|Home | About | Journals | Submit | Contact Us | Français|
West Nile virus (WNV), a mosquito-borne, single-stranded RNA flavivirus, causes significant human morbidity and mortality in the older population; thus, we investigated the effects of aging on infection with WNV in dendritic cells (DCs). We infected DCs with WNV in vitro and quantified cytokines and chemokines (type I IFN and CXCL10), pathogen recognition receptors RIG-I, and Toll-like receptors 3 and 7. The production of type I IFN was significantly lower in DCs from older donors, compared with younger donors. Although we observed no significant age-related difference in expression or nuclear translocation of signaling molecules in initial antiviral responses, DCs from older donors have diminished induction of late-phase responses (eg, STAT1, IRF7, and IRF1), suggesting defective regulation of type I IFN. Our results identify deficits in critical regulatory pathways in the antiviral response that may contribute to the enhanced susceptibility to viral infections observed in aging.
West Nile virus (WNV) is a mosquito-borne, enveloped, positive-strand RNA virus belonging to the family Flaviviridae, which includes Yellow fever and Dengue viruses . During 2005–2009, 12,975 cases were reported to the Centers for Disease Control and Prevention, including 496 fatalities. WNV infection in healthy humans are typically asymptomatic, but severe symptoms, which are more common in older patients (age, >55 years), include mengingoencephalitis and death [2, 3]. The multifactorial immune response to WNV includes adaptive responses via Treg and γδT cells and production of specific antibodies [4–7]. A critical role for innate immune responses has been shown by infection of mice depleted of macrophages or neutrophils or lacking key components of innate immunity, which exhibit higher and extended viremia and increased mortality [8–15].
Aging is associated with a progressive decrease in immune function, an increased susceptibility to infection, and decreased responsiveness to vaccines . In aging, humoral and cell-mediated immune responses of the adaptive immune system [17, 18] and innate responses  show a well-documented decrease. Our group showed age-related reduced expression of the pathogen recognition receptors, Toll-like receptors (TLRs), reduced TLR-signaling in primary monocytes and dendritic cells (DCs), which correlate with vaccine responsiveness, and a dysregulation of TLR3 in macrophages [20–22]. Here, we examine the effects of aging on responses to WNV infection of DCs, professional antigen-presenting cells that link innate and adaptive immunity .
Heparinized blood samples from healthy volunteers were obtained with written informed consent under the guidelines of the Human Investigations Committee of Yale University School of Medicine. Donors had no acute illness, took no antibiotics or nonsteroidal antiinflammatory drugs, and were negative for exposure to WNV infection . The mean age (± standard deviation [SD]) of young donors (n = 64) was 25.7 ± 2.8 years (range 21–33 years); 49% were female, 82% were white, 3% were black, 14% were Asian, and 1% were Hispanic. The mean age (± SD) of older donors (n = 64) was 73.6 ± 6.0 years (range 58–86 years); 43% were female, 84% were white, 14% were black, and 2% were Asian. Each donor was used only once, and because of limitations on sample size, it was not possible to assess each donor for each assay. Samples were randomly chosen for experiments >1 year for assays under study at the time of recruitment.
Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Hypaque (GE Healthcare) and purified by anti-CD14-conjugated magnetic microbeads (Miltenyibiotec). Monocyte-derived DCs (MDDCs) were derived after culture for 6 days in RPMI 1640 medium with 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL streptomycin, 50 ng/mL recombinant human granulocyte-macrophage colony-stimulating factor and 20 ng/mL recombinant human IL-4 (PeproTech) . Plasmacytoid DCs (pDCs) were enriched from PBMCs (~4-6 x105 pDCs/50 mL blood) by magnetic bead depletion of CD14- and CD19-positive cells and positive selection with anti-BDCA4 beads (Miltenyibiotec) .
Virulent WNV (CT-2741) provided by Dr. John Anderson, Connecticut Agricultural Experiment Station (New Haven, CT) , was passaged once in Vero cells (ATCC CCL-81), and viral plaque-forming units (PFUs) were quantified by plaque assays . WNV studies were conducted in a Biosafety Level 3 facility, licensed by the State of Connecticut and Yale University.
