Interaction of severe acute respiratory syndrome-associated coronavirus with dendritic cells

A new form of life-threatening, atypical pneumonia called severe acute respiratory syndrome (SARS) has emerged recently as a human disease involving over 8000 cases and 774 deaths in 30 countries, mainly in the Republic of China (WHO, 2004). A novel virus most probably of zoonotic origin, termed SARS-associated coronavirus (SARS-CoV), was isolated from patients and identified as the aetiological agent (Drosten et al., 2003; Fouchier et al., 2003; Ksiazek et al., 2003; Kuiken et al., 2003; Peiris et al., 2003b). Until then, human coronaviruses had been found to be associated with mild upper respiratory-tract infections. SARS-CoV, however, leads to a severe clinical outcome and has been isolated from stool and urine samples, indicating systemic infection (Peiris et al., 2003a).

Autopsies of patients who died of SARS-CoV infection revealed severe alveolar damage of the lungs and heavy injury of the lymphatic tissue (Ding et al., 2003, 2004; Lang et al., 2003; Nicholls et al., 2003). The latter includes massive necrosis in the white pulps and the marginal sinus, destruction of germinal centres and apoptosis of lymphocytes, accompanied by an infiltration of monocytic cells. These changes are strong evidence that immunopathogenesis is driving the severe outcome of the disease.

Dendritic cells (DCs) are key regulators of immune responses (Banchereau & Steinman, 1998). Their main function is to sample antigens in various body tissues, to migrate to draining lymph nodes and to present antigens to cells of the specific immune system. DC maturation is triggered by inflammatory cytokines, such as tumour necrosis factor alpha (TNF-α) and interleukin 1 (IL-1), or by products of pathogens, such as lipopolysaccharide (LPS) or double-stranded RNA (Cella et al., 1999). In the T-cell zone of the lymphatic organs, DCs present antigens to T cells and initiate the specific immune response (Banchereau & Steinman, 1998; Cella et al., 1997). Especially when viruses do not replicate primarily in the lymphatic tissue, the host has to rely on the migratory capacity and the function of DCs to initiate an immune response. Thus, DCs are excellent targets for pathogens to impair the initial steps of the immune response in early infection (Rinaldo & Piazza, 2004). Immature DCs (iDCs) and mature DCs (mDCs) of the myeloid type can be differentiated in vitro from peripheral blood monocytes (Sallusto & Lanzavecchia, 1994) and they are a highly suitable in vitro model to study the interaction between DCs and viruses. As SARS-CoV induces strong damage of the lymphatic system (Ding et al., 2003; Lang et al., 2003), we speculated that DCs might play an important role in this process.

In this study, we show that SARS-CoV infects both iDCs and mDCs, but that virus replication occurs only at a low level. Furthermore, infection with SARS-CoV of iDCs and the fibroblast cell line 293 leads to a delayed expression of alpha interferon (IFN-α), indicating that SARS-CoV circumvents the activation of the innate immune system.

Encountering virulent or UV-inactivated SARS-CoV, DCs were activated, but lacked major histocompatibility complex (MHC) class I upregulation. This indicates that mechanisms to escape the adaptive immune system are involved in the pathogenesis of SARS-CoV infection.

Cells and viruses.
Vero E6 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) plus 10 % fetal calf serum (FCS) supplemented with 100 IU penicillin and 100 µg streptomycin ml¹.

For virus-stock generation, Vero E6 cells were grown in cell-culture flasks until they reached 80 % confluence. The growth medium was removed and the cells were inoculated with 0.01 m.o.i. SARS-CoV strain FFM-1 in 5 ml infection medium (DMEM, 2 % FCS, 20 mM HEPES). After incubation for 1 h at 37 °C, the virus inoculum was removed and replaced by regular growth medium. At 72 h post-infection, the virus supernatants were harvested and cell debris was removed by centrifugation (3000 g for 5 min at 4 °C). Virus stocks were stored at 80 °C and thawed immediately before use. Virus titres were determined by a standard plaque assay as described previously (Spiegel et al., 2004).

