- Email: [email protected]
Recombinant respiratory syncytial virus lacking secreted glycoprotein G is attenuated, non-pathogenic but induces protective immunity
Microbes and Infection 6 (2004) 1049–1055 www.elsevier.com/locate/micinf
Original article
Recombinant respiratory syncytial virus lacking secreted glycoprotein G is attenuated, non-pathogenic but induces protective immunity Caroline F. Maher a, Tracy Hussell b, Edward Blair c, Christopher J.A. Ring c, Peter J.M. Openshaw a,* a
Respiratory Medicine, Imperial College, St. Mary’s Campus, Norfolk Place, London W2 1PG, UK b Biological Sciences, Imperial College London, Exhibition Road, London, UK c GlaxoSmithKline Medicines Research Center, Stevenage, UK Received 26 March 2004; accepted 9 July 2004 Available online 03 August 2004
Abstract Respiratory syncytial virus (RSV) causes intense pulmonary inflammatory responses in some infected infants. The surface attachment protein ‘G’ of RSV has membrane-bound and secreted forms and shows homology to the CX3C chemokine fractalkine. Using recombinant techniques, we generated replication-competent recombinant clonal RSV expressing normal G proteins (‘rRSV’) or only the membranebound form of G (‘Gmem rRSV’). Both recombinants grew well in HEp-2 cells, but after primary intranasal infection in mice, pulmonary Gmem rRSV replication was reduced tenfold compared to parental or rRSV; moreover, CCL2 and CCL5 production was greatly reduced and no apparent disease or pulmonary cellular infiltration was observed. However, Gmem rRSV-infected mice developed good antibody responses and were fully protected against subsequent intranasal challenge with parental virus. Even in mice sensitized to G by cutaneous infection with recombinant vaccinia expressing G, intranasal challenge with Gmem rRSV caused insignificant disease. We conclude that secreted G is a key viral product assisting virus replication in vivo, enhancing CCL2 and CCL5 production and promoting illness. Engineered RSV mutants lacking the ability to secrete G are thus promising vaccine candidates. © 2004 Elsevier SAS. All rights reserved. Keywords: Bronchiolitis, viral; Common cold; Asthma; Immunity, mucosal; Pneumovirinae
1. Introduction Respiratory syncytial virus (RSV) is the major cause of viral bronchiolitis, which is the commonest single cause of hospitalization during infancy. RSV causes common colds in older children and in adults and sometimes severe and even fatal pneumonia in the immunocompromised or elderly [1]. Repeated attempts to make effective vaccines have been unsuccessful, in part because immunity is transient but also because it can enhance disease [2]. With recent advances in knowledge of viral immunobiology, disease pathogenesis and molecular biology, it is now possible to manipulate features of RSV crucial to disease production [3], and to engineer specific changes in the virus that could produce vaccine strains that are more immunogenic but less pathologic than the natural variants [4].
In the mouse, it has been shown that exposure to the attachment protein G of RSV or vaccination with formalininactivated RSV predisposes mice to enhanced disease characterized by overproduction of Th2 cytokines (IL-4, IL-5 and IL-13) and lung eosinophilia during intranasal challenge with human strains of RSV [5]. Using recombinant vaccinia viruses expressing secreted or membrane-bound forms of attachment protein G, it appears that the secreted form may be involved in inducing this type of immunopathology [6–8]. Using novel recombinant techniques, mutant RSVs have been made that expresses only secreted or membrane-bound forms of G, the in vitro characteristics of which have been described [9]. However, the in vivo replication and pathologic effects of these mutants have been not been studied in detail.
* Corresponding author. Tel.: +44-207-594-3854; fax: +44-207-262-8913. E-mail address: [email protected] (P.J.M. Openshaw).
We now describe a recombinant RSV lacking the secreted form of G that induces excellent protective immunity but lacks the substantial pathogenic properties of the parental
1286-4579/$ - see front matter © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.micinf.2004.07.001
1050
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
strain, associated with an inability to induce CCL5 (RANTES) and CCL2 (MCP-1) in vivo, suggesting that secreted G causes specific chemokines to be produced and thereby contributes to the induction of inflammatory disease during infection with RSV.
2. Methods 2.1. Cells and viruses HEp-2 and BHK 21 cells were maintained in RPMI-1640 (Life Technologies) containing 10% FBS, 2 mM L-glutamine, penicillin (50 U/ml) and streptomycin (50 mg/ml). Modified vaccinia virus Ankara expressing bacteriophage T7 RNA polymerase (rMVA-T7) was provided by Dr. Bernard Moss (NIH, Bethesda, MD, USA) and was propagated in BHK 21 cells. rMVA-T7 virus titer was determined by TCID50 in BHK 21 cells. The RSV strain A2 (Dr. G. Wertz, University of Alabama, USA) was propagated and titrated for infectivity by plaque assay in HEp-2 monolayers. All stocks were free of mycoplasma infection by DNA hybridization (Gen-Probe Inc.). 2.2. Construction of a recombinant RSV cDNA clone The RSV cDNA plasmid was provided by Dr. Peter Collins (NIH). Its construction has been described [10]. Authentic non-mutated recombinant clonal RSV (designated ‘rRSV’) recovered from the infectious clone was used as a control for recombinants containing specific mutations. To produce a recombinant lacking secreted G, the G gene (flanked by naturally occurring Xho1 and BamH1 sites) was inserted into a pBlueScript® II vector (Stratagene) and subjected to site-directed mutagenesis (SDM) PCR to change the AUG initiation codon (methionine) at AA 48 to GGG (glycine). This change prevents synthesis of secreted G, producing an ORF encoding only the transmembrane form of G (denoted ‘Gmem rRSV’). The primers used contained the following sequence (underlining indicates mismatched sequence): sense 5′CATTATCCATTCTGGAGGGATAATCTCAACTTC-3′ and antisense5′-GAAGTTGAGATTATCCCTGCCAGAATGGATAATG-3′. The PCR products were transformed in Escherichia coli DH10B-competent cells (Invitrogen Ltd., Paisley, Scotland), and purified plasmids were sequenced between the Xho1 and BamH1 sites. Mutant fulllength recombinant antigenomic cDNA was constructed by substituting the wild-type G gene-containing 4019-bp fragment in the D53 plasmid for the mutated fragment. 2.3. Recovery of recombinant RSV (rRSV) Recombinant RSV was recovered from cDNA as described [10]. Briefly, HEp-2 cells in six-well plates at 80% confluence were transfected with recombinant antigenomic cDNA at 0.4 ug per well together with plasmids encoding the
RSV N protein (0.4 ug per well), P protein (0.3 ug per well) and L and M2–1 proteins (0.1 ug per well each) using LipofectACE reagent (Life Technologies). These four plasmids were provided by Dr. Peter Collins (NIH). The transfected cells were infected with rMVA-T7 at an MOI of 10 PFU per cell. The cells were incubated at 33 °C in a CO2 incubator in an overnight transfection. After 24 h the medium was changed and the plate incubated for 48 h. The culture supernatants were passaged onto fresh HEp-2 cells for 6 days. 2.4. ELISA assay for RSV G protein HEp-2 cells were infected with A2, rRSV or Gmem rRSV. At 4, 8, 16, 24, 36, 48, and 72 h; supernatants were removed, ultracentrifuged at 65 000 × g for 4 h, and supernatants and cell pellets were snap frozen. All samples were tested in an ELISA for viral antigen using monoclonal antibody 29 (kind gift of Geraldine Taylor, Compton, UK) to capture RSV G protein and polyclonal anti-RSV peroxidase (Biognesis, Poole, UK) as the detection antibody. 2.5. Evaluation of replication and immunogenicity of Gmem rRSV in BALB/c mice Eight- to 10-week-old BALB/c female mice were obtained from Harlan Olac (Oxon, UK) and kept in FELASA pathogen-free conditions. All procedures were subjected to ethical review and licensed by the Home Office, UK. Mice were challenged with normal or Gmem rRSV by intranasal administration of 106 PFU in 50 µl under light anesthesia. Mice were weighed daily after challenge; on d2, 4 and 7, five mice per group were sacrificed by injection of 3 mg pentobarbitone i.p. and bled out via the femoral vessels. Bronchoalveolar lavage (BAL) and lung tissue were recovered. 2.6. Real-time PCR for viral load RNA was extracted from lung homogenate using chloroform and RNA-STAT 60™, precipitated using 0.5 ml of isopropanol per ml of RNA-STAT 60™, washed in ethanol and vacuum dried. cDNA synthesis was carried out using Omniscript Reverse Transcriptase kit (QIAGEN, UK), and random hexamers (Promega Corporation, UK) were used to prime the reaction at 37 °C for 60 min. Real-time PCR (40 cycles) was performed in an ABI PRISM® 7700 using 1 × Universal Master Mix (Applied Biosystems, UK), forward primer at 900 nM, reverse primer at 300 nM and probe at 100 nM plus 2 µl of target cDNA per reaction. The primers and probes were: RSV L gene: forward, 5′-GAA CTC AGT GTA GGT AGA ATG TTT GCA-3′; reverse, 5′-TTC AGC TAT CAT TTT CTC TGC CAA T-3′, FAM/Tamra probe, 5′-TTT GAA CCT GTC TGA ACA TTC CCG GTT-3′ (MWG Biotech AG, UK). 2.7. Recovery of lung cells BAL was performed as described previously [11]; live cells were counted and stained with QR-conjugated CD4,
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
PE-conjugated CD8 and FITC-conjugated CD45RB for 30 min on ice. Samples were fixed and data collected on 40 000 lymphocytes on a Coulter EPICS Elite flow cytometer. Isotype-matched controls were used to determine background fluorescence. 2.8. RNase protection assay (RPA) Chemokine or chemokine receptor mRNA expression was determined using probe kits mCK5 and mCR5, respectively (Pharmingen, San Diego, CA). Probes were transcribed using T7 RNA polymerase and a-32P-labeled uridine triphosphate (Amersham Pharmacia Biotech, Uppsala, Sweden). RPAs were carried out using the RPA III kit (Ambion Inc., Austin, TX), analyzed by polyacrylamide gel electrophoresis and band intensity assessed electronically (Storm PhosphorImager using ImageQuant v4.2a software; Molecular Dynamics instruments, Sunnyvale, CA). Results were normalized to two housekeeping genes (L32 and GAPDH). It is important to note that the BALB/c genome contains a deletion in the 3′ untranslated region of the CXCL10 gene in a region to which the radiolabeled probe binds. Consequently, the protected probe is 162 residues long instead of 181 residues, as in other strains [12]. Previous studies of this infection model using this technique did not take account of this error in the manufacturer’s reagent, effectively swapping the protected bands for CXCL10 and CCL2 [13,14].
