Extracellular mRNA Induces Dendritic Cell Activation by Stimulating ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 277, No. 15, Issue of April 12, pp. 12689 –12696, 2002 Printed in U.S.A.

Extracellular mRNA Induces Dendritic Cell Activation by Stimulating Tumor Necrosis Factor-␣ Secretion and Signaling through a Nucleotide Receptor* Received for publication, November 8, 2001, and in revised form, January 9, 2002 Published, JBC Papers in Press, January 30, 2002, DOI 10.1074/jbc.M110729200

Houping Ni‡, John Capodici‡, Georgetta Cannon‡, Didier Communi§, Jean-Marie Boeynaems§, Katalin Kariko´¶, and Drew Weissman‡储 From the Divisions of ‡Infectious Diseases and ¶Neurosurgery, University of Pennsylvania, Philadelphia, Pennsylvania 19104 and the §Institute of Interdisciplinary Research and Department of Medical Chemistry, School of Medicine, Erasme Hospital, Universite´ Libre de Bruxelles, Brussels 1070, Belgium

We previously demonstrated that dendritic cell (DC) pulsing with antigen-encoded mRNA resulted in the loading of both major histocompatibility complex class I and II antigen presentation pathways and the delivery of an activation signal. Coculture of mRNA-pulsed DC with T cells led to the induction of a potent primary immune response. DC, in addition to recognizing foreign antigens through pattern recognition receptors, also must respond to altered self, transformed, or intracellularly infected cells. This occurs through cell surface receptors that recognize products of inflammation and cell death. In this report, we characterize two signaling pathways utilized by extracellular mRNA to activate DC. In addition, a novel ligand, poly(A), is identified that mediates signaling through a receptor that can be inhibited by pertussis toxin and suramin and can be desensitized by ATP and ADP, suggesting a P2Y type nucleotide receptor. The role of this signaling activity in vaccine design and the potential effect of mRNA released by damaged cells in the induction of immune responsiveness is discussed.

Dendritic cells (DC)1 are the sentinel cells of the adaptive immune system and function in the induction of primary and memory T cell immune responses (1, 2). They mainly populate tissues that interface with the environment and acquire antigens through high level constitutive macropinocytosis and endocytosis (3). Immature tissue DC sense invading organisms by recognizing evolutionarily conserved structures, known as * This work was supported by NHLBI, National Institutes of Health (NIH) Grant R01-HL-62060-1 and by NIAID, NIH Grant R21-AI-4531801A1. 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. 储 To whom correspondence should be addressed: Division of Infectious Diseases, University of Pennsylvania, 522B Johnson Pavilion, Philadelphia, PA 19104. Tel.: 215-614-0291; Fax: 215-349-5111; E-mail: [email protected]. 1 The abbreviations used are: DC, dendritic cell(s); Treg, T regulatory; LPS, lipopolysaccharide; dsRNA, double-stranded RNA, CD40L, CD40 ligand; GPCR, G-protein-coupled receptor; ssDNA, single-stranded DNA; PE, phycoerythrin; FITC, fluorescein isothiocyanate; TLR, Tolllike receptor; PBMC, peripheral blood mononuclear cells; CPK, creatine phosphokinase; CP, creatine phosphate; MHC, major histocompatibility complex; IL-12, interleukin-12; IFN-␣, interferon ␣; TNF-␣, tumor necrosis factor ␣; FCS, fetal calf serum; HIV, human immunodeficiency virus; PGE2, prostaglandin E2; RANTES, regulated on activation normal T cell expressed and secreted; ELISA, enzyme-linked immunosorbent assay; 2MeSATP, 2-(methylthio)adenosine 5⬘-triphosphate; 2MeSADP, 2-(methylthio)adenosine 5⬘-diphosphate. This paper is available on line at http://www.jbc.org

pathogen-associated molecular patterns, contained within microbial lipids, carbohydrates, and nucleic acid. This pattern recognition occurs through a set of germ-line encoded receptors, which are exemplified by the Toll-like receptor (TLR) family (4 – 6). Immature DC, upon receiving an activation signal, undergo phenotypical and functional changes, including: 1) decreased antigen acquisition with a coordinated increase in presentation of MHC-peptide complexes on the cell surface; 2) increased stability of MHC class II-peptide complexes; 3) increased expression of surface molecules that aid and promote T cell activation (7); 4) a changed pattern of release of chemokines and cytokines leading to attraction of T cells, promotion of T cell activation, and direction of their ultimate phenotype (Th0, Th1, Th2, or T regulatory (Treg)) (5, 8, 9); and 5) a shift in the repertoire of chemokine receptor expression that allows and directs DC migration to lymphoid organs (10, 11). Identified categories of DC activators include: 1) conserved constituents of bacteria (lipopolysaccharide (LPS), cell wall lipoproteins, flagellar proteins, and DNA), 2) host cell-derived molecules released during cell injury and death (proinflammatory cytokines and nucleotides), 3) intermediates of viral replication (double-stranded RNA (dsRNA)) (12), and 4) molecules on activated CD4⫹ T cells (CD40 ligand (CD40L)). The signaling by each DC activator leads to the transformation of a DC from antigen acquisition to antigen presentation but produces a DC that differs in the type of immune response it induces (13). DCs treated with TNF-␣, a pleiotropic stimulus of DC activation, and prostaglandin E produce a low level of IL-12 and induce a mixed population of Th0 and Th1 T cells. DCs exposed to dsRNA secrete high levels of IL-12 and IFN-␣ leading to a strong Th1 response. DCs exposed to apoptotic cells or malaria-infected red blood cells do not secrete IL-12 but produce IL-10 (14, 15) leading to the induction of Treg cells (16). Nucleotide receptors are comprised of two families, a metabotropic family (P2Y) that belong to the 7-transmembrane, G-protein coupled receptor (GPCR) superfamily, and a pore-forming, cation-selective family (P2X). Human DCs express mRNA for P2Y1, -2, -4, -6, -11 and P2X1, -2, -4, -5, -7 nucleotide receptors (17). Signaling through both families by selected nucleotides induces different aspects of DC activation. ATP, which can be released by damaged cells and has been demonstrated to synergize with TNF-␣ in the activation of DCs (18, 19), acts through P2Y11 receptor signaling. This signaling pathway leads to the generation of cAMP (19). The P2X7 receptor is important in cytokine secretion in human DC (20) and antigen presentation in murine DC (21). In addition to DC activation, it has also been observed that signaling through nucleotide receptors by ATP leads to DC apoptosis (22) and