Generation and infection of iDCs and mDCs.
DCs were prepared from peripheral blood mononuclear cells (PBMCs) of healthy individuals as described by Sallusto & Lanzavecchia (1994). PBMCs were purified by Ficoll gradients (Pharmacia). The adherent-cell fraction was further purified by using anti-CD2 and anti-CD19 immunomagnetic beads (Dynal). iDCs were produced by culturing 5x10⁵ cells ml¹ in 90 % RPMI 1640 medium, 10 % FCS, 2 mM glutamine, 100 IU penicillin ml¹ and 100 µg streptomycin ml¹ for 7 days in the presence of 50 ng granulocytemacrophage colony-stimulating factor ml¹ (Leukomax; Novartis Pharma) and 500 U IL-4 (Cellgenix). For maturation, a mixture of 10 ng IL-1β ml¹, 1000 U IL-6 ml¹ (Promocell), 10 ng TNF-α ml¹ and 10 µg PGE₂ ml¹ (Sigma) was added for 24 h. The purity of the cell cultures was approximately 95 %, as determined by flow-cytometry analysis showing expression of CD1a^high and CD14^low [CD1afluorescein isothiocyanate (FITC), CD14phycoerythrin (PE); BD Pharmingen]. iDCs and mDCs were infected with SARS-CoV (m.o.i. of 5) by adding infectious supernatant of SARS-CoV-infected Vero cells to the growth medium.

RNA extraction, quantitative SARS-CoV Taqman RT-PCR, IFN-α and γ-actin RT-PCR.
For RNA extraction, SARS-CoV-infected, mock-infected and UV-inactivated SARS-CoV-incubated iDCs, mDCs and 293 cells were collected at the indicated time points after infection and RNA was isolated by TRIzol extraction (Invitrogen).

SARS-CoV Taqman RT-PCR was used to determine the viral load and to measure the increase of viral RNA. All Taqman assays were performed on 5 µl RNA extract, with 15 pmol primers (NCCORFP, 5'-TGCCTCTGCATTCTTTGGA-3'; NCCORRP, 5'-TAAGTCAGCCATGTTCCCG-3') and 10 pmol probe (NCCORP, FAM-5'-CACGCATTGGCATGGAAGTCACA-3'-TAMRA) (TIB MolBiol) in a final volume of 20 µl by using a Lightcycler RNA Master Hybridization Probes kit (Roche). The Taqman RT-PCR was performed at 61 °C for 20 min, 95 °C for 5 min and 45 cycles of 95 °C for 15 s and 60 °C for 30 s (Weidmann et al., 2004).

For IFN-α and γ-actin RT-PCR, 1 µg RNA of each sample was subjected to treatment with DNase I (MBI Fermentas) followed by reverse transcription with SuperScript II (Invitrogen) using random-hexamer primers (Amersham Pharmacia Biotech). Amplification reactions were performed with 4 µl aliquots of each reverse transcription reaction with 10 pmol primers specific for IFN-α (forward primer, 5'-TCCATGAGATGATCCAGCAG-3'; reverse primer, 5'-ATTTCTGCTCTGACAACCTCCC-3') detecting the multiple subtypes of IFN-α (Larrea et al., 2001) or primers specific for γ-actin (forward primer, 5'-GCCGGTCGCAATGGAAGAAGA-3'; reverse primer, 5'-CATGGCCGGGGTGTTGAAGGTC-3') (Sigma-Ark). The reaction mixtures were subjected to an initial denaturation step for 2 min at 94 °C. Then, 0.25 U recombinant Taq polymerase (Eppendorf) was added and 35 cycles of denaturation (94 °C for 30 s), annealing (56 °C for 1 min) and extension (72 °C for 1 min) were performed, followed by a final extension step at 72 °C for 10 min. The amplification products were separated on a 2 % agarose gel containing 50 ng ethidium bromide ml¹ and visualized by UV transillumination in a Chemidoc XRS imager (Bio-Rad).