1051
3. Results RSV was recovered from the cloned wild-type (“rRSV”) and G gene-modified (“Gmem”) cDNAs and their viral titers determined. As previously reported [10,16,17], plaque size was reduced in viruses recovered from cDNA compared to parental A2 strain (data not depicted). Supernatants and cell pellets from HEp-2 cells infected with A2, rRSV or Gmem rRSV were tested by ELISA for RSV G protein; G was detectable in cell pellets of HEp-2 cells infected with A2 or rRSV or Gmem rRSV (Fig. 1, panel A). However, in culture supernatants, G was detected in samples from cells infected with A2, rRSV but not Gmem rRSV (Fig. 1, panel B). During primary intranasal infection of BALB/c mice, RSV A2 and rRSV (but not Gmem rRSV) produced slight illness and weight loss on d6 to 7 (not depicted). The number of cells recovered by BAL increased after A2 (9.5 ± 2 × 105 cells per ml; P < 0.01) or rRSV infection (8 ± 2 × 105 cells per ml; P < 0.01), but not after Gmem infection (3.5 ± 1 × 105) or inoculation with HEp-2 material (2 ± 0.5 × 105; Fig. 2A). Lung virus replication was detected on d4, A2 and rRSV growing to consistently higher titer than Gmem rRSV
2.9. Virus replication in BALB/c mice Lungs were homogenized and centrifuged at 400 × g for 5 min (Jouan CR 322). The supernatants were titrated in triplicate on HEp-2 cell monolayers in 96-well flat-bottomed plates; at 24 h cells were washed and incubated with biotinconjugated goat anti-RSV Ab for 1 h, washed and coated with streptavidin-HRP for 30 min followed by 3-amino-9ethylcarbazole. Infectious units were enumerated by light microscopy. 2.10. Evaluation of protective effıcacy of Gmem rRSV in BALB/c mice Eight- to 10-week-old BALB/c female mice were inoculated with normal, Gmem rRSV or HEp-2 material by i.n. administration of 106 PFU in 75 µl. After 3 weeks, mice were re-challenged i.n. with 106 PFU of RSV A2. The animals were sacrificed at 2 and 4 days; lungs and serum were collected and RSV-specific Ab assessed by ELISA [11]. Serum neutralizing Ab against RSV A2 was determined by 50% plaque reduction assay [15]. 2.11. Statistics Student’s t-tests were used assuming unequal variance. Where multiple comparisons were made the Bonferonni correction was applied.
Fig. 1. Detection of G protein in RSV-infected cell cultures. HEp-2 cells were infected with A2, rRSV or Gmem rRSV at t = 0. At 4, 8, 16, 24, 36, 48, and 72 h, supernatant was removed, ultracentrifuged at 65 000 × g for 4 h and snap frozen. Cell pellets (A) and supernatants (B) were tested for RSV antigen by ELISA using monoclonal antibody 29 (anti-G) to capture, and polyclonal anti-RSV peroxidase to detect the antigen. All infected cell pellets contained G, but supernatants contain G only in A2 (black circle) and rRSV (black triangle)-infected cultures. Supernatants from Gmem cultures (black square) only had small amounts of G in the supernatant at 72 h.
1052
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
Fig. 2. Response to primary intranasal infection with native and recombinant RSV. Groups of mice were infected intranasally with RSV A2 (black circle), engineered recombinant virus (rRSV; black triangle), recombinant lacking secreted G (Gmem rRSV, black square) or control material (HEp-2, white circle). A shows that mice infected with Gmem rRSV have little or no increase in cellularity of the BAL at 7 days. In B, multiplication of virus (determined by plaque assay of lung homogenates) is reduced about 10-fold in mice infected with Gmem rRSV.
(Fig. 2B). Similar trends were seen in viral RNA recovery measured by TaqMan PCR, which showed fewer viral mRNA copies in lungs from Gmem rRSV-infected mice (2.5 × 102 viral copies per ng RNA) than from A2 (6 × 103 copies) or rRSV-infected mice (2.5 × 103 copies) on d4 of infection. Since chemokines drive cell recruitment, we examined their expression by RPA. IFN-gamma-inducible protein 10 (IP-10), MCP-1 and CCL5 were all elevated on d2 of intranasal infection with RSV A2. Recombinant viruses induced little IP-10 (Fig. 3A), but Gmem rRSV induced much lower levels of MCP-1 and CCL5 than rRSV (Fig. 3B, C). Control inoculation with HEp-2 material induced no increase in chemokine mRNA expression (not depicted). Despite reduced viral replication and subsequent cellular infiltration, primary infection with Gmem rRSV produced strong serum antibody responses to RSV that were essentially indistinguishable from mice infected with A2 strain or rRSV (Fig. 4A). More importantly, Gmem rRSV infection completely protected against A2 RSV replication during secondary challenge 3 weeks later (Fig. 4B). In mice previously sensitized with recombinant vaccinia virus expressing authentic RSV G protein (VV-G), intranasal challenge with RSV A2 produces profound and augmented weight loss and illness [18,19]. We, therefore, tested whether Gmem rRSV challenge caused a similar effect in G-primed
Fig. 3. Lung chemokine mRNA in mice infected with Gmem rRSV. Whole lung mRNA expression for IP-10, CCL2 and CCL5 was determined on d2, 4 and 7 by RNAse protection. The mean and S.D. of four mice per group are shown, representative of two independent experiments. Chemokine expression is expressed as percentage of the average of L32 and GAPDH expression.
mice. Mice challenged with rRSV lost weight in the same way as those challenged with RSV A2 strain; those challenged with Gmem rRSV lost no weight and suffered no illness (Fig. 5A). Mice sensitized with rVV-G and challenged with Gmem rRSV had similar cell recruitment to mice sensitized with VV-bgal (control construct) and challenged with RSV A2 (Fig. 5B). However, very few eosinophils were recovered from BAL fluid in sensitized mice challenged with Gmem rRSV (5 ± 3%) compared to those challenged with RSV A2 (20 ± 8%; P < 0.01) or rRSV (14 ± 7%; P < 0.03). rRSV and parental A2 strain RSV induced a similar number of eosinophils in VV-G sensitized mice (Fig. 5C).