MDDCs (1.5 × 105 cells) were stained live at 4°C with surface markers CD1a-PE (eBioscience, CA), CD14-APC, HLA-DR-PerCP, CD80-FITC and CD86-FITC (BD Biosciences, CA), washed in PBS/2% FBS, fixed in 4% paraformaldehyde (PFA), and stored at -80°C for batched analysis in 90% FBS containing 10% DMSO. For blood pDC stimulation assays, PBMCs (1 × 106 cells/96 well) in RPMI/10% FBS and antibiotics were incubated for 4 h in medium alone or with TLR7 ligand R848 (10 μM; 3M Pharmaceuticals) with Brefeldin A (10μg/mL; Sigma) for the last 2 h. Cells were labeled with lineage cocktail (CD3, CD14, CD16, and CD19) APC-Cy7, HLA-DR-PerCP, CD11c-APC (BD Biosciences), and CD123-PE-Cy7 (eBioscience, CA); fixed; and stored at -80°C. For analysis, cells were thawed, permeabilized, and labeled for intracellular staining of IFN-α (Antigenix America) by fluorescence-activated cell sorter (FACS). Polystyrene compensation beads (BD CompBead; BD BioSciences) were used for compensation and normalization of instrument settings. Samples were acquired using an LSR II instrument (BD Biosciences) and analyzed using FlowJo software (Tree Star).
Total RNA was harvested from MDDCs with use of the RNeasy mini kit (Qiagen), and cDNA was synthesized using AffinityScript Multi Temperature cDNA Synthesis Kit (Stratagene). Primers and probes for qPCR were from Applied Biosystems or synthesized (Table S1). Amplification was performed in an iCycler (Bio-Rad) for 60 cycles with an annealing temperature at 60°C, and values were normalized to β-actin . qPCR assays over >1 year were batched together, and expression levels were determined in duplicate.
Supernatants were harvested and stored at -80°C, and IFN-α and IFN-β were quantified using ELISA of all samples together by batch analysis (PBL InterferonSource).
Infected and uninfected MDDCs were plated on glass cover slips, fixed, and permeabilized in 4% PFA/0.2% Triton-X at 37°C for 30 min and blocked in PBS/10% FBS for 1 h. Cells were labeled with rabbit anti-IRF3, IRF1, or STAT1 antibodies detected by Alexa 546 anti-rabbit antibody, and nuclei were counterstained with TO-PRO-3 (Molecular Probes) . Translocation was determined in ≥100 cells/sample by the equation [(% cells with transcription factors in the nucleus)/(total cells)] ×100. Images were collected using an LSM 510 confocal microscope (Carl Zeiss MicroImaging).
Total proteins from MDDCs were harvested for immunoblotting and were scanned for densitometric analysis, as described previously . For ChIP, protein-DNA complexes were cross-linked with 1% PFA, quenched with glycine, washed, and lysed (RIPA lysis buffer), and chromatin was sheared as described previously . Equal amounts of proteins were immunoprecipitated using anti-IRF1, complexes were cross-linked and digested with proteinase K, and DNA was purified (Qiagen). Input and output DNA were used in regular PCR and qPCR with primers for promoters of TLR7.
The Welch 2-sample t test was applied after log transformation to make the distribution more Gaussian, and the Bonferroni multiple testing correction was applied using the R statistical package (http://www.r-project.org/). A corrected P <.05 was considered to be statistically significant.
We investigated the effects of aging on antiviral responses of DCs, a key immune cell type conventionally challenging to study in humans because of limitations of sample size and frequency of only ~1% of PBMCs. Thus, we used primary monocytes differentiated into DCs as indicated by expression of CD1a and downregulation of CD14 . Derived DCs showed no significant age-related differences in purity of CD1a+/CD14- DC populations (young: 85.17% ± 2.00% vs older: 84.24% ± 2.06%; 20 persons/group; P = .74), showing that generation and differentiation of DCs from monocytes is intact in aging.
To assess the effects of aging on DC responses, we stimulated MDDCs with the model viral ligand, poly I:C, and analyzed cells by FACS for the expression of the DC maturation markers CD80, CD86, and HLA-DR. At baseline, the mean fluorescence intensities of these maturation markers were comparable between MDDCs of younger and older donors (Figure 1A, 1B). After 24 h, MDDCs from both age groups showed significant upregulation of CD80, CD86, and HLA-DR (Figure 1A; Figure 1B). However, the up-regulation of CD86 and CD80 was significantly lower in MDDCs from older donors (CD86 P < .01 and CD80 P < .05) (Figure 1A; Figure 1B) and was also observed when MDDCs from both age groups were infected with WNV for 48 h at MOI=1 (CD86 P < .05) (Figure 1C). In addition, we found that primary blood pDCs from older donors also expressed lower levels of CD86 and CD80 when PBMCs were stimulated with the TLR7 ligand R848 (10 μM) for 24 h (P < .05) (Figure 1D). These data suggest that maturation of DCs is partially altered in older subjects. Because CD86 and CD80 are critical for T cell costimulation, this may contribute to the reduced responses in aging.