Immunofluorescence microscopy.
DCs (5x10⁵) were harvested at different time points post-infection, washed in 5 ml PBS (Ca²⁺- and Mg²⁺-free) and fixed in 5 % paraformaldehyde for 10 min. Then, the cells were resuspended in 100 µl PBS and attached to SuperFrost Plus microscope slides (Shandon) by centrifugation for 2 min at 900 r.p.m. at high acceleration in a Cytospin 2 centrifuge (Shandon). Mouse mAb CMRF-56 (kindly provided by Derek Hart, Mater Medical Research Institute, Brisbane, Australia) was used for staining of DCs. For detection of viral nucleoprotein, cells were incubated with 1 : 1000-diluted anti-SARS-CoV N rabbit polyclonal antibody (Spiegel et al., 2005). Counterstaining for cell nuclei was performed with 1 : 200-diluted TO-PRO-3 iodide (Molecular Probes). After incubation for 1 h at room temperature in a humidified chamber, the cell samples were washed three times in PBS, followed by incubation with FITC-conjugated goat anti-mouse IgG1 and Cy3-conjugated donkey anti-rabbit IgG at a dilution of 1 : 200. The samples were again washed three times in PBS and then mounted by using FluorSave reagent (Calbiochem). Apoptotic cell death was monitored by TUNEL assay according to the manufacturer's instructions (Roche). Stained cell samples were examined by using a Leica confocal laser-scanning microscope with a x63 NA1.4 objective (detection of viral nucleoprotein) or a x10 objective (TUNEL assay).

Flow-cytometry analysis.
iDCs and mDCs, infected, mock-infected or incubated with UV-inactivated SARS-CoV, were collected at days 1, 4 and 6 after infection, washed in PBS and incubated with one or two of the following mAbs: anti-CD1aFITC (HI149; BD Pharmingen), anti-CD14PE (M5E2; BD Pharmingen), anti-CD40FITC (5C3; BD Pharmingen), anti-CD54FITC (84H10; Immunotech), anti-CD58PE (AICD58; Immunotech), anti-CD80FITC (BB1; BD Pharmingen), anti-CD83PE (HB15; Immunotech), anti-CD86PE (IT2.2; BD Pharmingen), anti-MHC class IPE (G46-2.6; BD Pharmingen) and anti-MHC class IIFITC (Tü39; BD Pharmingen). The samples were fixed with 5 % paraformaldehyde for 30 min before they were analysed on a FACSsort (Becton Dickinson) using CellQuest Pro software.

SARS-CoV replicates in iDCs and mDCs
iDC and mDC cultures were infected at an m.o.i. of 5. At day 3 post-infection, quantitative SARS-CoV Taqman PCR (Weidmann et al., 2004) was used to investigate virus replication in SARS-CoV-infected DCs. An increase of viral RNA molecules could be observed in SARS-CoV-infected iDCs and mDCs. No increase was documented in mock-infected controls and when UV-inactivated SARS-CoV was used (Fig. 1a).

(29K): Fig. 1. (a) Viral load in SARS-CoV-infected DCs as measured by quantitative real-time RT-PCR. Uninfected, SARS-CoV-infected and UV-inactivated SARS-CoV-incubated iDCs and mDCs (m.o.i. of 0.5) were harvested at day 3 post-infection. The RNA of 2x106 DCs was isolated and the number of SARS-CoV nucleocapsid RNAs was detected in a quantitative RT-PCR. The relative numbers of viral RNA of one representative experiment are given. Error bars show the intra-assay variation of the RT-PCR. (b) Replication of SARS-CoV in infected iDCs. SARS-CoV-infected iDCs were cultivated for 1, 4 and 6 days post-infection or were left untreated. Cytospin preparations of the infected cultures were used for detection of SARS-CoV N protein by immunofluorescence. Green, DC-specific 96 kDa early activation/differentiation antigen (mouse mAb CMRF-56 followed by FITC-conjugated goat anti-mouse IgG); red, SARS-CoV N protein (anti-SARS-CoV N rabbit polyclonal antibody followed by Cy3-conjugated donkey anti-rabbit IgG); blue, cell nuclei (TO-PRO-3 iodide).