12689

12690

RNA-mediated DC Signaling and Activation

aberrant DC activation. DC pretreated with low, non-toxic doses of ATP, produced lower amounts of IL-1␣, IL-1␤, TNF-␣, IL-6, and IL-12 after subsequent stimulation with LPS or CD40L (23). The targeting of DC for antigen delivery in vivo and in vitro represents an important approach in vaccine research. The first step in immune responsiveness induced by vaccines is the delivery of antigen in a form that the DC can acquire, process, and present to CD4⫹ T, CD8⫹ T, and/or B cells. With this delivery of antigen, the vaccine also must deliver an activation signal to the DC. This is often done by the inclusion of adjuvants, such as mycobacterium to complete Freund’s adjuvant, or with components of the antigen delivery system, such as CpG motifs in DNA vaccines (reviewed in Ref. 24). We previously reported that pulsing DC with mRNA encoding antigen led to loading of both CD4⫹ and CD8⫹ T cell antigen presentation pathways, delivery of an activation signal to DC, and the induction of potent antigen-specific T cell activation (25). In this report, we determined the mechanisms whereby mRNA activates DC. EXPERIMENTAL PROCEDURES

In Vitro RNA Transcription—Transcription was performed on Gagand luciferase-encoding plasmid templates linearized downstream from a stretch of dA50 using the T7 message machine kit (Ambion, Austin, TX) as described previously (25). Purification of the transcripts was performed by DNase I digestion followed by LiCl precipitation and 70% EtOH washing. Additional poly(A) tail was added to the transcripts with yeast poly(A) polymerase (Amersham Biosciences, Inc., Piscataway, NJ), and the mRNA was again purified. All Gag- or luciferase-encoding mRNA contained a poly(A) tail unless otherwise noted. Assays for LPS in RNA preparations using the Lumilus Amebocyte Lysate gel clot assay were negative with a sensitivity of 3 pg/ml (University of Pennsylvania, Department of Genetics, Cell Center Service Facility). The quality of each batch of mRNA was tested by agarose gel electrophoresis for degradation. Samples were stored in siliconized tubes at ⫺20 °C until use. Cell Culture—HL60, U937, and 293T cells (ATCC, Rockville, MD) and P2Y11 nucleotide receptor stably transfected Chinese hamster ovary and 1231N1 cells (26) were propagated in Dulbecco’s modified Eagle’s medium supplemented with glutamine (Invitrogen, Rockville, MD) and 10% FCS (HyClone, Ogden, UT). Leukapheresis samples were obtained from HIV-uninfected volunteers through an institutional review board-approved protocol. Peripheral blood mononuclear cells (PBMC) were purified by Ficoll-Hypaque density gradient purification. PBMC were cryopreserved in RPMI 1640 (Invitrogen) with 50% FCS and 10% Me2SO (Sigma Chemical Co., St. Louis, MO). DCs were produced as described previously with minor modifications (25). Monocytes were obtained by the RosetteSep method (Stem Cell Technologies, Vancouver, BC, Canada) from leukapheresis samples (containing red blood cells) as described by the manufacturer. Monocytes were cultured in AIM V serum-free medium (Invitrogen) supplemented with granulocyte macrophage/colony-stimulating factor (50 ng/ml) and IL-4 (100 ng/ml) (R&D Systems, Minneapolis, MN). Fresh medium containing cytokines was added on days 2 and 5. The resulting immature DCs were used between 6 and 9 days after initial culture of monocytes. DCs were treated with TNF-␣ (1 ng/ml) (R&D Systems) ⫹ PGE3 (500 nM) (Cayman Chemical Co., Ann Arbor, MI), LPS (1.0 –10 ␮g/ml) (Sigma), CD40L trimer (a kind gift of Elaine Thomas, Immunex, Seattle, WA), poly(A) and poly(U) single-stranded (ss) RNA, poly(I)䡠poly(C) dsRNA (20 ␮g/ml) (Sigma), Gag- or luciferase-encoding mRNA (0.22 ␮g/50 ␮l) complexed with Lipofectin (Invitrogen) (27) or Lipofectin without nucleic acid. Higher concentrations of LPS were required for DC maturation, because no serum was present in the cultures, thus LPS bound and activated DC in the absence of serum-derived LPSbinding protein. Similar activation profiles were observed for LPS when exogenous serum was added. 293T cells were transiently transfected with a cytomegalovirusdriven P2Y4 expression plasmid (28) complexed to Lipofectin in the same manner as RNA. After 24 h, cells, or control transfected 293T cells, were analyzed for the ability to flux Ca2⫹ in response to poly(A) or UTP (Sigma). Ca2⫹ Signal Fluorescence Determinations—Ca2⫹ flux was determined by measuring fluorescence spectral changes of Fura-2 (Molecular

