Increased Reactivity of Dendritic Cells from Aged Subjects to Self-Antigen, the Human DNA¹

Immunosenescence is characterized by a decline in immune functions and chronic inflammation (1, 2). They are also the major causes of morbidity and mortality associated with increased age. Atherosclerosis, diabetes, Parkinson’s disease, Alzheimer’s disease, and arthritis, which are characterized by a marked inflammatory component, are diseases prevalent in the elderly (3). The mechanisms underlying the decline in immune functions and chronic inflammation associated with aging are poorly understood. Accumulating evidences suggest a critical role of the innate immune cells, such as dendritic cells (DCs),³ in generating inflammatory and tolerogenic responses.

DCs are the key regulators of the immune system, coordinating both immunity and tolerance. They are present at various portals of entry of pathogens like skin, airways, etc., forming a network of APCs, ready to initiate and amplify immune responses. DCs present at these sites capture Ags, migrate to the lymphoid organs, and mature en route. During maturation, they lose their Ag-capturing capacity and up-regulate the expression of MHC and costimulatory molecules, thus becoming the efficient APCs (4). Under steady-state conditions, immature DCs continuously sample the self-Ags from apoptotic cells in the periphery leading to T cell tolerance without costimulation. However, upon stimulation with microbial products, inflammatory cytokines, or CD40 ligation, DCs begin to mature and migrate to the draining lymph nodes. Maturation of DCs is associated with phenotypic changes, including up-regulation of costimulatory molecules and MHC, and secretion of IL-12, transforming them into fully functional APCs (4, 5). These fully matured DCs are capable of stimulating T and B cells. DCs are thus critical for immune defense and any alteration in DC functions with age would compromise the proper functioning of the immune system.

The majority of prior studies in both mice (6, 7) and humans (8, 9) indicate that the immune responses are preserved in aged DCs. However, virtually nothing is known about the age-associated alterations in the capacity of DCs to maintain tolerance. Our previous studies had revealed that DCs from aged subjects are impaired in their capacity to phagocytose apoptotic cells when compared with DCs from young subjects (10). This would result in the accumulation of apoptotic cells leading to secondary necrosis and release of self-Ags such as DNA. Earlier studies from our laboratory have also shown that DCs from aged humans secrete increased quantities of proinflammatory cytokines in response to TLR ligands (10). This tendency of DCs from aged subjects to secrete enhanced levels of proinflammatory cytokine may not be restricted to foreign Ags but may also exist for self-Ags, which will contribute to the age-associated chronic inflammation. In the present study, we have compared the reactivity of DCs from aged and young subjects for a self-Ag, the human DNA. Our data show increased reactivity of aged DCs to human DNA.

Blood was collected from healthy elderly (65–90 years of age) and young (20–35 years of age) volunteer donors. Elderly subjects were of middle-class socio-economic status and are living independently. An extensive medical history was obtained to exclude individuals with any major diseases. None of these volunteers had any significant medical illness. Young, healthy subjects, matched for the gender were drawn from students, staff, and blood donors at the University of California, Irvine. The Institutional Review Board of the University of California, Irvine, approved this study.

DNA was isolated from human blood using Qiamp DNA Blood Midi Kit from Qiagen. RNase was added to remove RNA contamination. Purity and yield of DNA was measured by UV spectrophotometer. Preparations with 260/280 ratio above 1.9 were used for the experiments. The preparation was free of endotoxin contamination as determined by Limulus amebocyte lysate test (Lonza).

Monocyte-derived DCs from aged and young subjects were prepared using GM-CSF and IL-4 essentially as described (10, 11). Such DCs are of myeloid origin. Transfection reagent Lipofectamine 2000 (Invitrogen) was used to deliver human DNA to DCs. DNA was mixed with lipofectamine (lipo) in 100 μl of Opti-MEM for 20 min at room temperature, according to the protocol recommended by the manufacturer and added to 4 × 10⁵ DCs in 300 μl of RPMI 1640 medium containing 10% FCS. Final concentration of the DNA was 1 μg/ml. Cell viability was unaffected by this treatment. Twenty-four hours later, cells were harvested and stained for surface markers CD40, CD80, CD86, CD83, and HLA-DR using directly conjugated Abs (BD Pharmingen). Analysis was performed using the FlowJo software (TreeStar).