In addition, virus replication was shown by immunofluorescence assays detecting expression of the viral N protein at days 16 post-infection, as outlined in Fig. 1(b). It should be noted, however, that N protein expression was reduced strongly at days 4 and 6 post-infection compared with day 1. To confirm that the infected cells were DCs, cells were additionally stained for the 96 kDa early activation/differentiation antigen, which is expressed specifically by different DC populations, including monocyte-derived DCs (Highton et al., 2000; Hock et al., 1999). Indeed, nearly all of the SARS-CoV-infected cells expressed the 96 kDa early activation/differentiation antigen, indicating that the infected cells were DCs (Fig. 1b).

Apparently, DCs are susceptible to SARS-CoV infection. To investigate whether they support production and release of progeny virus, we determined the titre of supernatants of SARS-CoV-infected DC cultures by a standard plaque assay (Spiegel et al., 2004). We obtained low but reproducible titres (around 100 p.f.u. ml¹) at day 6 post-infection, indicating low-level replication of SARS-CoV in DCs (Fig. 2). To rule out the possibility that the observed titres represented residual input virus, we determined in parallel the long-term stability of SARS-CoV by inoculating growth medium with virus stock and testing for infectivity at different time points. At day 6, no viral infectivity remained, whereas supernatants of SARS-CoV-infected DCs were still infectious (Fig. 2).

(9K): Fig. 2. Virus titres of DC supernatants and long-term stability of SARS-CoV. Titres of supernatants of SARS-CoV-infected iDCs (m.o.i. of 5) were determined at days 1, 4 and 6 post-infection (bars). In parallel, residual infectivity of SARS-CoV incubated in cell-culture medium was determined at the indicated time points (curve). Data are representative of three experiments with similar results.

As the titres obtained directly from supernatants of infected DCs were low, we additionally performed recovery experiments. To this aim, intact SARS-CoV-infected DCs, as well as supernatants and cell lysates of SARS-CoV-infected DCs, were collected at day 6 after infection. To detect infectious virus, Vero cells were co-cultivated with the collected DCs or were incubated with the DC supernatants or cell lysates. After 3 days, the supernatants of the indicator Vero cells were then tested for infectious virus by plaque assay. We could detect infectious virus in all three experimental settings, indicating that DCs are infected productively with SARS-CoV (Table 1). However, the viral titres obtained from the recovery experiments using DC supernatants were lower (10⁶ p.f.u. ml¹) than the viral titres obtained with DC lysate or whole DCs (10⁷ p.f.u. ml¹). In summary, these data suggest that productive virus replication occurred in DCs, albeit at a low level.

Table 1. Virus recovery from SARS-CoV-infected DCs iDCs were infected with SARS-CoV (m.o.i. of 5) or were left untreated. DCs, DC supernatant and DC lysate were collected at day 6 post-infection and recovery experiments were performed by using Vero cells as indicator cells. Titres of recovered virus were determined by plaque assay. , Not detectable; ++, virus titre 1x106 p.f.u. ml1; +++, virus titre 1x107 p.f.u. ml1.

Virus replication does not induce cell death of DCs
To investigate the consequences of SARS-CoV infection for iDCs, we monitored cell death by light microscopy and TUNEL assay. As shown in the upper panel of Fig. 3, no difference in cell numbers was observed at day 6 post-infection when mock-infected cells, infected cells and cells inoculated with UV-inactivated virus were compared. Furthermore, no signs of apoptosis were detected by TUNEL staining in the corresponding Cytospin samples (Fig. 3, lower panel), indicating that SARS-CoV replication does not induce apoptotic cell death in DCs. Interestingly, SARS-CoV-infected cultures and cultures incubated with UV-inactivated virus exhibited a higher number of adherent cells than non-infected cultures. This indicates that SARS-CoV may mediate activation and maturation of iDCs (see below).

(77K): Fig. 3. Survival and growth of iDCs following infection with SARS-CoV. Immature DCs were either infected with SARS-CoV, incubated with UV-inactivated SARS-CoV or leftuntreated (mock). Cells were assayed for viability by light microscopy (LM) and TUNEL staining at day 6 post-infection.