Probes, Eugene, OR). The excitation wavelengths were 340 and 380 nm, and the measured emission wavelength was 510 nm. All experiments were performed at 37 °C using a Photon Technology International fluorescence spectrophotometer (London, Ontario, Canada) equipped with a magnetic stirrer. Cells were loaded with Fura-2AM for 1 h at 37 °C in RPMI 10% FCS followed by washing. One and one-half million Fura-2 loaded cells in 1 ml of Hanks’ balanced salt solution with Ca2⫹ and Mg2⫹ (Invitrogen) were added to a cuvette and analyzed to establish a baseline followed by addition of stimuli. Data are presented as the ratio of emissions at 510 nm. Certain cells were treated with pertussis toxin (30 ␮g/ml) (Sigma) 5 h before signaling or suramin (30 ␮M) or EGTA (200 ␮M) (Sigma) 10 min before signaling. Cells were stimulated with ATP (1–100 ␮M), 2MeSATP (1–5 ␮M), ADP (5–100 ␮M), UTP (20 –100 ␮M), poly(A) (1– 67 ␮M AMP equivalents), poly(U), poly(C), poly(G), poly(A,C), poly(A,G), poly(A,U) (67 ␮M AMP equivalents) (Sigma), Gagencoding mRNA with or without a poly(A) tail (30 ␮g added to 1.5 ml cells) or RANTES (BD PharMingen, San Jose, CA) (33 ng/ml), and the calcium flux was monitored for up to 600 s at 1-s intervals. In certain experiments, ATP and 2MeSATP (1 mM) were treated for 90 min with creatine phosphokinase (CPK) (20 units/ml) and creatine phosphate (CP) (10 mM) (Sigma) to convert contaminating ADP and 2MeSADP to ATP and 2MeSATP, respectively (29). Desensitization was measured by adding the second or third stimulus when the calcium flux subsided and the baseline was re-established. Analysis of DC Maturation and T Cell Activation by DC—DC were stained for the following markers with directly conjugated antibodies: CD83-phycoerythrin (PE) or fluorescein isothiocyanate (FITC), CD80-PE (Research Diagnostics Inc., Flanders, NJ), HLA-DR-FITC, HLA-A, B,CFITC, CD25-FITC, CCR5-FITC, CXCR4-PE, CD86-Cy-Chrome, CCR6-PE (BD PharMingen, San Diego, CA) and analyzed on a FACSCalibur (BD PharMingen) flow cytometer. Endocytosis/macropinocytosis was measured by incubating DCs with FITC-labeled bovine serum albumin (Sigma) for 1 h at 37 °C followed by washing and analysis. For studies of T cell activation, immature or activated DC were washed to remove the activating agent and cultured with autologous CD4⫹ T cells purified by negative selection (Human CD4⫹ T Cell Enrichment Column, R&D Systems) from cryopreserved PBMC at a ratio of 1 DC to 10 T cells and staphylococcal toxic shock surperantigen (Sigma) (0.01 ng/ml). One hour later, brefeldin A (20 ␮g/ml) (Sigma) was added. Five hours later, cells were stained with CD4-APC (BD PharMingen), permeabilized, and stained with CD69-Peridinin chlorophyll, IFN-␥-FITC, and IL-4-PE (BD PharMingen) as described previously (30) and analyzed on a FACSCalibur flow cytometer, 500,000 events per sample. ELISA Assays for cAMP and Cytokines—3-Isobutyl-1-methylxanthine (5 ␮M)-treated (30 min) DCs were activated for 30 min and lysed for cAMP quantitation. cAMP level was measured by competitive chemiluminescence assay (Applied Biosystems, Foster City, CA) as per manufacturer’s instructions. Supernatants of activated DC were collected at 48 h for cytokine measurement. IL-12 (p70) (BD PharMingen), TNF-␣, IFN-␣, and IL-8 (BIOSOURCE International, Camarillo, CA) were measured in supernatants by sandwich ELISA. Cultures were performed in duplicate and measured in duplicate. Statistical Analysis—Student’s t tests were performed using Microsoft Excel software. RESULTS

mRNA Induces DC Activation—DC activation is induced by classes of stimuli through specific receptors (TLR, TNF family, nucleotide, prostaglandin) and signaling cascades. Although each of these classes of agents induces activation and maturation of DC, the function of the resulting DC regarding the type of immune response induced differs. We previously demonstrated that HIV Gag-encoding mRNA activates DC as measured by expression of the DC maturation marker CD83 and antigen-specific T cell activation (25). To better define the phenotype of mRNA activated DC, we studied the following markers of DC activation: 1) MHC classes I and II and costimulatory (CD80, CD86) molecules necessary for T cell activation, 2) chemokine receptors (CCR5, CCR6, CXCR4) that traffic DC to peripheral tissues and lymphoid organs, 3) functional uptake of extracellular antigen, and 4) cytokines (IL-12, IL-8, IFN-␣, TNF-␣) released by DC that regulate and shape the resulting T cell response. mRNA in comparison to the standard DC activation agents, LPS, TNF-␣ ⫹ PGE3, dsRNA, and CD40L, demon-

RNA-mediated DC Signaling and Activation strated the greatest increase in surface proteins associated with T cell activation, CD80, CD86, and MHC classes I and II, and lesser regulation of chemokine receptor expression (Fig. 1 and Table I). Both Gag- and luciferase-encoding mRNAcontaining poly(A) tails were used, and no difference in activation of DC was observed. Poly(A) minimally increased CD83 and induced a smaller subset of markers of DC activation while poly(U) did not induce markers of DC activation (Fig. 1 and Table I). The cytokines produced by activated DC during interactions with T cells determines, in part, the phenotype of the resulting T cell response. IL-12, IFN-␣, TNF-␣, and IL-8 are differentially produced in response to DC activation stimuli. IL-12 in humans acts on T cells to induce them to produce IFN-␥ (Th1 response). IFN-␣ acts on DCs to down-regulate IL-12 and increase IL-10 production (16). The lack of IL-12 and the production of IL-10 favor Th2 and Treg T cells. IL-8 acts to attract inflammatory cells and potentiates the local immune response (31). DC cytokine production induced by the different forms of activation demonstrated that mRNA led to moderate levels of IL-12 production compared with the very high levels observed with CD40L and the low amounts observed with TNF-␣ ⫹ PGE3 stimulation (Fig. 2 and Table I). Gag encoding mRNA

12691

induced moderate levels of IFN-␣ and TNF-␣, whereas poly(A) did not increase IFN-␣ or TNF-␣ production but induced IL-12 suggesting that poly(A) represented a subset of mRNA activation of DC. Activated DCs are superior to immature DC in inducing T cell activation. A system that measures the ability of DC maturation agents to increase DC-induced T cell activation was employed. Unactivated T cells were cocultured with immature or activated DC in the presence of suboptimal concentrations of superantigen for 6 h and then analyzed for the early activation marker CD69. mRNA mediated activation of DC resulted in the induction of CD69 on T cells at similar levels observed for other potent DC-activating agents (Fig. 3 and Table II). Studies examining the cytokines made by T cells activated by DC demonstrated that immature DC induced both IL-4- and IFN␥-secreting T cells (Fig. 4). This assay measured the percentage of cells induced to make IL-4 or IFN-␥ by DC stimulation and differs from other studies that polyclonally stimulate T cells after coculture with DCs, which determine the potential to produce a cytokine induced by the DC (12, 32). LPS induced predominantly IFN-␥-producing T cells with a small population of IL-4 producing CD4⫹ T cells. Poly(I)䡠poly(C) activation of DC resulted in CD4⫹ T cells that produced almost entirely IFN-␥

FIG. 1. RNA activates DC as measured by the up-regulation of T cell activation molecules and modulation of chemokine receptors. Dotted lines, immature; thin solid lines, LPS (1.0 ␮g/ ml) treated; and heavy solid lines, Gagencoding mRNA pulsed DC were analyzed for expression of CD80, CD86, MHC classes I and II, CXCR4, and CCR5 after 24 h. Data presented are from one DC preparation and are representative of at least five preparations from at least five different donors.