IFN-α (PBL Biomedicals) and IL-6 (BD Pharmingen) in the supernatants were measured by specific ELISA kits as per the manufacturer’s protocol.

PBMCs from healthy individuals were serum-starved overnight. The next day, apoptosis was induced by exposing the cells to UV irradiation, at a dose of 80 mJ/cm² in a UV cross linker (Fisher Scientific). Irradiated cells were cultured in serum-free medium at 37°C in 5% CO₂ for another 24 h. At this time, ∼75% of the cells were in late apoptotic stage and 25% were early apoptotic as determined by Annexin V and propidium iodide staining. DCs from aged and young subjects were incubated with these apoptotic cells for 24 h. The up-regulation of costimulatory markers was determined by flow cytometry as described above. Supernatants collected were assayed for cytokines IL-6, TNF-α, and IFN-α by ELISA.

The expression of phosphorylated p65 unit of NFκB in DCs was determined by FACS as described (10). DCs from aged and young subjects were stimulated for 0 and 3 h with DNA plus Lipo, and fixed in phosflow buffer I (BD Biosciences) for 10 min at 37°C. Permeabilization was done with freshly prepared 90% ice-cold methanol for 30 min on ice and then cells were stained with the specific Ab for p65.

DCs from aged and young subjects were stimulated with DNA for 3 h. Cytoplasmic and nuclear fractions were prepared using the nuclear extract kit (Active Motif). Protein concentration was estimated using the Bio-Rad reagent. IRF-3 and NFκB-p65 binding to DNA to activate transcription was determined in the nuclear fraction using Trans AM ELISA kits (Active Motif) as per the manufacturer’s protocol. The results were normalized to protein levels.

DNA was labeled with Alexa Fluor 594 using ARES DNA Labeling Kit (Invitrogen). Labeled DNA (1 μg/ml) plus Lipo were added to DCs grown on poly l-lysine-coated coverslips for 3–18 h. DCs were permeabilized (PBS/1% FBS/0.1% saponin) and stained with FITC-labeled anti-lysosome-associated membrane protein 2 (LAMP-2; BD Biosciences) Abs.

Statistical analysis was done by using Graph Pad Prism software. Student’s t test or Wilcoxon signed rank test was used to measure significance. Values of p < 0.05 were considered significant.

Our earlier studies led us to hypothesize that increased reactivity of DCs from aged subjects to self-Ags may be one of the contributing factors to age-associated chronic inflammation. In the present study, we compared the reactivity of DCs from aged and young subjects to human DNA as a model self-Ag. In brief, human DNA (1 μg/ml) was delivered inside the human myeloid DCs from aged and young subjects using Lipofectamine 2000 (lipo) (Invitrogen). This dose of DNA was optimal for activation of DCs. Studies with DCs from both aged and young subjects show that higher concentrations of DNA is toxic while lower doses led to suboptimal activation as determined by secretion of IFN-α and IL-6 (Supplementary Fig. 1).⁴ Intracellular delivery of DNA led to activation of DCs as determined by up-regulation of costimulatory molecules CD80 and CD86 and secretion of cytokines IFN-α and IL-6 (Fig. 1, A and B). CD40, HLA-DR, and the maturation marker CD83 did not display significant (p > 0.05) up-regulation in response to intracellular human DNA. More importantly, DCs from aged subjects displayed increased activation compared with DCs from young subjects. The up-regulation of costimulatory molecules CD80 and CD86 was significantly higher (p < 0.05) in DCs from aged subjects when compared with DCs from young subjects (Fig. 1,A). The secretion of IFN-α (p = 0.0002) and IL-6 (p = 0.04) was also significantly higher in DCs from aged subjects as compared with DCs from young subjects in response to intracellular human DNA (Fig. 1 B). There was no secretion of TNF-α, IL-12p40, IL-12p70, and IL-10 as determined by ELISA. Introduction of lipo alone did not activate DCs. Stimulatory activity of the DNA was lost when delivered without lipo suggesting that exposure to DNA alone does not induce activation of DCs and that this requires its intracellular delivery. In summary, intracellular mammalian DNA led to partial maturation of DCs and induced the secretion of IFN-α and IL-6. Remarkably, DCs from aged subjects showed significantly increased reactivity toward human DNA compared with DCs from young subjects.