SARS-CoV infection induces IFN-α expression
Type I IFNs are key components of the innate immune system. They represent the first line of defence against viral infections and restrict the growth and replication of a number of viruses, including coronaviruses (Cinatl et al., 2003; Fuchizaki et al., 2003; Haagmans et al., 2004; Hensley et al., 2004; Kawamoto et al., 2003; Pei et al., 2001; Spiegel et al., 2004). Therefore, we investigated whether SARS-CoV infection of DCs activates the expression of IFN-α. iDCs were infected with SARS-CoV (m.o.i. of 5) and IFN-α expression was measured 24 and 48 h post-infection by using RT-PCR. Both treatment with UV-inactivated virus and infection with replicating virus led to the production of detectable levels of IFN-α-specific transcripts 24 h post-infection (Fig. 4, upper panel). Signals specific for IFN-α could no longer be observed 48 h post-infection, indicating the downregulation of IFN-α expression at later time points of infection. Similar experiments were performed with a clone of the IFN-competent fibroblast cell line 293, which supports efficient SARS-CoV replication (Spiegel et al., 2005). In contrast to the infection of DCs, infection of 293 cells with SARS-CoV led to a sustained activation of IFN-α expression, which was even more pronounced at 48 h post-infection (Fig. 4, lower panel). Taken together, these findings demonstrate that SARS-CoV infection activates IFN-α expression in both DCs and 293 cells; however, the activation is only transient in the case of SARS-CoV-infected DCs.

(44K): Fig. 4. IFN-α expression following SARS-CoV infection. Immature DCs and 293 cells were infected with SARS-CoV, incubated with UV-inactivated SARS-CoV (SARS-CoV/UV) or left untreated (mock) for 24 and 48 h. Total RNA was isolated and RT-PCRs with primers specific for γ-actin and IFN-α cDNA were performed as described in Methods.

Phenotype of DCs in SARS-CoV infection
To further investigate the interaction between DCs and SARS-CoV, we analysed the expression of antigen-presenting molecules (MHC class I, MHC class II and CD1a), costimulatory molecules (CD40, CD80 and CD86), adhesion molecules (CD54 and CD58), the maturation marker CD83 and the LPS receptor CD14 by flow-cytometry analysis. For mDCs, similar expression patterns were observed for uninfected cells and for cells either incubated with UV-inactivated SARS-CoV or infected with SARS-CoV at days 1, 4 and 6 after infection (data not shown). For iDCs, SARS-CoV infection had a clear effect. When compared with uninfected iDCs, enhanced expression of CD40, CD54, CD58, CD80, CD83, CD86 (Fig. 5b) and MHC class II (Fig. 5a) was detected at day 4 after infection, whereas no differences were observed for MHC class I, CD1a (Fig. 5a) or CD14 (Fig. 5b). Similar results were obtained at days 1 and 6 after infection (data not shown). Interestingly, iDCs incubated with UV-inactivated SARS-CoV displayed the same expression pattern as SARS-CoV-infected iDCs, indicating that productive infection is not necessary for the upregulation of surface-molecule expression. The expression pattern of iDCs that came into contact with SARS-CoV particles is typical of mature DCs and the lack of CD14 expression shows clearly that the cell populations analysed did not differentiate to macrophages during cell culture. Therefore, we conclude that SARS-CoV is able to activate iDCs. However, the missing upregulation of antigen-presenting MHC class I molecules indicates that the virus is able to impair the function of DCs.

(24K): Fig. 5. Flow-cytometry analysis of surface-marker expression on iDCs at day 4 after infection. The black lines represent the isotype control, the green lines the marker expression of uninfected iDCs, the red lines the marker expression of SARS-CoV-infected iDCs and the blue lines the marker expression of iDCs co-cultivated with UV-inactivated SARS-CoV. (a) Uninfected iDCs and mDCs, SARS-CoV-infected and UV-inactivated SARS-CoV-incubated iDCs (m.o.i. of 5) were stained with anti-human MHC class IPE, MHC class IIFITC or CD1aFITC. The light-blue line represents the MHC class I expression of mDCs. (b) Uninfected, SARS-CoV-infected and UV-inactivated SARS-CoV-incubated iDCs (m.o.i. of 5) were stained with anti-human CD40FITC, CD80FITC, CD86PE, CD54FITC, CD58PE, CD83PE or CD14PE.

