL-17A or IL-17F, respectively, at >50% inhibition. Neither IL-17RA.Fc nor IL-17RC.Fc alone could significantly inhibit the activity of the IL-17F/IL-17A heterodimer. Interestingly, a combination of soluble IL-17RA and IL-17RC neutralized ∼50% of the activity of IL-17F/IL-17A heterodimer. These findings suggest that a combination of both receptors is required to neutralize the activity of the IL-17F/IL-17A heterodimer whereas only the higher binding affinity receptor, either IL-17RA or IL-17RC, is required for neutralizing the activity of IL-17A or IL-17F homodimers, respectively.
In this study, we show that human IL-17F/IL-17A heterodimer, the recently identified member of the IL-17 cytokine family, utilizes the same receptor complex as the IL-17F and IL-17A cytokines. Using various experimental approaches, including surface plasmon resonance and siRNA gene knockdown, we characterized the physical and functional interactions of IL-17F/IL-17A, IL-17F, and IL-17A with the IL-17RA and IL-17RC receptors and propose that all three cytokines require both receptors for their biological activity.
Toy et al. demonstrated that human IL-17A or IL-17F could not induce CXCL1 expression in IL-17RA−/− fibroblast cells and that transfection of human IL-17RA did not rescue the expression of CXCL1 as was seen in wild-type cells (26). When the cells were cotransfected with both human IL-17RA and IL-17RC and treated with either human IL-17A or IL-17F, production of CXCL1 was restored, suggesting that a heterodimeric IL-17RA/IL-17RC receptor was required for signaling (26). Recently, it has been reported that IL-17RA is also required for murine IL-17F/IL-17A signaling (21). We confirm and extend these findings by assessing the function of human IL-17F/IL-17A heterodimer in more physiological primary human foreskin fibroblast cells. We find that Ab blockade or siRNA knockdown of either receptor component leads to loss of both homodimeric and heterodimeric cytokine activity. These data are consistent with a model in which both receptor chains are required to form a functional receptor complex. Although IL-17A and IL-17F homodimers can bind independently to IL-17RA and IL-17RC (Table II⇑), both chains are required for IL-17A and IL-17F biological function.
In our study, direct binding measurements with purified IL-17 cytokines and receptors indicate that while all three IL-17 cytokines bind with similar affinities to IL-17RC, they bind very differently to IL-17RA. The order of binding to the receptors and the binding affinities for receptor chains may determine the molecular mechanisms that differentiate the biological functions of these cytokines. Based on the binding kinetics, we propose that IL-17F is likely to associate with IL-17RC and not IL-17RA, unless IL-17RA is in great abundance. On the other hand, IL-17A would be predicted to preferentially bind IL-17RA complexes. The IL-17F/IL-17A heterodimer is likely to interact equally with either receptor chain complex. The pattern of binding affinities that we demonstrate for IL-17A and IL-17F is the same as in previous reports (17, 27), but the dissociation constants differ. The difference in relation to IL-17RA may be attributed to the monomeric form of the IL-17RA protein used in previous experiments (17) vs the dimeric IL-17RA receptor Ig Fc fusion construct used in the present study. Consistent with the higher affinity of IL-17A compared with IL-17F for IL-17RA, we show that the IL-17RA.Fc fusion protein inhibits cellular responses mediated by IL-17A better than those mediated by either IL-17F or IL-17F/IL-17A heterodimer (Fig. 5⇑B and Ref. 30). Both soluble receptors, IL-17RA.Fc and IL-17RC.Fc, which have approximately the same binding affinity for the IL-17F/IL-17A heterodimer, are required to inhibit the activity of IL-17F/IL-17A. The affinities we report for IL-17A and IL-17F binding to IL-17RC are weaker than a previous report even though both studies used the same spliced form of IL-17RC and both used a receptor Ig Fc reagent (27). However, these studies differed in how the receptor was immobilized on the sensor chip surface, which, in turn, may affect binding affinity of the ligand (31).