TABLE I Rank order change of DC activation markers by stimuli The changes in DC activation markers induced by each stimuli after 24 – 48 h of treatment of immature DC were quantitated (mean fluorescence by flow cytometry for surface molecule expression and amount in supernatant by ELISA for cytokines) and rank-ordered with 1 being change compared to immature (no treatment). Rank order placement of each activator was determined if a statistically different result was observed in comparing at least five separate experiments between activators. An equal rank was given if no statistically significant difference was observed between activators. DC activators

aCD83 aHLA-ABC aHLA-DR aCD80 aCD86 aCD25 sCCR5 sCCR6 aCXCR4 sEndocytosis aIL-12 aIFN-␣ aTNF-␣ aIL-8

TNF ⫹ PGE3 LPS CD40L Poly(I) 䡠 poly(C) Gag-encoding mRNA Poly(A) Poly(U) None a

ND, not done.

4 5 4 4 3

4 4 4 3 5

5 5 5 4 6

3 3 3 3 4

3 3 3 3 4

3 4 3 3 2

4 4 4 4 3

2 2 2 2 2

4 4 4 4 3

4 4 4 4 4

2 5 6 3 4

1 2 1 4 3

NDa 3 3 2 2

3 3 3 3 3

2 1 1

2 1 1

3 2 1

2 1 1

2 1 1

1 1 1

2 1 1

1 1 1

2 1 1

3 2 1

3 1 1

1 1 1

1 1 1

2 1 1

12692


FIG. 3. DC activated by Gag-encoding mRNA are similar to poly(I)䡠poly(C)- and LPS-activated DC in the extent of expression of the early T cell activation marker CD69 in cocultured T cells. Immature DC treated for 24 h with medium, poly(I)䡠poly(C) (20 ␮g/ml), LPS (1 ␮g/ml), or Gag-encoding mRNA (4.4 ␮g/ml) were washed and cocultured with autologous CD4⫹ T cells and suboptimal concentrations of TSST-1 superantigen (0.01 ng/ml). Intracellular CD69 expression on CD4⫹ T cells was measured 6 h later by flow cytometry. Data presented are from one donor and are representative of two different donors. TABLE II CD69 expression on CD4⫺ T cells was determined after coculture with immature or activated DC CD4⫹ T cells, cocultured with activated DC for 6 h in the presence of superantigen, were stained for CD4 and CD69 and analyzed using flow cytometry. DC activator

CD69⫹ CD4⫹ T cells %

TNF-␣ ⫹ PGE3 LPS CD40L Poly(I) 䡠 poly(C) Gag-encoding mRNA Medium

FIG. 2. DC activation agents stimulate different patterns of cytokine secretion. Immature DC were treated with medium, Lipofectin, Lipofectin-complexed Gag-encoding mRNA, poly(A) and poly(U) ssRNA, LPS, poly(I)䡠poly(C) dsRNA, or CD40L for 48 h. Supernatants from duplicate cultures were analyzed in duplicate for TNF-␣, IFN-␣, IL-12 (p70) content by ELISA. Error bars represent the standard error of the mean of the four samples for each measurement. Data presented are from one DC preparation and are representative of six DC preparations from six different donors.

but consistently had a lower percentage of total T cells making IFN-␥ compared with other stimuli. mRNA-treated DC activated CD4⫹ T cells to produce mainly IFN-␥, similar to poly(I)䡠poly(C) and CD40L (Fig. 4 and data not shown). RNA Induces a Calcium Flux in DC—We observed that poly(A) partially activated DC whereas mRNA fully induced DC maturation and that poly(A) did not whereas mRNA induced TNF-␣ secretion. It was recently demonstrated that ATP or TNF-␣ treatment of DC induced low level expression of the DC maturation marker CD83. The addition of both TNF-␣ and

6.8 5.5 7.0 6.2 5.6 1.5

ATP resulted in a synergistic activation of DC (18). A similar synergy was also observed when TNF-␣ and poly(A) were used to activate DC (data not shown). ATP signals cells through nucleotide receptors that are divided into two families, one G-protein linked (P2Y) and the second through selective cation pore formation (P2X). Poly(A) at similar molar AMP equivalents as ATP stimulated a calcium flux in Fura-2-loaded DC (Fig. 5). Dose-response analysis demonstrated that 1.0 ␮M AMP equivalents of poly(A) could flux calcium in DC. In vitro transcribed mRNA encoding the HIV Gag protein and containing a 50-nucleotide or longer poly(A) tail also fluxed Ca2⫹ in DC, whereas mRNA lacking a poly(A) tail did not (Fig. 5). The P2X receptors flux extracellular calcium through plasma membrane pores, and their signaling is inhibited by removal of extracellular Ca2⫹ with EGTA. RNA signaling of DC was not inhibited by the absence of extracellular calcium, suggesting that it did not signal through P2X receptors (Fig. 5). These data suggested that poly(A) signaled through P2Y receptors present on DC whose ligands were previously identified as nucleotides. P2Y nucleotide and many G-protein coupled 7-transmem-


12693

FIG. 4. mRNA activation of DC results in CD4ⴙ T cells that produce IFN-␥. DC activated with mRNA (4.4 ␮g/ ml), LPS (1 ␮g/ml), or poly(I)䡠poly(C) (20 ␮g/ml) were cocultured with autologous CD4⫹ T cells, TSST-1 superantigen (0.01 ng/ml), and brefeldin A. After 6 h, cells were permeabilized, stained for IFN-␥, IL-4, and CD69, and flow cytometrically analyzed. Presented histograms showing IFN-␥ versus IL-4 were gated on CD69⫹ cells. Although the percentage of cytokine-positive cells appears similar for each condition, as CD69⫹ cells were analyzed under each condition, the number of cytokine-positive CD4⫹ T cells in the immature DC coculture were approximately one quarter of that observed under other conditions (Fig. 3). Data from one subject are representative of two donors.