Under normal physiological conditions, apoptotic cells are rapidly and efficiently cleared by macrophages and DCs without induction of inflammation. However, inefficient clearance of apoptotic cells, as observed in aging, leads to accumulation of late apoptotic cells that undergo secondary necrosis resulting in the release of self Ags such as heat shock proteins (HSPs) and DNA that may gain entry into the DCs to activate them. Intracellular delivery of self, human DNA by transfection to DCs was used to mimic this physiological condition. However, to confirm whether the increased reactivity of DCs from aged subjects to self-Ags exists even when DCs are exposed to them under physiological conditions, DCs from aged and young subjects were incubated with the late-phase apoptotic cells at a ratio of 1:2 for 24 h. This led to a significantly higher up-regulation (p < 0.05) of CD80 and CD86 on DCs from aged subjects as compared with DCs from young subjects (Fig. 2,A). This increased activation of DCs from aged subjects was also reflected in cytokine levels. The level of cytokines IFN-α, IL-6, and TNF-α was significantly higher (p < 0.05) in the supernatants of DCs from aged subjects exposed to late apoptotic cells as compared with supernatants from exposed DCs from young subjects (Fig. 2 B). These findings suggest that DCs from aged subjects are more reactive to self-Ags than DCs from young subjects and their increased reactivity to human DNA is not an artifact of the artificial delivery system.

Next, we determined whether these human DNA-activated DCs were capable of inducing proliferation of T cells. For this purpose, DCs from aged and young individuals were activated with human DNA for 24 h as described in Fig. 1. DCs were washed and cultured with allogeneic, magnetic bead purified CFSE-labeled T cells from young subjects for 5 days and proliferation was determined by measuring the dilution of CFSE dye by flow cytometry. Titrating doses of DCs (10,000–100) were used for the experiments. DCs from young subjects induced significantly higher (p < 0.05) level of T cell proliferation compared with DCs from aged subjects in both stimulated and unstimulated groups. However, human DNA-stimulated DCs from aged subjects induced significantly higher (p < 0.05) proliferation of T cells (∼4% higher) as compared with unstimulated DC and lipo-stimulated DCs without DNA from the aged subjects (Fig. 3,A). In contrast, human DNA-stimulated DCs from young subjects did not increase T cell proliferation relative to unstimulated young DC controls (Fig. 3,A). Secretion of IFN-γ correlated with the proliferation data. Akin to proliferation, IFN-γ levels were overall lower in the aged subjects compared with the young subjects in all groups (Fig. 3,B). However, IFN-γ levels were significantly higher (p < 0.05) in the Lipo plus DNA-stimulated group compared with unstimulated and lipo stimulated DCs in the aged subjects (Fig. 3,B). Similar to proliferation, IFN-γ levels were comparable in DNA stimulated and unstimulated DCs from young subjects (Fig. 3 B) at all doses of DCs. These data further substantiate our findings that DCs from aged subjects are more activated compared with young subjects in response to human DNA.

To investigate the mechanisms responsible for the increased reactivity of DCs from aged subjects to human DNA, first we determined the distribution of the intracellular DNA in the DCs. Red-labeled DNA was found to localize in the cytosol after delivery. The picture was similar at all time points tested (3 h, 8–18 h, Fig. 4,A). No colocalization of DNA was observed with the lysosomal marker, LAMP-2 suggesting that DNA does not enter the lysosomes (Fig. 4 A).

The presence of DNA in the cytosol suggests that it may be using a cytosolic sensor of DNA. Recent studies have identified DAI (DNA-dependent activator of IFN-regulatory factors, also known as DLM-1/ZBP-1) as a cytosolic receptor that senses both foreign and self DNA (12, 13). We determined whether age-associated increased expression of this receptor is responsible for the increased reactivity of DCs from aged subjects to human DNA. The expression of DAI in DCs was comparable between the aged and young individuals both at the gene and protein levels (Fig. 4, B and C).