4. Discussion We have shown that an engineered recombinant RSV that produces only membrane-bound viral glycoprotein G is less pathogenic than a control recombinant virus in vivo infection and challenge. Although of greatly reduced pathogenicity,
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
Fig. 4. Serum antibody response and protection against re-infection after intranasal infection with wild-type or recombinant viruses. Mice were infected intranasally with the indicated virus or control material and serum harvested after 10 days. The presence of RSV-specific antibody was determined by ELISA. Each point represents the mean and standard deviation of 4 mice per group (A). Prior intranasal infection with Gmem rRSV completely protected against parental A2 RSV viral replication during secondary challenge. RSV A2 or Gmem rRSV-infected mice were challenged 2–3 weeks later with RSV A2. Virus titer was determined in lung homogenates by plaque assay. Results are representative of two experiments (B).
this recombinant induced a powerful antibody response and completely protected against re-challenge with virulent human RSV A2. It is so non-pathogenic that it even fails of induce disease in previously sensitized mice. These findings are in direct contrast to the conclusions reached by a recent study of a spontaneous RSV mutant, termed RSV-DsG, shown to be deficient in secreted G protein [20]. These authors found that the mutant virus not only replicated more efficiently in lungs of infected mice, but also induced a higher infiltration of eosinophils and macrophages and higher concentrations of IFN-c and IL-10 in mouse lungs. However, RSV-DsG virus is a spontaneous mutant, and may differ in other ways from the Long strain used as a comparator. By contrast, we modified the G gene in a cloned cDNA copy of the viral genome and generated a recombinant mutant RSV, which does not express the secreted form of the viral G protein. The phenotype of this modified virus was compared with that of virus recovered from the unmodified parental RSV cDNA clone (rRSV) and the A2 strain of RSV. Because Gmem rRSV was generated by modification of the
1053
Fig. 5. Body weight in previously sensitized, RSV-challenged mice. Groups of four to five mice were scarified with vaccinia virus containing the attachment protein of RSV (VV-G) or control VV-bgal and challenged i.n. with RSV A2 or Gmem rRSV 3 weeks later. Weight is expressed as percentage of the starting weight (A). Total viable BAL cell recovery is shown in B, and BAL eosinophils in C. The results are typical from one of three independent experiments. rVV-G, challenge with A2 (black circle); control VV, challenge with A2 (white triangle); rVV-G, challenge with rRSV (black triangle); rVV-G, challenge with Gmem rRSV (black square); control VV, challenge with Gmem rRSV (white square).
D53 RSV cDNA clone, rRSV rescued from this same clone forms an ideal control against which to compare Gmem rRSV, allowing us to test the exact effects of the secreted product of G. The virus tested by Schwarze and Schauer would be likely to have had other mutations in addition to the ability to secrete G. In our studies, plaque size was reduced in viruses recovered from cDNA compared to parental A2 strain (data not depicted). Such a reduction has been seen in other studies of RSV recovered from cloned cDNA [10,16,17] This may be due to the mutations introduced into the viral leader sequences, or may be a more general feature of clonal RSV produced from one genetic sequence. Since Gmem RSV was rescued from cDNA following site-specific mutagenesis of the G gene, any difference in the phenotype of Gmem RSV relative to virus derived from the unmodified (parental) cDNA clone is attributable to that specific modification. Mutation of the G gene sequence was predicted to result in the absence of the secreted form of the G protein, confirmed to be the case by ELISA analysis of cell culture supernatants.
1054
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
The reduction in lung inflammation and chemokine production that we observed in mice infected with the recombinant lacking secreted G strongly suggests that secreted G is involved in promoting inflammatory disease during RSV infection, but that it is not essential to induction of protective immune responses. RSV has several intriguing features that might explain these unusual immunological phenomena. First, the attachment glycoprotein G is exceptionally heavily glycosylated, mostly with O-linked carbohydrate. Indeed, some 63% of the mass of G is carbohydrate [21]. This level of O-linked glycosylation is exceptional amongst viral glycoproteins but is seen amongst some filoviruses, which are closely related to RSV in terms of sequence homology and genome structure. We and others have previously shown that CD8+ T cell priming prevents Th2-driven eosinophilia by producing IFN-c [19,22,23]. Secreted G is likely to be processed as a soluble antigen and to be endocytosed by professional antigen-presenting cells. In these cells, endocytic vesicles fuse with those carrying MHC class II molecules, leading to presentation to CD4 T cells. Membrane-bound G, however, would be processed and presented as a newly formed intracellular protein and targeted by ubiquitination for destruction in the proteasome with subsequent loading into MHC class I molecules. Although these processing pathways are not mutually exclusive, qualitative differences are likely to exist for the processing and presentation of different G protein species. In support of this explanation, we found no increase in CD8+ T cell numbers or in IFN-c production by cells in Gmem rRSV-challenged mice (not depicted). Second, the central conserved region of G shows sequence and structural homology to a component of the TNF receptor [24], raising the possibility that secreted G could bind TNF or similar molecules and modify the immune response to infection. These observations suggest that the RSV G protein has evolved to interact with the immune system and produce effects on the host response that are advantageous to the virus. Third, G has intriguing homologies to the CX3C chemokine fractalkine; both have transmembrane and cytoplasmic domains plus a serine/threonine-rich mucin-like stalk connected to a highly conserved non-glycosylated region containing four cysteines, and have secreted form and membrane-bound form. It has recently been shown that RSV G protein binds to the human fractalkine receptor and has some fractalkine-like activity in vitro [25]. The engagement of secreted G with CX3CR1 on CD4 T cells may affect cell recruitment and activation [25]. Activated CD4+ T cells are required for lung eosinophilia in RSV-infected mice [19], and a reduction in CD4+ T cell number or activation, by depletion or interference with pro-inflammatory cytokines, prevents eosinophilic lung inflammation [26–28]. If secreted G recruits and/or activates CD4+ T cells, Gmem rRSV infection would lead to reduced inflammation with a reduction in chemokine expression and cell recruitment. Reduced numbers of activated CD4+ T cells in mice infected with Gmem would contribute to reduced disease severity. It is interesting to note that CCL2 and CCL5 mRNA were reduced in lung samples from mice infected with Gmem
rRSV. Clearly, this could be simply due to reduced infectivity and replication, but this seems unlikely because rRSV was also attenuated in its replication, yet induced abundant CCL5 and CCL2 was found. Furthermore, reduced chemokine expression in the absence of secreted G protein is unlike that described for an attenuated cold passaged RSV strain (CP52, lacking G and SH proteins [14]), indicating that the absence of soluble G alone has a specific modulating effect on inflammatory responses in the lung. The induction of CCL5 during RSV infection of both children and mice has been extensively documented [29,30]. In RSV bronchiolitis, dsRNA intermediates may induce CCL5 secretion from airway epithelial cells, which may predict the later development of recurrent wheeze [31,32]. A reduction of CCL5 by antiCCL-5 [33] or Perflubron [13] treatment reduces inflammation and/or airway hyperreactivity. Since CCL5 recruits and activates monocytes, lymphocytes and eosinophils, a reduction of this chemokine during Gmem rRSV infection may explain the improved clinical outcome in the current study. In conclusion, improved knowledge of the immunobiology of RSV disease and the development of new methods of constructing mutant viruses now enable the production of new live vaccines against RSV disease that induce strong protective immune responses but lack the pathogenic potential of natural RSV variants. Such novel vaccine candidates should finally overcome the barriers to empirical RSV vaccine development.
Acknowledgments Supported by Welcome Trust Program grants 054797 and 071381 and MRC (UK) Case Award G78/5996. We thank Dr. Peter Collins, NIAID, Bethesda, USA for generously providing the cDNA plasmid, reagents and expertise.