brane receptors can be desensitized to subsequent signaling. This increase in EC50 required for subsequent signaling can be observed for the same or different ligands that share a receptor (33–35). DCs were sequentially stimulated with ATP, ADP, or UTP; poly(A); and RANTES (Fig. 6). UTP stimulation did not inhibit subsequent signaling by poly(A) or RANTES (Fig. 6). ATP, at low concentrations (1–5 ␮M), desensitized DC to poly(A), with little or no effect on subsequent RANTES signaling (Fig. 6). Higher concentrations of ATP were toxic as previously demonstrated (20) and observed in this assay by a loss of RANTES signaling. Two new P2Y receptors that are coupled to Gi and signal to ADP have recently been identified (29, 36). ADP completely desensitized DC to poly(A) (Fig. 6). ATP is rapidly degraded to ADP by ecto-nucleotidases on the surface of DC (17). To determine whether ATP or ADP derived from degraded ATP was responsible for desensitization, a non-hydrolyzable analog of ATP, 2MeSATP, was used after treatment with CPK and CP to convert contaminating 2MeSADP into 2MeSATP and desensitization was observed (Fig. 6). These data suggest that poly(A) likely signaled through a nucleotide receptor whose ligands included ATP and ADP and may be a member of a new family of ADP receptors more closely related to the UDP-glucose receptor (29, 36). GPCR desensitization is dependent on both the receptor and the ligand, and certain receptors are not desensitized by all of its ligands. An example of this is the lack of desensitization of CCR5 by MIP-1␤ (37). Additional studies were performed to analyze each known P2Y receptor for poly(A) and mRNA signaling. A recent report demonstrated that ATP activated DC by signaling through the P2Y11 receptor. Signaling through this receptor, unlike other P2Y or P2X receptors expressed by DC, resulted in increased cAMP generation (19). To determine if RNA used this receptor, DCs were stimulated by poly(A) and cAMP was measured by competitive ELISA. Poly(A) did not increase cAMP with levels similar to unstimulated DC. ATP,

which signals through P2Y11, increased the cAMP level 30-fold over baseline (Fig. 7). To confirm that RNA did not signal through P2Y11, cell lines (Chinese hamster ovary and 1231N1) stably expressing human P2Y11 (26) were loaded with Fura-2 and analyzed for Ca2⫹ flux. ATP but not poly(A) induced calcium fluxes, confirming that RNA did not signal through the P2Y11 nucleotide receptor. In addition, the observation that poly(A) induced IL-12 clearly distinguished it from ATP-induced DC maturation, which has been demonstrated to inhibit IL-12 secretion (19). We next sought to determine which P2Y nucleotide receptor signaled in response to poly(A) by utilizing the differential sensitivity to specific inhibitors of the members of this family. Pertussis toxin inhibits Gi-proteins and completely blocks betachemokine-mediated signaling through their respective receptors (38). P2Y2, -4, -12, and -13 but not P2Y1, -6, or -11 are sensitive to pertussis toxin-mediated inhibition when ligand concentration is low (39 – 41). Poly(A) and RANTES but not ATP or UTP signaling were blocked by pretreatment of DC with pertussis toxin (Fig. 5 and data not shown). Suramin is a synthetic polysulfonated naphthylurea that blocks P2Y2 but not P2Y4 receptor signaling (40). When poly(A) was added to DCs that were preincubated with suramin, calcium flux was completely blocked, suggesting that P2Y2 was not responsible for RNA signaling (Fig. 5). Confirmatory investigation utilized cell lines (HL60 and U937) that expressed P2Y1 (HL60 only), P2Y2, and P2Y6 receptors and 293T cells transiently transfected with a P2Y4 expression plasmid. Both cell lines and P2Y4-transfected 293T cells fluxed calcium upon stimulation with ATP or UTP, respectively, but none responded to poly(A), suggesting that poly(A) signaled through a nucleotide receptor that could be desensitized with ADP and ATP and whose signaling was blocked by pertussis toxin and suramin. The distribution of the poly(A) signaling activity was limited because neither poly(A) nor mRNA sig-

12694


FIG. 7. Poly(A) does not increase cAMP generation in DC. 3-Isobutyl-1-methylxanthine-treated DC were stimulated with medium, ATP (100 ␮M), poly(A) (67 ␮M AMP equivalents), and poly(U) (67 ␮M AMP equivalents), lysed 30 min later, and analyzed for cAMP content. Samples were performed in duplicate and analyzed in duplicate. Data from one subject’s DC are representative three donors’ DC.

FIG. 6. Poly(A)-induced Ca2ⴙ flux is blocked by a preliminary stimulation with ATP, 2MeSATP, or ADP but not UTP. Fura-2loaded DC were sequentially stimulated with ATP (3 ␮M), UTP (20 ␮M), ADP (5 ␮M), and 2MeSATP (1 ␮M) treated with CPK and CP to convert contaminating 2MeSADP, poly(A) (67 ␮M AMP equivalents), and RANTES (33 ng/ml). The ratio of emissions at 510 nm versus time is presented. A low concentration of ATP was used to avoid toxic effects observed at higher ATP concentrations. Data are representative of three preparations of DC from three donors. Portions of data from two donors are shown.

FIG. 5. Poly(A) induces a Ca2ⴙ flux in DC, which can be inhibited by pertussis toxin (PTX) and suramin. DC were preincubated with medium or pertussis toxin (30 ␮g/ml) for 5 h, loaded with Fura-2, treated with EGTA (200 ␮M) or suramin (30 ␮M) for 10 min, where indicated, and then stimulated with Gag-encoding mRNA with or without (noted with an asterisk) a poly(A) tail (30 ␮g/ml), ATP (100 ␮M), poly(A) (67 ␮M AMP equivalents), poly(U) (67 ␮M AMP equivalents), or RANTES (33 ng/ml) and followed for Fura-2 spectral changes induced by Ca2⫹ release. The ratio of emissions at 510 nm is given versus time. The same scale of ratio of light emission and time was used for each assay shown. Ca2⫹ fluxes presented were obtained from multiple samples of DC from different donors. Three to ten repetitions for each condition were performed.

naled T, B, and monocytic cells derived from PBMC. Our in vitro synthesized mRNA was prepared from nucleotide triphosphates and purified by LiCl precipitation and EtOH washing to remove free nucleotides. The lack of signaling of these RNA preparations on cell lines expressing P2Y1, -2, -4, -6, and -11 suggested that the RNA preparations did not contain contaminating nucleotides that were responsible for Ca2⫹ signaling. In addition, homopolymers other than poly(A) made by polynucleotide phosphorylase catalyzed polymerization of ADP, UDP, CDP, and/or GDP also did not signal DC, suggesting that the receptor likely recognized stretches of A in RNA and not contaminating ADP. The data suggested that poly(A) represented a subset, moderate induction of markers of DC activation and IL-12 secretion but no TNF-␣ or IFN-␣ (Table I), of the DC maturation activity of mRNA. The ability of in vitro transcribed encoding mRNA to induce IFN-␣ and TNF-␣ suggested that part of the mRNA’s DC maturation ability might be mediated by the formation of

regions of dsRNA. This dsRNA would activate DC in a similar manner as poly(I)䡠poly(C), although it likely may not provide the same intensity of activation due to the lower levels of dsRNA content of the mRNA (12). It was also possible that the poly(A) signaling altered the dsRNA DC activation effect. This was supported by the differences in DC maturation markers induced by poly(I)䡠poly(C) and mRNA (Fig. 1 and Table I) and directly tested by demonstrating that adding poly(A) to poly(I)䡠poly(C) stimulation of DC altered the activation markers CD80 and CD86 such that it resembled mRNA activation (Fig. 8). The magnitude of change in the level of the activation markers, although relatively small, was reproduced in three separate experiments. The presented data normalized each experiment to equal untreated mean fluorescence for each activation marker and averaged the results of three experiments. Gag-encoding mRNA lacking a poly(A) tail, in addition to not fluxing Ca2⫹ in DC, induced fewer markers of DC maturation than mRNA containing a poly(A) tail suggesting that the additional signal delivered by poly(A) both increased DC activation and altered the phenotype of the resulting DC. DISCUSSION