When we examined the response of MDDCs to infection with WNV (MOI= 1, 5, and 10; 0–72 h), no significant age-related difference in viability was observed (24 h: 93% [range, 91%–95%]; 48 h: 86% [range, 83%–88%]; and 72 h: 65% [range, 61%–68%]). Viability of MDDCs in both age groups decreased at 72 h, and data from this time are not presented. Equivalent increases in MDDC maturation markers were noted by infections with MOI=1 and MOI=5, with MOI=1 chosen for optimal viability. MDDCs infected with WNV showed increased mRNA for cytokines and chemokines (Figure 2A, B), but we observed significantly lower induction of mRNA for type I interferon (IFN-β) in MDDCs from older individuals at 48 h (P < .01) (Figure 2B). The lower level of IFN-β did not appear to be attributable to differential viral replication, because quantification of WNV-E gene from MDDCs showed similar increases in viral replication for both age groups (Figure 2C) and similar viral loads in plaque assays of supernatants at 48 h (younger 2.13 ± .55 x106 PFU/mL vs older 2.09 ± .52 x106 PFU/mL; 8 subjects/group; P = .8905). The production of IFN-β likely combines effects of directly infected and neighboring uninfected cells, because activation was observed in both populations (Fig. S1) . Although levels of IL-12 production in MDDCs after poly I:C stimulation (Fig. S2A) and WNV infection (Fig. S2B) trended lower in older donors, the decreased values did not reach statistical significance, and no statistically significant differences were observed in levels of IL-10 between the 2 populations (Fig. S2). The decrease in type I IFN production from MDDCs of older donors was also significant at the protein level (IFN-α P < .001 and IFN-β P < .01) (Figure 2D) and, moreover, was observed when MDDCs were stimulated by the TLR3 agonist poly I:C (P < .05) (Figure 2E).
pDCs are thought to be the first line of antiviral defense and produce abundant type I IFN . Primary blood pDCs from older donors infected in vitro with WNV (MOI = 1 for 48 h) showed a significant age-associated decrease in type I IFN production, compared with younger donors, confirming our findings with derived DCs both by ELISA of WNV-infected pDCs (P < .01) (Figure 2F) and by intracellular cytokine staining of IFN-α by FACS in pDCs from PBMCs treated with TLR7 ligand R848 (P < .01) (Figure 2G). These data suggest that IFN production may be impaired in aging.
Induction of type I IFNs by viruses is mediated both by TLRs and the cytosolic helicase RIG-I. To assess whether their expression contributes to the reduced responses from older persons, MDDCs were infected with WNV and TLRs and RIG-I were quantified. Although infection with WNV led to a dramatic increase in the mRNA level of RIG-I, TLR3, and TLR7, the degree of up-regulation was reduced in MDDCs from older donors and reached significance for TLR7 (P < .01) (Figure 3C) and inversely correlated with age (Pearson r=-.5754; P = .0001) (Figure 3D). The trend is consistent with the age-associated differences in IFN production in DCs, and the significant correlation between the mRNA levels of IFN-β and TLR7 (Pearson r = .3622; P = .0234) suggests the importance of IFN production in the induction of the TLR7 pathway. In addition, we quantified expression levels by qPCR in MDDCs of TLR1, TLR2, TLR4, and TLR9, TLRs not considered to be critical for recognition of RNA viruses. We found that these TLRs were not upregulated after WNV infection, and we did not observe statistically significant differences between younger and older donors (Fig. S3).
To dissect mechanisms underlying the age-related defect in production of IFN, we assessed factors involved in the generation of IFN responses. In the early phase of viral infection, IFN-β is produced in an IRF3-dependent manner . We observed no statistically significant age-related difference in nuclear translocation of IRF3 after WNV infection (younger 7.6 ± 1.8% vs older 6.8 ± 1.2%; 6 subjects/group; P = .3725) or in total IRF3 compared by immunoblot analysis of MDDCs at baseline or after WNV infection (Figure 4).