SARS was the first emerging infectious disease of the 21st century. It was introduced from an animal reservoir and met an immunologically naïve human population. There is evidence that, despite severe lung injury, the virus hardly activates the immune system, leading to an impairment of immunity (Ding et al., 2003, 2004).

Here, we investigated whether DCs, the key players of antigen presentation, could be involved in SARS-CoV pathogenesis and whether virus-driven immune-escape mechanisms contribute to viral pathogenesis.

The replication of SARS-CoV in DCs was shown by detection of viral RNA and the expression of viral proteins using quantitative real-time RT-PCR and immunofluorescence assays, respectively. The results were confirmed by plaque assay using supernatants of infected DCs and recovery experiments using supernatants of infected DCs, DC cell lysate and co-culture experiments. Infectious virus could be detected in all experimental settings. Because viral titres were much lower when supernatants of DCs were tested directly for infectious virus compared with titres obtained in recovery experiments, we concluded that only a low-level replication of SARS-CoV occurred in DCs. To confirm that it was not just input virus that we recovered from the cells, we demonstrated that all input virus infectivity was destroyed at the time point that our titration experiments were performed. In line with our results, a recent study reported low SARS-CoV titres (10² TCID₅₀ ml¹) in supernatants of infected DCs 5 days post-infection, whereas no virus at all could be detected in supernatants of infected macrophages (Tseng et al., 2005). However, we cannot formally rule out the possibility that DCs may stabilize input virus instead of supporting complete replication. Nevertheless, we think that this is less likely, as we could detect viral RNA at day 3 post-infection, as well as expression of viral nucleoprotein in SARS-CoV-infected DCs up to day 6 post-infection.

It may well be that SARS-CoV uptake into DCs is mediated by macropinocytosis. The functional SARS-CoV receptor ACE-2, described recently (Li et al., 2003), is not expressed on DCs and thus cannot be involved (Hofmann & Pohlmann, 2004; Law et al., 2005). Our results suggest that other receptor molecules are involved in virus uptake. The entry of SARS-CoV into DCs may be mediated through C-type lectins, such as CD209 (DC-SIGN), CD209L (L-SIGN) or CD206 (mannose receptor) (Jeffers et al., 2004; Marzi et al., 2004; Yang et al., 2004). The S protein of SARS-CoV contains mannose structures (Han et al., 2004) and retroviral vectors pseudotyped with SARS-CoV S protein can enter DCs via CD209 (Yang et al., 2004). Therefore, certain C-type lectins might serve as an alternative receptor for the cellular entry of SARS-CoV, which has indeed been shown for CD209L (Jeffers et al., 2004).

The human coronavirus 229E, which is related to SARS-CoV, is known to induce apoptosis in monocytes/macrophages (Collins, 2002), For SARS-CoV, however, we could not observe any cell death of infected DCs during an infection period of 6 days. Thus, the immune dysfunction observed in SARS-infected humans is probably not due to SARS-CoV-mediated cell death of DCs.

Viruses have acquired many different mechanisms to escape the immune attack of the host (Alcami & Koszinowski, 2000; Beck et al., 2003; Weber et al., 2004). Here, we investigated whether SARS-CoV has developed immune-evasion mechanisms to modulate the innate and the specific immune responses. Surprisingly, treatment of iDCs and 293 cells with UV-inactivated virus was sufficient to induce IFN-α, suggesting that SARS-CoV replication is not necessary for the activation of IFN-α expression. Replication-competent virus, however, induced a stronger IFN-α signal in both iDCs and 293 cells. Interestingly, induction of IFN-α in iDCs at 24 h post-infection was no longer present at 48 h post-infection. In contrast, in SARS-CoV-infected 293 fibroblast cells, the induced IFN-α expression persisted for at least 48 h post-infection, but the virus replicated in these cells to high titres, despite IFN-α expression. This might be explained by different kinetics of virus replication versus IFN-α expression. We have shown previously that SARS-CoV infection of 293 cells does not induce IFN-β for up to 16 h after infection (Spiegel et al., 2005) and the same applies for IFN-α (unpublished data). As the virus replication cycle of SARS-CoV is completed in approximately 6 h (Ng et al., 2003), the virus simply appears to replicate to high titres well before type I IFNs are induced. Thus, IFN-α expression, presumably induced by replicating virus, occurs too late to hamper efficient virus production. Whether this delayed activation of type I IFN expression requires the action of a virus-encoded IFN antagonist remains to be determined.