In the current report, we characterize a new signaling activity by extracellular mRNA that activates DC and determine the resulting phenotype and functional ability of mRNA-matured DC to activate T cells and compared it to other DC maturation stimuli. mRNA-activated DC expressed MHC classes I and II and coactivation molecules CD80 and CD86 at similar or greater levels than the potent DC activators, LPS, dsRNA, and CD40L. mRNA also induced IL-12, IFN-␣, IL-8, and TNF-␣ production by DC. In vitro transcribed RNA containing a cod-


FIG. 8. The addition of poly(A) to poly(I)䡠poly(C) alters the expression of CD80 and CD86 to levels observed with mRNA activation. DCs were activated with the indicated agents for 24 h followed by staining with directly conjugated specific antibodies and flow cytometrically analyzed. The mean fluorescence for each marker was calculated for each activation agent. P values were calculated by normalizing mean fluorescence based on the untreated samples from the three replicates and applying a one-tailed t test. A single-tailed t test was used as the predicted results of change for each activation marker was to only increase after activation. Error bars represent plus or minus two standard deviations.

ing sequence extended with poly(A) tail and homopolymer poly(A) but not homopolymers made of U, C, or G, or mRNA lacking a poly(A) tail induced a calcium flux in DC. This Ca2⫹ flux was inhibited by pretreatment with ATP, ADP, pertussis toxin, and suramin, suggesting the involvement of a P2Y type, Gi-linked nucleotide receptor. Although poly(A) and mRNA lacking a poly(A) tail could induce some markers of DC activation, full maturation of DC, however, required mRNA that could induce TNF-␣ most likely by regions of dsRNA- and poly(A)-mediated signaling. The data suggest that mRNA signals a Gi-protein-linked nucleotide receptor that can be desensitized by ATP and ADP. Two members of a new family of ADP P2Y receptors have recently been identified. This family demonstrates more homology to the UDP-glucose receptor than other members of the P2Y family, responds to concentrations of ADP 5- to 1000-fold lower than the respective ligands of other P2Y receptors, and generally inhibits adenylate cyclase production (29, 36, 42). The tissue distribution of P2Y12, platelets, and brain, and the ability of ATP to desensitize the poly(A) receptor suggest that P2Y12 and -13 receptors

12695

do not signal to mRNA, but as new members of this subclass of P2Y receptors are identified, they will need to be screened for DC expression and mRNA signaling. Based on our data, we cannot determine whether mRNA uses a new nucleotide receptor present on DC but absent on other lymphoid cell types (T cells, B cells, monocyte/macrophages) or on cell lines constructed to express most known P2Y receptors or if mRNA is a ligand for a heterodimer of known GPCR. Heterodimers of GPCR have been observed in multiple systems, including chemokine (43), ␥-aminobutyric acid (44), opioid receptors (45, 46), and nucleotide receptors (P2Y) (47). The functional ␥-aminobutyric acid type B receptor is a heterodimer of two GPCR receptors with low homology to each other. The expression of both receptors is required for signal transduction following ligand binding (44). Thus, GPCR heterodimers can significantly change the binding affinity for ligands and form receptors capable of binding ligands to which neither receptor alone sends signal. The ability of poly(A) to signal a GPCR on DC is highly suggestive that this signaling activity is responsible for the DC-activating properties we observed. We cannot exclude that poly(A) has another activity in addition to GPCR signaling. It is interesting that ATP and poly(A) both activate DC through GPCR but do so through different receptors and signaling mechanisms with different results. ATP signaling through P2Y11 increases cAMP concentrations in DC, a property shared with prostaglandin E2, another DC-activating agent that synergizes with TNF-␣ (48). This activation of cAMP leads to a reduced ability to secrete IL-12 (19, 23). Poly(A), whose signaling activity is likely due to an ADP family, P2Y type nucleotide receptor, does not increase cAMP levels and induces IL-12 secretion. A recent report highlighted the ability of ligands utilizing similar signal transduction pathways to differentially activate DC. In this report, signaling through TLR 2 and 4, which share signaling pathways, including NF-␬B and mitogen-activated protein kinase family member activation, led to DC that differed in the cytokines produced and the resulting phenotype of T cells activated (13). A number of host cell-derived molecules cooperate with TNF-␣, which by itself does not completely activate DC (49), to induce DC activation (18, 19, 23, 49 –51). Some of these molecules (ATP, PGE2) alter the cytokine secretion patterns of the mature DC and the phenotype of the resulting T cells (18, 23, 51). We also observed that the addition of poly(A) to poly(I)䡠poly(C)-activated DC changed the resulting DC phenotype, such that it more closely resembled mRNA activation of DC compared with dsRNA. dsRNA has recently been demonstrated to signal TLR3 at concentrations at least 4-fold higher than that required for mRNA to induce DC activation (53). Further studies will be required to determine whether mRNA, containing a coding sequence, signals through TLR3. mRNA lacking a poly(A) tail induced markers of DC activation but at a level below that observed for mRNA with a poly(A) tail. The signaling by the poly(A) tail both increased the mRNA ability to mature DC and alter the phenotype of the activated DC. These studies identify a new class of ligands for P2Y nucleotide receptors, mRNA. Originally described as a receptor that bound and induced a response to ATP, seven 7-transmembrane, Gprotein linked nucleotide receptors have been identified in humans. Although nucleotide receptors are present on nearly every cell type, including immunologic, neurologic, cardiac, salivary, and bone, we have been unable to identify another cell type that fluxes Ca2⫹ in response to mRNA, suggesting the mRNA nucleotide receptor is restricted in its expression, as has been described for the members of a new family of ADP receptors, P2Y12 (36) and P2Y13 (29). In addition to nucleotide receptors, a class of