References [1] [2]
[3]
[4]
[5] [6]
C.B. Hall, Respiratory syncytial virus and parainfluenza virus, New Engl. J. Med. 344 (2001) 1917–1928. K. McIntosh, J.M. Fishaut, Immunopathologic mechanisms in lower respiratory tract disease of infants due to respiratory syncytial virus, Prog. Med. Virol. 26 (1980) 94–118. T.E. Sparer, S. Matthews, T. Hussell, A.J. Rae, B. García-Barreno, J.A. Melero, P.J.M. Openshaw, Eliminating a region of respiratory syncytial virus attachment protein allows induction of protective immunity without vaccine-enhanced lung eosinophilia, J. Exp. Med. 187 (1998) 1921–1926. A. Bukreyev, M.H. Skiadopoulos, J. McAuliffe, B.R. Murphy, P.L. Collins, A.C. Schmidt, More antibody with less antigen: can immunogenicity of attenuated live virus vaccines be improved? Proc. Natl. Acad. Sci. USA 99 (2002) 16987–16991. P.J.M. Openshaw, F.J. Culley, W. Olszewska, Immunopathogenesis of vaccine-enhanced RSV disease, Vaccine 20 (2001) s27–s31. G.P. Bembridge, J.A. Lopez, R. Bustos, J.A. Melero, R. Cook, H. Mason, G. Taylor, Priming with a secreted form of the fusion protein of respiratory syncytial virus (RSV) promotes interleukin-4 (IL-4) and IL-5 production but not pulmonary eosinophilia following RSV challenge, J. Virol. 73 (1999) 10086–10094.
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055 [7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
T.R. Johnson, B.S. Graham, Secreted respiratory syncytial virus G glycoprotein induces interleukin-5 (IL-5), IL-13, and eosinophilia by an IL-4-independent mechanism, J. Virol. 75 (1999) 8485–8495. T.R. Johnson, J.E. Johnson, S.R. Roberts, G.W. Wertz, R.A. Parker, B.S. Graham, Priming with secreted glycoprotein G of respiratory syncytial virus (RSV) augments interleukin-5 production and tissue eosinophilia after RSV challenge, J. Virol. 72 (1998) 2871–2880. M.N. Teng, S.S. Whitehead, P.L. Collins, Contribution of the respiratory syncytial virus G glycoprotein and its secreted and membranebound forms to virus replication in vitro and in vivo, Virology 289 (2001) 283–296. P.L. Collins, M.G. Hill, E. Camargo, H. Grosfeld, R.M. Chanock, B.R. Murphy, Production of infectious human respiratory syncytial virus from cloned cDNA confirms an essential role for the transcription elongation factor from the 5′ proximal open reading frame of the M2 mRNA in gene expression and provides a capability for vaccine development, Proc. Natl. Acad. Sci. USA 92 (1995) 11563–11567. T. Hussell, U. Khan, P.J.M. Openshaw, IL-12 treatment attenuates Th2 and B cell responses but does not improve vaccine-enhanced lung illness, J. Immunol. 159 (1997) 328–334. W. Hallensleben, L. Biro, C. Sauder, J. Hausmann, V.C. Asensio, I.L. Campbell, P. Staeheli, A polymorphism in the mouse crg-2/IP-10 gene complicates chemokine gene expression analysis using a commercial ribonuclease protection assay, J. Immunol. Methods 234 (2000) 149–151. H.A. Haeberle, W.A. Kuziel, H.J. Dieterich, A. Casola, Z. Gatalica, R.P. Garofalo, Inducible expression of inflammatory chemokines in respiratory syncytial virus-infected mice: role of MIP-1alpha in lung pathology, J. Virol. 75 (2001) 878–890. R.A. Tripp, L. Jones, L.J. Anderson, Respiratory syncytial virus G and/or SH glycoproteins modify CC and CXC chemokine mRNA expression in the BALB/c mouse, J. Virol. 74 (2000) 6227–6229. H.V. Coates, D.W. Alling, R.M. Chanock, An antigenic analysis of respiratory syncytial virus isolates by a plaque reduction neutralization test, Am. J. Epidemiol. 83 (1966) 299–313. H. Jin, D. Clarke, H.Z. Zhou, X. Cheng, K. Coelingh, M. Bryant, S. Li, Recombinant human respiratory syncytial virus (RSV) from cDNA and construction of subgroup A and B chimeric RSV, Virology 251 (1998) 206–214. M.B. Elliott, K.S. Pryharski, Q. Yu, C.L. Parks, T.S. Laughlin, C.K. Gupta, R.A. Lerch, V.B. Randolph, N.A. LaPierre, K.M. Dack, G.E. Hancock, Recombinant respiratory syncytial viruses lacking the C-terminal third of the attachment (G) protein are immunogenic and attenuated in vivo and in vitro, J. Virol. 78 (2004) 5773–5783. P.J.M. Openshaw, S.L. Clarke, F.M. Record, Pulmonary eosinophilic response to respiratory syncytial virus infection in mice sensitized to the major surface glycoprotein G, Int. Immunol. 4 (1992) 493–500. W.H. Alwan, W.J. Kozlowska, P.J.M. Openshaw, Distinct types of lung disease caused by functional subsets of antiviral T cells, J. Exp. Med. 179 (1994) 81–89.
1055
[20] J. Schwarze, U. Schauer, Enhanced virulence, airway inflammation and impaired lung function induced by respiratory syncytial virus deficient in secreted G protein, Thorax 59 (2004) 517–521. [21] G.W. Wertz, M. Krieger, L.A. Ball, Structure and cell surface maturation of the attachment glycoprotein of human respiratory syncytial virus in a cell line deficient in O glycosylation, J. Virol. 63 (1989) 4767–4776. [22] T. Hussell, P.J.M. Openshaw, IL-12 activated NK cells reduce lung eosinophilia to the attachment protein of respiratory syncytial virus but do not enhance the severity of illness after lung challenge in CD8 cell-immunodeficient conditions, J. Immunol. 165 (2000) 7109–7115. [23] A. Srikiatkhachorn, T.J. Braciale, Virus specific CD8+ T lymphocytes downregulate T helper cell type 2 cytokine secretion and pulmonary eosinophilia during experimental murine respiratory syncytial virus Infection, J. Exp. Med. 186 (1997) 421–432. [24] J.P. Langedijk, B.L. de Groot, H.J. Berendsen, J.T. Van Oirschot, Structural homology of the central conserved region of the attachment protein G of respiratory syncytial virus with the fourth subdomain of 55-kDa tumor necrosis factor receptor, Virology 243 (1998) 293–302. [25] R.A. Tripp, L.P. Jones, L.M. Haynes, H. Zheng, P.M. Murphy, L.J. Anderson, CX3C chemokine mimicry by respiratory syncytial virus G glycoprotein, Nat. Immun. 2 (2001) 732–738. [26] T. Hussell, A. Pennycook, P.J.M. Openshaw, Inhibition of tumour necrosis factor reduces the severity of virus-specific lung immunopathology, Eur. J. Immunol. 31 (2001) 2566–2573. [27] G. Walzl, S. Matthews, S. Kendall, J.C. Gutiérrez-Ramos, A.J. Coyle, P.J.M. Openshaw, T. Hussell, Inhibition of T1/ST2 during respiratory syncytial virus infection prevents Th2- but not Th1-driven immunopathology, J. Exp. Med. 193 (2001) 785–792. [28] T. Hussell, C.J. Baldwin, A. O’Garra, P.J.M. Openshaw, CD8+ T-cells control Th2-driven pathology during pulmonary respiratory syncytial virus infection, Eur. J. Immunol. 27 (1997) 3341–3349. [29] Y. Zhang, B.A. Luxon, A. Casola, R.P. Garofalo, M. Jamaluddin, A.R. Brasier, Expression of respiratory syncytial virus-induced chemokine gene networks in lower airway epithelial cells revealed by cDNA microarrays, J. Virol. 75 (2001) 9044–9058. [30] R.A. Tripp, D. Moore, A. Barskey, L. Jones, C. Moscatiello, H. Keyserling, L.J. Anderson, Peripheral blood mononuclear cells from infants hospitalized because of respiratory syncytial virus infection express T helper-1 and T helper-2 cytokines and CC chemokine messenger RNA, J. Infect. Dis. 185 (2002) 1388–1394. [31] H.L. Chung, S.G. Kim, RANTES may be predictive of later recurrent wheezing after respiratory syncytial virus bronchiolitis in infants, Ann. Allergy Asthma Immunol. 88 (2002) 463–467. [32] J.E. Gern, W.W. Busse, Relationship of viral infections to wheezing illnesses and asthma, Nat. Rev. Immunol. 2 (2002) 132–138. [33] K.K. Tekkanat, H. Maassab, A. Miller, A.A. Berlin, S.L. Kunkel, N.W. Lukacs, RANTES (CCL5) production during primary respiratory syncytial virus infection exacerbates airway disease, Eur. J. Immunol. 32 (2002) 3276–3284.