The type I IFN secreted in the early phase is recognized by type I IFN receptor (IFNAR) for autocrine activation of its downstream signaling pathway, JAK - STAT . Although the expression of IFNAR1 and IFNAR2 increased from basal levels after infection with WNV, the magnitude of induction was comparable between age groups (data not shown), confirming preservation of early antiviral responses in MDDCs from aged donors. Although not exhaustive, these data indicate that early antiviral responses are largely intact in DCs of aging donors.
We next examined key components of the late phase of antiviral responses in DCs, including amplification of IFN responses via the signaling cascade of STAT1, STAT2, and IRF9 . No statistically significant differences were noted between untreated MDDCs at baseline or after 48 h (mock) of incubation for STAT1 or β-actin protein (densitometry of STAT1 mock 48h/0h; younger 1.07 ± .12 vs older 1.08 ± .08; 3 subjects/group; P = .9575). Although infection with WNV increased STAT1 protein in MDDCs from younger donors by 24 h, the corresponding increase in STAT1 in MDDCs from older donors was both significantly lower and delayed, as noted in analysis of total STAT1, activated phospho-STAT1 (Tyr701) and phospho-STAT1 (Ser727), and in the ratio of phosphorylated to total STAT1 (Figure 5B). These findings indicate that STAT1 expression in response to WNV infection is impaired with aging.
After activation of the JAK - STAT pathway, upregulation of IRF7 expression establishes a positive feedback loop that stimulates the later antiviral responses and IFN genes . To evaluate the effects of reduced levels of STAT1 in MDDCs of older donors, we quantified levels of IRF7 by immunoblot. No statistically significant age-related differences in IRF7 were noted between untreated MDDCs at baseline or after 48 h (mock) of incubation (densitometry of IRF7 mock 48h/0h; younger .93 ± .06 vs older .93 ± .04; 3 subjects/group; P = .9883). Although IRF7 was induced in MDDCs after infection (Figure 5C), MDDCs of older donors showed significantly lower levels of IRF7 (P < .05) (Figure 5D)P. Our data suggest an age-associated deficit in the IFN positive-feedback pathway, such as upregulation of IRF7, in response to WNV infection.
Another target of the JAK - STAT pathway is IRF1, which translocates into the nucleus and is a positive regulator of transcription of type I IFN. No statistically significant differences in IRF1 were noted between untreated MDDCs at baseline or after 48 h (mock) of incubation (IRF1 48 h mock to 0 h; younger .96 ± .04 vs older .93 ± .07; 3 subjects/group; P = .7521). Although we noted equivalent percentage of IRF1 nuclear translocation cells between the age groups after infection with WNV (Figure 6A), the magnitude of induction of IRF1 was decreased in MDDCs of older donors, analogous to the results for IRF7 (P < .05) (Figure 6B, C).
An important consequence of the transcriptional activity of IRF1 is induction of TLR7, because there is an IRF1 binding site located in the TLR7 promoter . To examine the role of IRF1 in induction of TLR7, MDDCs were incubated with poly I:C for 24 h and subjected to chromatin immunoprecipitation (ChIP) analyses using an antibody to IRF1 and PCR primers that include the IRF1 binding site in the TLR7 promoter. We observed differential binding of IRF1 in the presence and absence of poly I:C (Figure 6D). We quantified the binding site using qPCR and observed a 2.3-fold enrichment after ChIP in samples stimulated with poly I:C, compared with unstimulated (mock) samples (n = 3), indicating that IRF1 is a transcription factor for TLR7. The lower expression of IRF1 in DCs of older donors also suggests that IRF1 contributes to the reduction in TLR7 expression. These results indicate that age-associated defects in STAT1 expression led to reduced upregulation of the STAT1 target genes IRF7 and IRF1 and, consequently, diminish signaling efficiency of the positive-feedback regulation of type I IFN.
In addition to stimulation of antiviral mechanisms through IFN production, a successful antiviral immune response includes overcoming inhibitory pathways, such as TAM (Tyro3, Axl, and Mer) receptor tyrosine kinases, which can hijack the IFNAR/STAT1 signaling cassette to broadly inhibit both TLRs and TLR-induced cytokine receptor cascades . To assess negative regulation in age-related deficiencies in anti-WNV responses, we examined expression of Axl in DCs. After infection with WNV, MDDCs from older donors had higher expression of Axl than did MDDCs from younger donors (P < .05) (Figure 7A) and similar increases of other inhibitory molecules in the JAK - STAT pathway, including SOCS1 (P< .05) (Figure 7B) and PIAS1 (P< .01) (Figure 7B), suggesting that negative signaling is enhanced in aging.