12696


receptors responsive to adenine dinucleotides has also been described (54). The ligand binding characteristics and the lack of sensitivity to inhibitors (pertussis toxin and suramin) of these receptors do not parallel those of poly(A)-mediated signaling. A number of orphan receptors related to nucleotide receptors have been identified. Studies are now in progress to identify whether an identified orphan or new nucleotide receptor, specific for poly(A) and mRNA containing a poly(A) tail, is present on DC or if receptor heterodimerization between known GPCR results in RNA signaling. DCs obtain antigen from microbial pathogens through germline-encoded pattern recognition receptors. DCs are also responsible for developing immune responses to altered self where antigen loading occurs through the endocytosis and processing of infected or transformed cells. The use of mRNA to load DC with antigen as a vaccine has been observed to be an efficient and potent method for loading antigen-processing pathways for CD4⫹ T cells, CD8⫹ T cells (24, 25, 55– 64), and B cells2 that exceeds those observed for DNA, viral, and protein immunization methods (52, 63– 66). A hypothesis, that RNA may be used physiologically to load DC with antigen and activate them to induce immune responses to intracellularly infected and transformed self cells, can be forwarded. In this model, cell injury and non-apoptotic cell death releases mRNA that loads antigen-processing pathways and promotes DC activation. Support for this model includes the DC’s superior ability to take up and translate extracellular mRNA (1000- to 10,000-fold better than other antigen presenting cell) (25), and the ability of mice injected with DC, pulsed with naked extracellular mRNA or immunized with RNA, to mount a primary immune response (25, 55). The finding of this report, that extracellular RNA has a specific signaling activity that potently activates DC, adds to this hypothesis. Acknowledgment—We thank Dr. Skip Brass for the use of his fluorescence spectrophotometer. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

21.

22. 23.

2

Steinman, R. M., and Inaba, K. (1999) J. Leukoc. Biol. 66, 205–208 Steinman, R. M. (1991) Annu. Rev. Immunol. 9, 271–296 Lanzavecchia, A. (1996) Curr. Opin. Immunol. 8, 348 –354 Imler, J. L., and Hoffmann, J. A. (2000) Rev. Immunogenet. 2, 294 –304 Pulendran, B., Palucka, K., and Banchereau, J. (2001) Science 293, 253–256 Akira, S., Takeda, K., and Kaisho, T. (2001) Nat. Immunol. 2, 675– 680 Lanzavecchia, A., and Sallusto, F. (2001) Curr. Opin. Immunol. 13, 291–298 Liu, Y. J., Kanzler, H., Soumelis, V., and Gilliet, M. (2001) Nat. Immunol. 2, 585–589 Wan, Y., and Bramson, J. (2001) Curr. Pharm. Des. 7, 977–992 Sallusto, F., and Lanzavecchia, A. (2000) Immunol. Rev. 177, 134 –140 Allavena, P., Sica, A., Vecchi, A., Locati, M., Sozzani, S., and Mantovani, A. (2000) Immunol. Rev. 177, 141–149 Cella, M., Salio, M., Sakakibara, Y., Langen, H., Julkunen, I., and Lanzavecchia, A. (1999) J. Exp. Med. 189, 821– 829 Re, F., and Strominger, J. L. (2001) J. Biol. Chem. 276, 37692–37699 Urban, B. C., Ferguson, D. J., Pain, A., Willcox, N., Plebanski, M., Austyn, J. M., and Roberts, D. J. (1999) Nature 400, 73–77 Urban, B. C., Willcox, N., and Roberts, D. J. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 8750 – 8755 Ito, T., Amakawa, R., Inaba, M., Ikehara, S., Inaba, K., and Fukuhara, S. (2001) J. Immunol. 166, 2961–2969 Berchtold, S., Ogilvie, A. L., Bogdan, C., Muhl-Zurbes, P., Ogilvie, A., Schuler, G., and Steinkasserer, A. (1999) FEBS Lett. 458, 424 – 428 Schnurr, M., Then, F., Galambos, P., Scholz, C., Siegmund, B., Endres, S., and Eigler, A. (2000) J. Immunol. 165, 4704 – 4709 Wilkin, F., Duhant, X., Bruyns, C., Suarez-Huerta, N., Boeynaems, J. M., and Robaye, B. (2001) J. Immunol. 166, 7172–7177 Ferrari, D., La Sala, A., Chiozzi, P., Morelli, A., Falzoni, S., Girolomoni, G., Idzko, M., Dichmann, S., Norgauer, J., and Di Virgilio, F. (2000) FASEB. J. 14, 2466 –2476 Mutini, C., Falzoni, S., Ferrari, D., Chiozzi, P., Morelli, A., Baricordi, O. R., Collo, G., Ricciardi-Castagnoli, P., and Di Virgilio, F. (1999) J. Immunol. 163, 1958 –1965 Coutinho-Silva, R., Perfettini, J. L., Persechini, P. M., Dautry-Varsat, A., and Ojcius, D. M. (2001) Am. J. Physiol. Cell. Physiol. 280, C81–C89 la Sala, A., Ferrari, D., Corinti, S., Cavani, A., Di Virgilio, F., and Girolomoni, G. (2001) J. Immunol. 166, 1611–1617