Next, we investigated the signaling pathways activated in DCs in response to human DNA. Recent reports suggest an involvement of IRFs particularly IRF-3 (14) and IRF-7 (13) in IFN-α secretion via cytosolic DNA sensors (12). Therefore, we determined whether IRF-3 and IRF-7 are activated in response to human DNA in DCs. DNA-binding activity of IRF-3 was assessed using a transcription factor ELISA kit for IRF-3 according to the manufacturer’s protocol. This kit measures the binding of IRF-3 to consensus DNA binding sequences, which activate transcription. Significant activity (p < 0.05) of IRF-3 was observed in the DNA activated group at three hours post stimulation compared with unstimulated controls (Fig. 5). Earliest IRF-3 activity was detected 2 hours after addition of DNA, peaked at ∼3 h, and declined to baseline level by 4–5 h (data not shown). More importantly, IRF-3 activity was significantly increased (p < 0.05) in DC from aged subjects exposed to human DNA as compared with DCs from young subjects exposed to human DNA (Fig. 4). Basal level of IRF-3 activation was not significantly different (p > 0.05) between the aged and young subjects. No phosphorylation of p38, AKT, or IRF-7 was observed at all time points tested (data not shown). Increased activation of IRF-3 in DNA-treated DCs from aged may thus account for an enhanced secretion of IFN-α.

Besides IRFs, NF-κB signaling pathway is also reported to be activated by intracellular DNA sensing (12, 15). It has been shown that IRF-3 is activated by ds-B-form-DNA through the NF-κB dependent signaling pathway (16). Moreover, members of the NF-κB family of transcription factors p50, p52, p65 (RelA), c-Rel, and RelB are expressed at relatively high level in DCs (17). To varying degrees, the NF-κB transcription factors are associated with DC development, maturation, and functions. Therefore, to further understand the mechanisms of increased activation of DCs from aged subjects to human DNA, we determined the activation of p65 subunit of NF-κB signaling pathway using flow cytometry and transcription factor ELISA. Our studies revealed that there was an increased activation of the p65 subunit of NF-κB at 3 h post activation with DNA (Fig. 6,A). Furthermore, the level of activation of p65 at the basal level and after activation with DNA was significantly higher (p < 0.05) in DCs from aged subjects compared with young subjects (Fig. 6,A). These data were further confirmed with flow cytometry (Fig. 6, B and C) by determining the phosphorylation of the p65 subunit using specific Abs.

The predominant form of inducible NF-κB in APCs is considered to be the p50/p65 heterodimer. However, the p50 subunit cannot initiate transcription because it does not contain a DNA-binding sequence. To further investigate a role of NF-κB in the activation of DCs by human DNA, we used a cell-permeable, specific, peptide inhibitor, NF-κB SN50 (18) (BIOMOL) that inhibits the translocation of p50 unit to the nucleus preventing the formation of active p50/p65 heterodimers. Addition of the SN50 inhibitor (50 μg/ml) to human DNA-stimulated DCs resulted in complete inhibition of p65 phosphorylation compared with both control peptide (SN50m) treated and no peptide treated human DNA-stimulated DCs (Fig. 6,D). Furthermore, addition of the inhibitor resulted in a significant inhibition (p < 0.05) of IFN-α and IL-6 secretion from human DNA-stimulated DCs (Fig. 6,E). No inhibition in the production of these cytokines was observed using a control peptide at the same concentration (Fig. 6 E). These data suggest that increased level of NF-κB activation in DCs from aged subjects may be responsible for their increased reactivity to human DNA.

Aging is considered a state of chronic inflammation, which is believed to be underlying cause of several diseases and mortality associated with aging (19). However, the mechanisms responsible for chronic inflammation in aging are not well understood. In addition, aging is also associated with increased reactivity to self Ags as evidenced by the presence of autoantibodies to a variety of self Ags (20, 21, 22). In this study, we show that DCs from aged subjects display increased reactivity to human DNA, resulting in increased proinflammatory cytokine production and induction of T cell proliferation.