Original article
Recombinant respiratory syncytial virus lacking secreted glycoprotein G is attenuated, non-pathogenic but induces protective immunity Caroline F. Maher a, Tracy Hussell b, Edward Blair c, Christopher J.A. Ring c, Peter J.M. Openshaw a,* a
Respiratory Medicine, Imperial College, St. Mary’s Campus, Norfolk Place, London W2 1PG, UK b Biological Sciences, Imperial College London, Exhibition Road, London, UK c GlaxoSmithKline Medicines Research Center, Stevenage, UK Received 26 March 2004; accepted 9 July 2004 Available online 03 August 2004
Abstract Respiratory syncytial virus (RSV) causes intense pulmonary inflammatory responses in some infected infants. The surface attachment protein ‘G’ of RSV has membrane-bound and secreted forms and shows homology to the CX3C chemokine fractalkine. Using recombinant techniques, we generated replication-competent recombinant clonal RSV expressing normal G proteins (‘rRSV’) or only the membranebound form of G (‘Gmem rRSV’). Both recombinants grew well in HEp-2 cells, but after primary intranasal infection in mice, pulmonary Gmem rRSV replication was reduced tenfold compared to parental or rRSV; moreover, CCL2 and CCL5 production was greatly reduced and no apparent disease or pulmonary cellular infiltration was observed. However, Gmem rRSV-infected mice developed good antibody responses and were fully protected against subsequent intranasal challenge with parental virus. Even in mice sensitized to G by cutaneous infection with recombinant vaccinia expressing G, intranasal challenge with Gmem rRSV caused insignificant disease. We conclude that secreted G is a key viral product assisting virus replication in vivo, enhancing CCL2 and CCL5 production and promoting illness. Engineered RSV mutants lacking the ability to secrete G are thus promising vaccine candidates. © 2004 Elsevier SAS. All rights reserved. Keywords: Bronchiolitis, viral; Common cold; Asthma; Immunity, mucosal; Pneumovirinae
1. Introduction Respiratory syncytial virus (RSV) is the major cause of viral bronchiolitis, which is the commonest single cause of hospitalization during infancy. RSV causes common colds in older children and in adults and sometimes severe and even fatal pneumonia in the immunocompromised or elderly [1]. Repeated attempts to make effective vaccines have been unsuccessful, in part because immunity is transient but also because it can enhance disease [2]. With recent advances in knowledge of viral immunobiology, disease pathogenesis and molecular biology, it is now possible to manipulate features of RSV crucial to disease production [3], and to engineer specific changes in the virus that could produce vaccine strains that are more immunogenic but less pathologic than the natural variants [4].
In the mouse, it has been shown that exposure to the attachment protein G of RSV or vaccination with formalininactivated RSV predisposes mice to enhanced disease characterized by overproduction of Th2 cytokines (IL-4, IL-5 and IL-13) and lung eosinophilia during intranasal challenge with human strains of RSV [5]. Using recombinant vaccinia viruses expressing secreted or membrane-bound forms of attachment protein G, it appears that the secreted form may be involved in inducing this type of immunopathology [6–8]. Using novel recombinant techniques, mutant RSVs have been made that expresses only secreted or membrane-bound forms of G, the in vitro characteristics of which have been described [9]. However, the in vivo replication and pathologic effects of these mutants have been not been studied in detail.
* Corresponding author. Tel.: +44-207-594-3854; fax: +44-207-262-8913. E-mail address: [email protected] (P.J.M. Openshaw).
We now describe a recombinant RSV lacking the secreted form of G that induces excellent protective immunity but lacks the substantial pathogenic properties of the parental
1286-4579/$ - see front matter © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.micinf.2004.07.001
1050
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
strain, associated with an inability to induce CCL5 (RANTES) and CCL2 (MCP-1) in vivo, suggesting that secreted G causes specific chemokines to be produced and thereby contributes to the induction of inflammatory disease during infection with RSV.
2. Methods 2.1. Cells and viruses HEp-2 and BHK 21 cells were maintained in RPMI-1640 (Life Technologies) containing 10% FBS, 2 mM L-glutamine, penicillin (50 U/ml) and streptomycin (50 mg/ml). Modified vaccinia virus Ankara expressing bacteriophage T7 RNA polymerase (rMVA-T7) was provided by Dr. Bernard Moss (NIH, Bethesda, MD, USA) and was propagated in BHK 21 cells. rMVA-T7 virus titer was determined by TCID50 in BHK 21 cells. The RSV strain A2 (Dr. G. Wertz, University of Alabama, USA) was propagated and titrated for infectivity by plaque assay in HEp-2 monolayers. All stocks were free of mycoplasma infection by DNA hybridization (Gen-Probe Inc.). 2.2. Construction of a recombinant RSV cDNA clone The RSV cDNA plasmid was provided by Dr. Peter Collins (NIH). Its construction has been described [10]. Authentic non-mutated recombinant clonal RSV (designated ‘rRSV’) recovered from the infectious clone was used as a control for recombinants containing specific mutations. To produce a recombinant lacking secreted G, the G gene (flanked by naturally occurring Xho1 and BamH1 sites) was inserted into a pBlueScript® II vector (Stratagene) and subjected to site-directed mutagenesis (SDM) PCR to change the AUG initiation codon (methionine) at AA 48 to GGG (glycine). This change prevents synthesis of secreted G, producing an ORF encoding only the transmembrane form of G (denoted ‘Gmem rRSV’). The primers used contained the following sequence (underlining indicates mismatched sequence): sense 5′CATTATCCATTCTGGAGGGATAATCTCAACTTC-3′ and antisense5′-GAAGTTGAGATTATCCCTGCCAGAATGGATAATG-3′. The PCR products were transformed in Escherichia coli DH10B-competent cells (Invitrogen Ltd., Paisley, Scotland), and purified plasmids were sequenced between the Xho1 and BamH1 sites. Mutant fulllength recombinant antigenomic cDNA was constructed by substituting the wild-type G gene-containing 4019-bp fragment in the D53 plasmid for the mutated fragment. 2.3. Recovery of recombinant RSV (rRSV) Recombinant RSV was recovered from cDNA as described [10]. Briefly, HEp-2 cells in six-well plates at 80% confluence were transfected with recombinant antigenomic cDNA at 0.4 ug per well together with plasmids encoding the
RSV N protein (0.4 ug per well), P protein (0.3 ug per well) and L and M2–1 proteins (0.1 ug per well each) using LipofectACE reagent (Life Technologies). These four plasmids were provided by Dr. Peter Collins (NIH). The transfected cells were infected with rMVA-T7 at an MOI of 10 PFU per cell. The cells were incubated at 33 °C in a CO2 incubator in an overnight transfection. After 24 h the medium was changed and the plate incubated for 48 h. The culture supernatants were passaged onto fresh HEp-2 cells for 6 days. 2.4. ELISA assay for RSV G protein HEp-2 cells were infected with A2, rRSV or Gmem rRSV. At 4, 8, 16, 24, 36, 48, and 72 h; supernatants were removed, ultracentrifuged at 65 000 × g for 4 h, and supernatants and cell pellets were snap frozen. All samples were tested in an ELISA for viral antigen using monoclonal antibody 29 (kind gift of Geraldine Taylor, Compton, UK) to capture RSV G protein and polyclonal anti-RSV peroxidase (Biognesis, Poole, UK) as the detection antibody. 2.5. Evaluation of replication and immunogenicity of Gmem rRSV in BALB/c mice Eight- to 10-week-old BALB/c female mice were obtained from Harlan Olac (Oxon, UK) and kept in FELASA pathogen-free conditions. All procedures were subjected to ethical review and licensed by the Home Office, UK. Mice were challenged with normal or Gmem rRSV by intranasal administration of 106 PFU in 50 µl under light anesthesia. Mice were weighed daily after challenge; on d2, 4 and 7, five mice per group were sacrificed by injection of 3 mg pentobarbitone i.p. and bled out via the femoral vessels. Bronchoalveolar lavage (BAL) and lung tissue were recovered. 2.6. Real-time PCR for viral load RNA was extracted from lung homogenate using chloroform and RNA-STAT 60™, precipitated using 0.5 ml of isopropanol per ml of RNA-STAT 60™, washed in ethanol and vacuum dried. cDNA synthesis was carried out using Omniscript Reverse Transcriptase kit (QIAGEN, UK), and random hexamers (Promega Corporation, UK) were used to prime the reaction at 37 °C for 60 min. Real-time PCR (40 cycles) was performed in an ABI PRISM® 7700 using 1 × Universal Master Mix (Applied Biosystems, UK), forward primer at 900 nM, reverse primer at 300 nM and probe at 100 nM plus 2 µl of target cDNA per reaction. The primers and probes were: RSV L gene: forward, 5′-GAA CTC AGT GTA GGT AGA ATG TTT GCA-3′; reverse, 5′-TTC AGC TAT CAT TTT CTC TGC CAA T-3′, FAM/Tamra probe, 5′-TTT GAA CCT GTC TGA ACA TTC CCG GTT-3′ (MWG Biotech AG, UK). 2.7. Recovery of lung cells BAL was performed as described previously [11]; live cells were counted and stained with QR-conjugated CD4,
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
PE-conjugated CD8 and FITC-conjugated CD45RB for 30 min on ice. Samples were fixed and data collected on 40 000 lymphocytes on a Coulter EPICS Elite flow cytometer. Isotype-matched controls were used to determine background fluorescence. 2.8. RNase protection assay (RPA) Chemokine or chemokine receptor mRNA expression was determined using probe kits mCK5 and mCR5, respectively (Pharmingen, San Diego, CA). Probes were transcribed using T7 RNA polymerase and a-32P-labeled uridine triphosphate (Amersham Pharmacia Biotech, Uppsala, Sweden). RPAs were carried out using the RPA III kit (Ambion Inc., Austin, TX), analyzed by polyacrylamide gel electrophoresis and band intensity assessed electronically (Storm PhosphorImager using ImageQuant v4.2a software; Molecular Dynamics instruments, Sunnyvale, CA). Results were normalized to two housekeeping genes (L32 and GAPDH). It is important to note that the BALB/c genome contains a deletion in the 3′ untranslated region of the CXCL10 gene in a region to which the radiolabeled probe binds. Consequently, the protected probe is 162 residues long instead of 181 residues, as in other strains [12]. Previous studies of this infection model using this technique did not take account of this error in the manufacturer’s reagent, effectively swapping the protected bands for CXCL10 and CCL2 [13,14].