Significant evidence links aging with the deleterious and cumulative effects of reactive oxygen species generated throughout the lifespan [32, 33], and Axl, which inhibits the inflammatory response, can be regulated and activated by oxidative stress in muscle cells . When we assessed the effect of oxidative stress on expression of Axl, we detected higher expression of Axl in MDDCs treated with H2O2 (P < .01) (Figure 7C) and concomitantly lower production of IFN-β in response to poly I:C stimulation (P < .01) (Figure 7D). Although not conclusive, these data suggest that oxidative stress, as noted in aging, may influence immune responses through the upregulation of Axl and subsequent inhibition of IFN-β production.
Induction of type I IFN is a key feature of the innate immune antiviral response, and we show here an age-related impairment in production of type I IFN by DCs from older donors after infection with WNV in vitro. This is consistent with our previous findings of an age-dependent reduction in TLR ligand-stimulated cytokine production by monocytes, macrophages, and blood DCs [20–22] and impairments in PI3K signaling, critical for type I IFN production , which has recently been shown by others to be impaired in aging in humans and mice [36, 37]. Induction of type I IFN-dependent genes occurs as a highly regulated process . Although we did not observe any age-related deficiencies in the initial IFN response (IRF3), we showed an impaired increase in STAT1, IRF7, and IRF1 in DCs from older donors infected with WNV, suggesting inhibition of the positive-feedback regulation of type I IFN. Our results with DCs are distinct from our previous studies in macrophages, which showed higher levels and increased nuclear translocation of STAT1 in the early responses (3 h) to WNV infection by older donors . It is tempting to speculate that the different age-related responses of the 2 cell types may reflect either impairment of different mechanisms or related or compensatory effects.
Murine studies of age-associated defects in IFN-α production in pDCs, consistent with our current findings, showed reduced upregulation of IRF7 in pDCs of aged mice that is related to oxidative stress . Oxidative stress inhibits IFN-α-induced antiviral gene expression by blocking the JAK–STAT pathway , and macrophages deficient in the antioxidant molecule myeloid HO-1 show reduced expression of IFN-β and of primary IRF3 target genes . Our study suggests a further role for oxidative stress in inhibition of IFN-β production by DCs through age-dependent upregulation of Axl.
Impairments in innate immunity in aging, such as the reduced DC functions shown here, may contribute to age-related defects in T cell immunity, which have been reported to play an important role in the susceptibility of old mice to WNV . We have shown here that DCs from older donors have reduced costimulatory markers after stimulation, which although modest, is consistent with our previous finding of lower costimulatory markers in older subjects; this correlated with responsiveness to vaccination against influenza . Induction of type I IFN is a key feature of DC function; thus, the deficiencies shown here, although only 2–4-fold lower in these populations in vitro, might be greatly amplified in vivo and impair responses at numerous levels, including maturation of DCs , upregulation of TLR7, stimulation of allogenic CD4+ and CD8+ T cells [43, 44], activation of natural killer cell , and enhancing mature lymphocyte survival and B cell class switching . In support of that is our previous result showing that lower levels of IFN-α production by DC correlate with reduced efficiency of vaccination in older donors .
Cells recognize and respond to RNA viruses through nucleic acid sensors, which activate downstream signaling, IFN-responsive genes, and establish a positive feedback loop. Deficiency of TLR3 or RIG-I in murine models or human viral infections shows impaired induction of host responses and enhanced viral replication [47, 48]. We have identified significant age-related reduction in upregulation of TLR7 in DCs after WNV infection and reduced production of IFN from pDCs stimulated with agonists of TLR7, consistent with our previous study showing increased susceptibility to WNV by TLR7-deficient mice . Reduction of the IFN-dependent induction of TLR7 may be critical to the reduced immune responses observed in older patients. Thus, triggering the TLR7 pathway, as has previously been shown to enhance host responses to hepatitis C virus  and reduction in HIV replication , may provide a potential therapeutic approach in WNV infection or in older patients.
This work was supported the National Institutes of Health (AI 50031, AI 070343, and the NCRR/GCRC Program M01-RR00125). EF is an Investigator of the Howard Hughes Medical Institute.
We thank Dr. John Anderson, for WNV strain CT-2741; Donna Carrano and Mary Lou Breitenstein, for valuable assistance; and the Yale IMAGIN team, for insightful discussions.