D. Weissman, G. Cannon, and K. Kariko´ , unpublished observations.

24. Leitner, W. W., Ying, H., and Restifo, N. P. (1999) Vaccine 18, 765–777 25. Weissman, D., Ni, H., Scales, D., Dude, A., Capodici, J., McGibney, K., Abdool, A., Isaacs, S. N., Cannon, G., and Kariko, K. (2000) J. Immunol. 165, 4710 – 4717 26. Communi, D., Govaerts, C., Parmentier, M., and Boeynaems, J. M. (1997) J. Biol. Chem. 272, 31969 –31973 27. Kariko, K., Kuo, A., Barnathan, E. S., and Langer, D. J. (1998) Biochim. Biophys. Acta 1369, 320 –334 28. Communi, D., Motte, S., Boeynaems, J. M., and Pirotton, S. (1996) Eur. J. Pharmacol. 317, 383–389 29. Communi, D., Suarez Gonzalez, N., Detheux, M., Brezillon, S., Lannoy, V., Parmentier, M., and Boeynaems, J. M. (2001) J. Biol. Chem. 276, 41479 – 41485 30. Scales, D., Ni, H., Shaheen, F., Capodici, J., Cannon, G., and Weissman, D. (2001) J. Immunol. 166, 6437– 6443 31. Vissers, J. L., Hartgers, F. C., Lindhout, E., Teunissen, M. B., Figdor, C. G., and Adema, G. J. (2001) J. Leukoc. Biol. 69, 785–793 32. Rissoan, M. C., Soumelis, V., Kadowaki, N., Grouard, G., Briere, F., de Waal Malefyt, R., and Liu, Y. J. (1999) Science 283, 1183–1186 33. Otero, M., Garrad, R. C., Velazquez, B., Hernandez-Perez, M. G., Camden, J. M., Erb, L., Clarke, L. L., Turner, J. T., Weisman, G. A., and Gonzalez, F. A. (2000) Mol. Cell. Biochem. 205, 115–123 34. Sabala, P., Czajkowski, R., Przybylek, K., Kalita, K., Kaczmarek, L., and Baranska, J. (2001) Br. J. Pharmacol. 132, 393– 402 35. Santiago-Perez, L. I., Flores, R. V., Santos-Berrios, C., Chorna, N. E., Krugh, B., Garrad, R. C., Erb, L., Weisman, G. A., and Gonzalez, F. A. (2001) J. Cell. Physiol. 187, 196 –208 36. Hollopeter, G., Jantzen, H. M., Vincent, D., Li, G., England, L., Ramakrishnan, V., Yang, R. B., Nurden, P., Nurden, A., Julius, D., and Conley, P. B. (2001) Nature 409, 202–207 37. Oppermann, M., Mack, M., Proudfoot, A. E., and Olbrich, H. (1999) J. Biol. Chem. 274, 8875– 8885 38. Luther, S. A., and Cyster, J. G. (2001) Nat. Immunol. 2, 102–107 39. Communi, D., Janssens, R., Suarez-Huerta, N., Robaye, B., and Boeynaems, J. M. (2000) Cell. Signal 12, 351–360 40. von Kugelgen, I., and Wetter, A. (2000) Naunyn-Schmiedeberg’s Arch. Pharmacol. 362, 310 –323 41. Di Virgilio, F., Chiozzi, P., Ferrari, D., Falzoni, S., Sanz, J. M., Morelli, A., Torboli, M., Bolognesi, G., and Baricordi, O. R. (2001) Blood. 97, 587– 600 42. Zhang, F. L., Luo, L., Gustafson, E., Lachowicz, J., Smith, M., Qiao, X., Liu, Y. H., Chen, G., Pramanik, B., Laz, T. M., Palmer, K., Bayne, M., and Monsma, F. J., Jr. (2001) J. Biol. Chem. 276, 8608 – 8615 43. Mellado, M., Rodriguez-Frade, J. M., Vila-Coro, A. J., Fernandez, S., Martin de Ana, A., Jones, D. R., Toran, J. L., and Martinez, A. C. (2001) EMBO J. 20, 2497–2507 44. Jones, K. A., Tamm, J. A., Craig, D. A., Ph, D., Yao, W., and Panico, R. (2000) Neuropsychopharmacology 23, S41–S49 45. Jordan, B. A., and Devi, L. A. (1999) Nature 399, 697–700 46. Gomes, I., Jordan, B. A., Gupta, A., Trapaidze, N., Nagy, V., and Devi, L. A. (2000) J. Neurosci. 20, RC110 47. Yoshioka, K., Saitoh, O., and Nakata, H. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 7617–7622 48. Rieser, C., Bock, G., Klocker, H., Bartsch, G., and Thurnher, M. (1997) J. Exp. Med. 186, 1603–1608 49. Granucci, F., Vizzardelli, C., Virzi, E., Rescigno, M., and Ricciardi-Castagnoli, P. (2001) Eur. J. Immunol. 31, 2539 –2546 50. Delneste, Y., Herbault, N., Galea, B., Magistrelli, G., Bazin, I., Bonnefoy, J. Y., and Jeannin, P. (1999) J. Immunol. 163, 3071–3075 51. Steinbrink, K., Paragnik, L., Jonuleit, H., Tuting, T., Knop, J., and Enk, A. H. (2000) Arch. Dermatol. Res. 292, 437– 445 52. Morse, M. A., Lyerly, H. K., Gilboa, E., Thomas, E., and Nair, S. K. (1998) Cancer Res. 58, 2965–2968 53. Alexopoulou, L., Holt, A. C., Medzhitov, R., and Flavell, R. A. (2001) Nature 413, 732–738 54. Pintor, J., and Miras-Portugal, M. T. (1995) Br. J. Pharmacol. 115, 895–902 55. Nair, S. K., Boczkowski, D., Morse, M., Cumming, R. I., Lyerly, H. K., and Gilboa, E. (1998) Nat. Biotechnol. 16, 364 –369 56. Gilboa, E., Nair, S. K., and Lyerly, H. K. (1998) Cancer. Immunol. Immunother. 46, 82– 87 57. Ashley, D. M., Faiola, B., Nair, S., Hale, L. P., Bigner, D. D., and Gilboa, E. (1997) J. Exp. Med. 186, 1177–1182 58. Boczkowski, D., Nair, S. K., Snyder, D., and Gilboa, E. (1996) J. Exp. Med. 184, 465– 472 59. Murphy, G. P., Tjoa, B. A., Simmons, S. J., Ragde, H., Rogers, M., Elgamal, A., Kenny, G. M., Troychak, M. J., Salgaller, M. L., and Boynton, A. L. (1999) Prostate 39, 54 –59 60. Nair, S. K., Heiser, A., Boczkowski, D., Majumdar, A., Naoe, M., Lebkowski, J. S., Vieweg, J., and Gilboa, E. (2000) Nat. Med. 6, 1011–1017 61. Ying, H., Zaks, T. Z., Wang, R. F., Irvine, K. R., Kammula, U. S., Marincola, F. M., Leitner, W. W., and Restifo, N. P. (1999) Nat. Med. 5, 823– 827 62. Zhang, W., He, L., Yuan, Z., Xie, Z., Wang, J., Hamada, H., and Cao, X. (1999) Hum. Gene. Ther. 10, 1151–1161 63. Van Tendeloo, V. F., Ponsaerts, P., Lardon, F., Nijs, G., Lenjou, M., Van Broeckhoven, C., Van Bockstaele, D. R., and Berneman, Z. N. (2001) Blood 98, 49 –56 64. Strobel, I., Berchtold, S., Gotze, A., Schulze, U., Schuler, G., and Steinkasserer, A. (2000) Gene. Ther. 7, 2028 –2035 65. Saenz-Badillos, J., Amin, S. P., and Granstein, R. D. (2001) Exp. Dermatol. 10, 143–154 66. Andersson, C., Vasconcelos, N. M., Sievertzon, M., Haddad, D., Liljeqvist, S., Berglund, P., Liljestrom, P., Ahlborg, N., Stahl, S., and Berzins, K. (2001) Scand J. Immunol. 54, 117–124