1051
3. Results RSV was recovered from the cloned wild-type (“rRSV”) and G gene-modified (“Gmem”) cDNAs and their viral titers determined. As previously reported [10,16,17], plaque size was reduced in viruses recovered from cDNA compared to parental A2 strain (data not depicted). Supernatants and cell pellets from HEp-2 cells infected with A2, rRSV or Gmem rRSV were tested by ELISA for RSV G protein; G was detectable in cell pellets of HEp-2 cells infected with A2 or rRSV or Gmem rRSV (Fig. 1, panel A). However, in culture supernatants, G was detected in samples from cells infected with A2, rRSV but not Gmem rRSV (Fig. 1, panel B). During primary intranasal infection of BALB/c mice, RSV A2 and rRSV (but not Gmem rRSV) produced slight illness and weight loss on d6 to 7 (not depicted). The number of cells recovered by BAL increased after A2 (9.5 ± 2 × 105 cells per ml; P < 0.01) or rRSV infection (8 ± 2 × 105 cells per ml; P < 0.01), but not after Gmem infection (3.5 ± 1 × 105) or inoculation with HEp-2 material (2 ± 0.5 × 105; Fig. 2A). Lung virus replication was detected on d4, A2 and rRSV growing to consistently higher titer than Gmem rRSV
2.9. Virus replication in BALB/c mice Lungs were homogenized and centrifuged at 400 × g for 5 min (Jouan CR 322). The supernatants were titrated in triplicate on HEp-2 cell monolayers in 96-well flat-bottomed plates; at 24 h cells were washed and incubated with biotinconjugated goat anti-RSV Ab for 1 h, washed and coated with streptavidin-HRP for 30 min followed by 3-amino-9ethylcarbazole. Infectious units were enumerated by light microscopy. 2.10. Evaluation of protective effıcacy of Gmem rRSV in BALB/c mice Eight- to 10-week-old BALB/c female mice were inoculated with normal, Gmem rRSV or HEp-2 material by i.n. administration of 106 PFU in 75 µl. After 3 weeks, mice were re-challenged i.n. with 106 PFU of RSV A2. The animals were sacrificed at 2 and 4 days; lungs and serum were collected and RSV-specific Ab assessed by ELISA [11]. Serum neutralizing Ab against RSV A2 was determined by 50% plaque reduction assay [15]. 2.11. Statistics Student’s t-tests were used assuming unequal variance. Where multiple comparisons were made the Bonferonni correction was applied.
Fig. 1. Detection of G protein in RSV-infected cell cultures. HEp-2 cells were infected with A2, rRSV or Gmem rRSV at t = 0. At 4, 8, 16, 24, 36, 48, and 72 h, supernatant was removed, ultracentrifuged at 65 000 × g for 4 h and snap frozen. Cell pellets (A) and supernatants (B) were tested for RSV antigen by ELISA using monoclonal antibody 29 (anti-G) to capture, and polyclonal anti-RSV peroxidase to detect the antigen. All infected cell pellets contained G, but supernatants contain G only in A2 (black circle) and rRSV (black triangle)-infected cultures. Supernatants from Gmem cultures (black square) only had small amounts of G in the supernatant at 72 h.
1052
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
Fig. 2. Response to primary intranasal infection with native and recombinant RSV. Groups of mice were infected intranasally with RSV A2 (black circle), engineered recombinant virus (rRSV; black triangle), recombinant lacking secreted G (Gmem rRSV, black square) or control material (HEp-2, white circle). A shows that mice infected with Gmem rRSV have little or no increase in cellularity of the BAL at 7 days. In B, multiplication of virus (determined by plaque assay of lung homogenates) is reduced about 10-fold in mice infected with Gmem rRSV.
(Fig. 2B). Similar trends were seen in viral RNA recovery measured by TaqMan PCR, which showed fewer viral mRNA copies in lungs from Gmem rRSV-infected mice (2.5 × 102 viral copies per ng RNA) than from A2 (6 × 103 copies) or rRSV-infected mice (2.5 × 103 copies) on d4 of infection. Since chemokines drive cell recruitment, we examined their expression by RPA. IFN-gamma-inducible protein 10 (IP-10), MCP-1 and CCL5 were all elevated on d2 of intranasal infection with RSV A2. Recombinant viruses induced little IP-10 (Fig. 3A), but Gmem rRSV induced much lower levels of MCP-1 and CCL5 than rRSV (Fig. 3B, C). Control inoculation with HEp-2 material induced no increase in chemokine mRNA expression (not depicted). Despite reduced viral replication and subsequent cellular infiltration, primary infection with Gmem rRSV produced strong serum antibody responses to RSV that were essentially indistinguishable from mice infected with A2 strain or rRSV (Fig. 4A). More importantly, Gmem rRSV infection completely protected against A2 RSV replication during secondary challenge 3 weeks later (Fig. 4B). In mice previously sensitized with recombinant vaccinia virus expressing authentic RSV G protein (VV-G), intranasal challenge with RSV A2 produces profound and augmented weight loss and illness [18,19]. We, therefore, tested whether Gmem rRSV challenge caused a similar effect in G-primed
Fig. 3. Lung chemokine mRNA in mice infected with Gmem rRSV. Whole lung mRNA expression for IP-10, CCL2 and CCL5 was determined on d2, 4 and 7 by RNAse protection. The mean and S.D. of four mice per group are shown, representative of two independent experiments. Chemokine expression is expressed as percentage of the average of L32 and GAPDH expression.
mice. Mice challenged with rRSV lost weight in the same way as those challenged with RSV A2 strain; those challenged with Gmem rRSV lost no weight and suffered no illness (Fig. 5A). Mice sensitized with rVV-G and challenged with Gmem rRSV had similar cell recruitment to mice sensitized with VV-bgal (control construct) and challenged with RSV A2 (Fig. 5B). However, very few eosinophils were recovered from BAL fluid in sensitized mice challenged with Gmem rRSV (5 ± 3%) compared to those challenged with RSV A2 (20 ± 8%; P < 0.01) or rRSV (14 ± 7%; P < 0.03). rRSV and parental A2 strain RSV induced a similar number of eosinophils in VV-G sensitized mice (Fig. 5C).