Biological activity of the IL-17 family may differ among cell types as a result of IL-17 receptor composition (homodimeric and heterodimeric forms). For instance, we observed that soluble IL-17RC.Fc, when used at 50-fold molar excess, can neutralize the biological activity of IL-17F but not IL-17A on primary human foreskin fibroblast cells even though they have a similar binding affinity for the IL-17RC (Fig. 5⇑B). On human bronchial epithelial cells, Kuestner et al. (27) have reported neutralization of IL-17A activity (in combination with TNF-α) by IL-17RC used at a 217-fold molar excess. These disparate results may be partially attributed to differences in receptor composition among cell lines tested (primary human foreskin fibroblasts vs bronchial epithelial cells), cytokine activity examined (IL-17A vs IL-17A plus TNF-α), and/or molar excess of receptor to cytokine employed (50 vs 217 molar excess). Thus, further characterization of receptors in these cell lines is warranted.
IL-17A and IL-17F share 50% identity at the amino acid level. The crystal structure of IL-17F contains a cysteine knot fold characterized by two sets of paired β strands with disulfide bonds connecting the second and fourth strand (17). Conservation of the disulfide linkage in both the IL-17A homodimer and the IL-17F/IL-17A heterodimer suggests that these cytokines have similar structures (20). Although reports have characterized the TGF-β superfamilies as containing a cysteine knot motif, the lack of the third disulfide bond is unique to IL-17F. The crystal structure also showed that IL-17F contains a large cavity (one per monomer) at the cytokine dimer interface (17), as was seen with nerve growth factor (19). These cavities have been proposed as likely sites of cytokine/receptor interaction, based on the binding of nerve growth factor to its high affinity receptor, TrkA (17). Recently, it has been shown that IL-17RA is a preformed multimeric complex and that receptor association is reduced in the presence of IL-17A or IL-17F, suggesting a conformational change of the receptor due to ligand binding (25). Furthermore, it has been proposed that ligand binding alters the conformation of IL-17RA to facilitate a functional, heterotypic interaction with IL-17RC (26). It is tempting to speculate that one of the receptors binds to the large cavity present in the homodimer or heterodimer interface, leading to a conformational change and recruitment of the second receptor.
The recent emergence of the effector Th17 T cell lineage has reshaped the Th1/Th2 landscape and presents a new arm of adaptive immunity that provides explanations to some aspects of immune regulation and immune pathogenesis. While considerable data related to the expression of IL-17A and IL-17F cytokines have been generated recently, detailed functional studies of the IL-17 receptor family have not yet been reported. Currently, it is not known if IL-17F, IL-17A, and IL-17F/IL-17A have redundant roles in certain disease settings. Along these lines, studies have been shown that IL-17A-deficient mice are resistant to the development of collagen-induced arthritis, suggesting that IL-17F cannot compensate for IL-17A in the pathogenesis of collagen-induced arthritis (15). By contrast, intranasal administration of either IL-17A or IL-17F adenovirus leads to pulmonary neutrophilia and inflammatory gene expression in the lungs (32).
Based on our finding of a shared use of receptors by IL-17A, IL-17F, and IL-17F/IL-17A, an interesting question arises from our studies as to whether IL-17F/IL-17A contributes partially to the activity observed for IL-17A in vitro and in vivo. Tissue expression comparing IL-17F/IL-17A heterodimer, IL-17F, and IL-17A might define the role of the heterodimer. In conclusion, we have shown that IL-17F/IL-17A, IL-17F, and IL-17A share a common receptor complex. However, due to binding affinity differences between the IL-17 cytokines and soluble receptors, each receptor component may be exploited to target either IL-17A, IL-17F, or IL-17F/IL-17A cytokine. Elucidating the complexity related to regulation of the expression of various IL-17A–F family members and the biologic significance of their interactions with their cognate receptor(s) will enhance their potential as therapeutic targets in inflammation and autoimmunity.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Address correspondence and reprint requests to Dr. Jill F. Wright, Department of Inflammation, Wyeth Research, 200 Cambridge Park Drive, Cambridge, MA 02140. E-mail address:
↵2 F.B. and B.L. contributed equally to this work.
↵3 Current address: Acceleron Pharma, 24 Emily Street, Cambridge, MA 02139.
↵4 Current address: Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064.
↵5 Current address: Washington University School of Medicine, Division of Oncology, 660 South Euclid Avenue, Campus Box 8007, St. Louis, MO 63110.
↵6 Abbreviations used in this paper: siRNA, small interfering RNA; GRO-α, growth-related oncogene-alpha.
- Received August 23, 2007.
- Accepted June 2, 2008.
- Copyright © 2008 by The American Association of Immunologists