SARS-CoV infection of DCs is much less effective than infection of 293 cells, probably due to the lack of the authentic SARS-CoV receptor ACE-2. The induction of IFN-α at 24 h post-infection appears to be sufficient to restrict virus growth in DCs, which is in line with the observation of a strongly reduced expression of viral nucleoprotein at days 4 and 6 post-infection. Apparently, the restriction of virus growth leads to the downregulation of IFN-α observed at 48 h post-infection. Nevertheless, SARS-CoV was able to infect DCs productively, as demonstrated by the successful transfer of infectious virus into susceptible Vero cells. In summary, SARS-CoV has developed mechanisms to induce a delayed response of the innate immune system in both 293 fibroblast cells and DCs, which allows the production of infectious progeny virus in both cell types.

Immature DCs undergo maturation and migrate to lymphatic tissue after uptake of pathogens or antigen. As SARS-CoV replicates in these cells, iDCs may play a key role in promoting viral dissemination within the host, offering a shuttle for the virus to enter the lymphatic tissue. This might contribute to the severe damage seen in the lymphatic tissues obtained from SARS-CoV-infected subjects (Ding et al., 2003).

To study the influence of SARS-CoV infection on DC function, we analysed the expression of antigen-presenting molecules, costimulatory molecules, adhesion molecules and maturation markers. With the marked exception of MHC class I upregulation, SARS-CoV-infected iDCs were clearly activated. In contrast, two other studies (Law et al., 2005; Ziegler et al., 2005) reported the lack of enhanced CD83, CD86 and MHC class II expression in SARS-CoV-infected DCs and it was postulated that an abortive SARS-CoV infection might prevent the activation of iDCs (Ziegler et al., 2005). Productive replication, however, is not required, as we observed activation of iDCs even when UV-inactivated SARS-CoV was used and similar results were obtained for DCs treated with γ-irradiated SARS-CoV (Tseng et al., 2005).

Interestingly, neither replication-competent nor UV-inactivated SARS-CoV induced MHC class I surface expression. As virus replication is not a prerequisite for the inhibition of MHC class I upregulation, one may speculate that high viraemia per se may enhance the impairment of antigen-presenting cells. This may be due to a bystander effect mediated by viral antigen. This hypothesis could explain in part the immune dysfunction seen in the course of human SARS-CoV infection. Indeed, a lack of MHC class I upregulation, together with a complete lack of cytokine expression, was observed when PBMCs derived from SARS patients were analysed (Reghunathan et al., 2005). The molecular mechanisms driving the inhibition of MHC class I upregulation remain to be elucidated.

Taken together, our studies demonstrated that SARS-CoV has the ability to circumvent the innate as well as the adaptive immune system. The transport of virus to the lymphatic tissue by infected DCs followed by the infection of susceptible target cells might play a crucial role in the impairment of the immune response seen in SARS patients. Identification of the underlying mechanisms may help to develop effective strategies for the treatment of SARS.

We thank Stephan Becker from the University of Marburg, Marburg, Germany, for the generous gift of SARS-CoV isolate FFM-1, Adolfo García-Sastre and Luis Martínez-Sobrido from the Mount Sinai School of Medicine, New York, USA, for kindly providing anti-SARS-CoV N rabbit polyclonal antibody and Derek Hart from the Mater Medical Research Institute, Brisbane, Australia, for kindly providing monoclonal mouse antibody CMRF-56.

Footnotes

^,†, Kerstin Schneider² †These authors contributed equally to this work.