4. Discussion We have shown that an engineered recombinant RSV that produces only membrane-bound viral glycoprotein G is less pathogenic than a control recombinant virus in vivo infection and challenge. Although of greatly reduced pathogenicity,
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
Fig. 4. Serum antibody response and protection against re-infection after intranasal infection with wild-type or recombinant viruses. Mice were infected intranasally with the indicated virus or control material and serum harvested after 10 days. The presence of RSV-specific antibody was determined by ELISA. Each point represents the mean and standard deviation of 4 mice per group (A). Prior intranasal infection with Gmem rRSV completely protected against parental A2 RSV viral replication during secondary challenge. RSV A2 or Gmem rRSV-infected mice were challenged 2–3 weeks later with RSV A2. Virus titer was determined in lung homogenates by plaque assay. Results are representative of two experiments (B).
this recombinant induced a powerful antibody response and completely protected against re-challenge with virulent human RSV A2. It is so non-pathogenic that it even fails of induce disease in previously sensitized mice. These findings are in direct contrast to the conclusions reached by a recent study of a spontaneous RSV mutant, termed RSV-DsG, shown to be deficient in secreted G protein [20]. These authors found that the mutant virus not only replicated more efficiently in lungs of infected mice, but also induced a higher infiltration of eosinophils and macrophages and higher concentrations of IFN-c and IL-10 in mouse lungs. However, RSV-DsG virus is a spontaneous mutant, and may differ in other ways from the Long strain used as a comparator. By contrast, we modified the G gene in a cloned cDNA copy of the viral genome and generated a recombinant mutant RSV, which does not express the secreted form of the viral G protein. The phenotype of this modified virus was compared with that of virus recovered from the unmodified parental RSV cDNA clone (rRSV) and the A2 strain of RSV. Because Gmem rRSV was generated by modification of the
1053
Fig. 5. Body weight in previously sensitized, RSV-challenged mice. Groups of four to five mice were scarified with vaccinia virus containing the attachment protein of RSV (VV-G) or control VV-bgal and challenged i.n. with RSV A2 or Gmem rRSV 3 weeks later. Weight is expressed as percentage of the starting weight (A). Total viable BAL cell recovery is shown in B, and BAL eosinophils in C. The results are typical from one of three independent experiments. rVV-G, challenge with A2 (black circle); control VV, challenge with A2 (white triangle); rVV-G, challenge with rRSV (black triangle); rVV-G, challenge with Gmem rRSV (black square); control VV, challenge with Gmem rRSV (white square).
D53 RSV cDNA clone, rRSV rescued from this same clone forms an ideal control against which to compare Gmem rRSV, allowing us to test the exact effects of the secreted product of G. The virus tested by Schwarze and Schauer would be likely to have had other mutations in addition to the ability to secrete G. In our studies, plaque size was reduced in viruses recovered from cDNA compared to parental A2 strain (data not depicted). Such a reduction has been seen in other studies of RSV recovered from cloned cDNA [10,16,17] This may be due to the mutations introduced into the viral leader sequences, or may be a more general feature of clonal RSV produced from one genetic sequence. Since Gmem RSV was rescued from cDNA following site-specific mutagenesis of the G gene, any difference in the phenotype of Gmem RSV relative to virus derived from the unmodified (parental) cDNA clone is attributable to that specific modification. Mutation of the G gene sequence was predicted to result in the absence of the secreted form of the G protein, confirmed to be the case by ELISA analysis of cell culture supernatants.
1054
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055
The reduction in lung inflammation and chemokine production that we observed in mice infected with the recombinant lacking secreted G strongly suggests that secreted G is involved in promoting inflammatory disease during RSV infection, but that it is not essential to induction of protective immune responses. RSV has several intriguing features that might explain these unusual immunological phenomena. First, the attachment glycoprotein G is exceptionally heavily glycosylated, mostly with O-linked carbohydrate. Indeed, some 63% of the mass of G is carbohydrate [21]. This level of O-linked glycosylation is exceptional amongst viral glycoproteins but is seen amongst some filoviruses, which are closely related to RSV in terms of sequence homology and genome structure. We and others have previously shown that CD8+ T cell priming prevents Th2-driven eosinophilia by producing IFN-c [19,22,23]. Secreted G is likely to be processed as a soluble antigen and to be endocytosed by professional antigen-presenting cells. In these cells, endocytic vesicles fuse with those carrying MHC class II molecules, leading to presentation to CD4 T cells. Membrane-bound G, however, would be processed and presented as a newly formed intracellular protein and targeted by ubiquitination for destruction in the proteasome with subsequent loading into MHC class I molecules. Although these processing pathways are not mutually exclusive, qualitative differences are likely to exist for the processing and presentation of different G protein species. In support of this explanation, we found no increase in CD8+ T cell numbers or in IFN-c production by cells in Gmem rRSV-challenged mice (not depicted). Second, the central conserved region of G shows sequence and structural homology to a component of the TNF receptor [24], raising the possibility that secreted G could bind TNF or similar molecules and modify the immune response to infection. These observations suggest that the RSV G protein has evolved to interact with the immune system and produce effects on the host response that are advantageous to the virus. Third, G has intriguing homologies to the CX3C chemokine fractalkine; both have transmembrane and cytoplasmic domains plus a serine/threonine-rich mucin-like stalk connected to a highly conserved non-glycosylated region containing four cysteines, and have secreted form and membrane-bound form. It has recently been shown that RSV G protein binds to the human fractalkine receptor and has some fractalkine-like activity in vitro [25]. The engagement of secreted G with CX3CR1 on CD4 T cells may affect cell recruitment and activation [25]. Activated CD4+ T cells are required for lung eosinophilia in RSV-infected mice [19], and a reduction in CD4+ T cell number or activation, by depletion or interference with pro-inflammatory cytokines, prevents eosinophilic lung inflammation [26–28]. If secreted G recruits and/or activates CD4+ T cells, Gmem rRSV infection would lead to reduced inflammation with a reduction in chemokine expression and cell recruitment. Reduced numbers of activated CD4+ T cells in mice infected with Gmem would contribute to reduced disease severity. It is interesting to note that CCL2 and CCL5 mRNA were reduced in lung samples from mice infected with Gmem
rRSV. Clearly, this could be simply due to reduced infectivity and replication, but this seems unlikely because rRSV was also attenuated in its replication, yet induced abundant CCL5 and CCL2 was found. Furthermore, reduced chemokine expression in the absence of secreted G protein is unlike that described for an attenuated cold passaged RSV strain (CP52, lacking G and SH proteins [14]), indicating that the absence of soluble G alone has a specific modulating effect on inflammatory responses in the lung. The induction of CCL5 during RSV infection of both children and mice has been extensively documented [29,30]. In RSV bronchiolitis, dsRNA intermediates may induce CCL5 secretion from airway epithelial cells, which may predict the later development of recurrent wheeze [31,32]. A reduction of CCL5 by antiCCL-5 [33] or Perflubron [13] treatment reduces inflammation and/or airway hyperreactivity. Since CCL5 recruits and activates monocytes, lymphocytes and eosinophils, a reduction of this chemokine during Gmem rRSV infection may explain the improved clinical outcome in the current study. In conclusion, improved knowledge of the immunobiology of RSV disease and the development of new methods of constructing mutant viruses now enable the production of new live vaccines against RSV disease that induce strong protective immune responses but lack the pathogenic potential of natural RSV variants. Such novel vaccine candidates should finally overcome the barriers to empirical RSV vaccine development.
Acknowledgments Supported by Welcome Trust Program grants 054797 and 071381 and MRC (UK) Case Award G78/5996. We thank Dr. Peter Collins, NIAID, Bethesda, USA for generously providing the cDNA plasmid, reagents and expertise.
References [1] [2]
[3]
[4]
[5] [6]
C.B. Hall, Respiratory syncytial virus and parainfluenza virus, New Engl. J. Med. 344 (2001) 1917–1928. K. McIntosh, J.M. Fishaut, Immunopathologic mechanisms in lower respiratory tract disease of infants due to respiratory syncytial virus, Prog. Med. Virol. 26 (1980) 94–118. T.E. Sparer, S. Matthews, T. Hussell, A.J. Rae, B. García-Barreno, J.A. Melero, P.J.M. Openshaw, Eliminating a region of respiratory syncytial virus attachment protein allows induction of protective immunity without vaccine-enhanced lung eosinophilia, J. Exp. Med. 187 (1998) 1921–1926. A. Bukreyev, M.H. Skiadopoulos, J. McAuliffe, B.R. Murphy, P.L. Collins, A.C. Schmidt, More antibody with less antigen: can immunogenicity of attenuated live virus vaccines be improved? Proc. Natl. Acad. Sci. USA 99 (2002) 16987–16991. P.J.M. Openshaw, F.J. Culley, W. Olszewska, Immunopathogenesis of vaccine-enhanced RSV disease, Vaccine 20 (2001) s27–s31. G.P. Bembridge, J.A. Lopez, R. Bustos, J.A. Melero, R. Cook, H. Mason, G. Taylor, Priming with a secreted form of the fusion protein of respiratory syncytial virus (RSV) promotes interleukin-4 (IL-4) and IL-5 production but not pulmonary eosinophilia following RSV challenge, J. Virol. 73 (1999) 10086–10094.
C.F. Maher et al. / Microbes and Infection 6 (2004) 1049–1055 [7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
T.R. Johnson, B.S. Graham, Secreted respiratory syncytial virus G glycoprotein induces interleukin-5 (IL-5), IL-13, and eosinophilia by an IL-4-independent mechanism, J. Virol. 75 (1999) 8485–8495. T.R. Johnson, J.E. Johnson, S.R. Roberts, G.W. Wertz, R.A. Parker, B.S. Graham, Priming with secreted glycoprotein G of respiratory syncytial virus (RSV) augments interleukin-5 production and tissue eosinophilia after RSV challenge, J. Virol. 72 (1998) 2871–2880. M.N. Teng, S.S. Whitehead, P.L. Collins, Contribution of the respiratory syncytial virus G glycoprotein and its secreted and membranebound forms to virus replication in vitro and in vivo, Virology 289 (2001) 283–296. P.L. Collins, M.G. Hill, E. Camargo, H. Grosfeld, R.M. Chanock, B.R. Murphy, Production of infectious human respiratory syncytial virus from cloned cDNA confirms an essential role for the transcription elongation factor from the 5′ proximal open reading frame of the M2 mRNA in gene expression and provides a capability for vaccine development, Proc. Natl. Acad. Sci. USA 92 (1995) 11563–11567. T. Hussell, U. Khan, P.J.M. Openshaw, IL-12 treatment attenuates Th2 and B cell responses but does not improve vaccine-enhanced lung illness, J. Immunol. 159 (1997) 328–334. W. Hallensleben, L. Biro, C. Sauder, J. Hausmann, V.C. Asensio, I.L. Campbell, P. Staeheli, A polymorphism in the mouse crg-2/IP-10 gene complicates chemokine gene expression analysis using a commercial ribonuclease protection assay, J. Immunol. Methods 234 (2000) 149–151. H.A. Haeberle, W.A. Kuziel, H.J. Dieterich, A. Casola, Z. Gatalica, R.P. Garofalo, Inducible expression of inflammatory chemokines in respiratory syncytial virus-infected mice: role of MIP-1alpha in lung pathology, J. Virol. 75 (2001) 878–890. R.A. Tripp, L. Jones, L.J. Anderson, Respiratory syncytial virus G and/or SH glycoproteins modify CC and CXC chemokine mRNA expression in the BALB/c mouse, J. Virol. 74 (2000) 6227–6229. H.V. Coates, D.W. Alling, R.M. Chanock, An antigenic analysis of respiratory syncytial virus isolates by a plaque reduction neutralization test, Am. J. Epidemiol. 83 (1966) 299–313. H. Jin, D. Clarke, H.Z. Zhou, X. Cheng, K. Coelingh, M. Bryant, S. Li, Recombinant human respiratory syncytial virus (RSV) from cDNA and construction of subgroup A and B chimeric RSV, Virology 251 (1998) 206–214. M.B. Elliott, K.S. Pryharski, Q. Yu, C.L. Parks, T.S. Laughlin, C.K. Gupta, R.A. Lerch, V.B. Randolph, N.A. LaPierre, K.M. Dack, G.E. Hancock, Recombinant respiratory syncytial viruses lacking the C-terminal third of the attachment (G) protein are immunogenic and attenuated in vivo and in vitro, J. Virol. 78 (2004) 5773–5783. P.J.M. Openshaw, S.L. Clarke, F.M. Record, Pulmonary eosinophilic response to respiratory syncytial virus infection in mice sensitized to the major surface glycoprotein G, Int. Immunol. 4 (1992) 493–500. W.H. Alwan, W.J. Kozlowska, P.J.M. Openshaw, Distinct types of lung disease caused by functional subsets of antiviral T cells, J. Exp. Med. 179 (1994) 81–89.
1055
[20] J. Schwarze, U. Schauer, Enhanced virulence, airway inflammation and impaired lung function induced by respiratory syncytial virus deficient in secreted G protein, Thorax 59 (2004) 517–521. [21] G.W. Wertz, M. Krieger, L.A. Ball, Structure and cell surface maturation of the attachment glycoprotein of human respiratory syncytial virus in a cell line deficient in O glycosylation, J. Virol. 63 (1989) 4767–4776. [22] T. Hussell, P.J.M. Openshaw, IL-12 activated NK cells reduce lung eosinophilia to the attachment protein of respiratory syncytial virus but do not enhance the severity of illness after lung challenge in CD8 cell-immunodeficient conditions, J. Immunol. 165 (2000) 7109–7115. [23] A. Srikiatkhachorn, T.J. Braciale, Virus specific CD8+ T lymphocytes downregulate T helper cell type 2 cytokine secretion and pulmonary eosinophilia during experimental murine respiratory syncytial virus Infection, J. Exp. Med. 186 (1997) 421–432. [24] J.P. Langedijk, B.L. de Groot, H.J. Berendsen, J.T. Van Oirschot, Structural homology of the central conserved region of the attachment protein G of respiratory syncytial virus with the fourth subdomain of 55-kDa tumor necrosis factor receptor, Virology 243 (1998) 293–302. [25] R.A. Tripp, L.P. Jones, L.M. Haynes, H. Zheng, P.M. Murphy, L.J. Anderson, CX3C chemokine mimicry by respiratory syncytial virus G glycoprotein, Nat. Immun. 2 (2001) 732–738. [26] T. Hussell, A. Pennycook, P.J.M. Openshaw, Inhibition of tumour necrosis factor reduces the severity of virus-specific lung immunopathology, Eur. J. Immunol. 31 (2001) 2566–2573. [27] G. Walzl, S. Matthews, S. Kendall, J.C. Gutiérrez-Ramos, A.J. Coyle, P.J.M. Openshaw, T. Hussell, Inhibition of T1/ST2 during respiratory syncytial virus infection prevents Th2- but not Th1-driven immunopathology, J. Exp. Med. 193 (2001) 785–792. [28] T. Hussell, C.J. Baldwin, A. O’Garra, P.J.M. Openshaw, CD8+ T-cells control Th2-driven pathology during pulmonary respiratory syncytial virus infection, Eur. J. Immunol. 27 (1997) 3341–3349. [29] Y. Zhang, B.A. Luxon, A. Casola, R.P. Garofalo, M. Jamaluddin, A.R. Brasier, Expression of respiratory syncytial virus-induced chemokine gene networks in lower airway epithelial cells revealed by cDNA microarrays, J. Virol. 75 (2001) 9044–9058. [30] R.A. Tripp, D. Moore, A. Barskey, L. Jones, C. Moscatiello, H. Keyserling, L.J. Anderson, Peripheral blood mononuclear cells from infants hospitalized because of respiratory syncytial virus infection express T helper-1 and T helper-2 cytokines and CC chemokine messenger RNA, J. Infect. Dis. 185 (2002) 1388–1394. [31] H.L. Chung, S.G. Kim, RANTES may be predictive of later recurrent wheezing after respiratory syncytial virus bronchiolitis in infants, Ann. Allergy Asthma Immunol. 88 (2002) 463–467. [32] J.E. Gern, W.W. Busse, Relationship of viral infections to wheezing illnesses and asthma, Nat. Rev. Immunol. 2 (2002) 132–138. [33] K.K. Tekkanat, H. Maassab, A. Miller, A.A. Berlin, S.L. Kunkel, N.W. Lukacs, RANTES (CCL5) production during primary respiratory syncytial virus infection exacerbates airway disease, Eur. J. Immunol. 32 (2002) 3276–3284.