The immune system is normally protected from exposure to self dsDNA during apoptosis due to the rapid engulfment of apoptotic cells, and the abundance of extra- and intracellular DNases (23, 24). However, phagocytic cells may be exposed to cellular DNA following tissue necrosis, inflammation, or viral infection. Defective clearance of apoptotic cells would also result in an accumulation of late phase apoptotic cells. Previous study from our laboratory in humans (10) and a recent study (22) in mice suggest that apoptotic cell clearance is decreased with age. The late phase apoptotic cells are associated with additional proteolytic degradation of specific autoantigens, which may release endogenous danger signals like nuclear structures clustered in apoptotic blebs (chromatin and dsDNA) and proteins such as HSPs causing maturation of DCs resulting in T cell immunity to self (24, 25, 26). Our results indicate that intracellular human DNA delivered via transfection agents such as lipofectamine can stimulate DCs to secrete cytokine and up-regulate costimulatory molecules. This is in keeping with recent studies where intracellular mammalian DNA delivered via fugene (27) stimulated murine bone marrow derived dendritic cells to mature and secrete IFN-α. Intracellular administration of double-stranded B-form DNA also triggered antiviral responses, including production of type I IFNs and chemokines (16). Furthermore, macrophages from mouse embryos deficient in DNase II contain intracellular undigested DNA that causes lethal type I IFN production (28). Mammalian DNA or RNA complexed with Abs, RNA binding proteins, or antimicrobial peptides are also able to activate APCs to secrete type I IFNs (29, 30, 31) and have been shown to be involved in autoimmune disorders such as systemic lupus erythromatous (32).

We have previously reported that DCs from aged subjects secreted increased levels of proinflammatory cytokines TNF-α and IL-6 in response to TLR ligands (10). Present results show that this enhanced proinflammatory cytokine secretion by DCs from aged subjects is not restricted to TLR ligands but is also seen with self-Ags such as human DNA (Fig. 1, A and B). There is increased up-regulation of costimulatory molecules CD80 and CD86 and enhanced secretion of IFN-α and IL-6 by DCs from aged donors in response to self-DNA. These results are further substantiated by our findings that DCs from aged subjects also show increased response to late apoptotic cells (Fig. 2, A and B). Activation of DCs with apoptotic cells resulted in the secretion of TNF-α along with IFN-α and IL-6. Apoptotic cells contain other self-Ags besides DNA such as HSPs, HMGBP1 that can stimulate DCs. This may be responsible for the observed TNF-α secretion.

Activation of DCs converts them into fully functional APCs capable of priming T cell responses. We show that human DNA-activated DCs from aged subjects primed T cells to proliferate (Fig. 3,A) while the less activated DCs from young subjects did not induce T cell proliferation (Fig. 3,A). The levels of IFN-γ in the DNA stimulated DCs from aged subjects was also significantly above (p < 0.05) the unstimulated DCs, while no such difference was observed in IFN-γ secretion from young subjects. Previous studies have reported that priming of T cells by DCs requires optimal level of costimulation as well as cytokine production (33) and this may account for the increased proliferation of T cell observed in the aged subjects. Though the DNA-stimulated DCs from aged subjects display increased capacity to prime T cells compared with their young counterparts, nevertheless overall they displayed reduced level of T cell proliferation (Fig. 3 A) compared with the young subjects. This supports previous report where reduced alloimmunity was observed in aged mice compared with young mice during solid organ transplant (34). However, the reasons for this are presently unclear. We have observed the same phenomenon in other experiments and the underlying mechanisms are being investigated. One of the possible mechanisms may be the impaired formation of DC-T cell synapse in aging.

Earlier studies with intracellular DNA delivered via transfection reagents have shown that DNA signals through non-TLR receptors (16, 27). This is in agreement with our results in that the DNA was localized in the cytosol (Fig. 4 A) and was not accessible to intracellular TLRs in the endosomes. The nucleic acid sensing TLR3 and TLR8 are found in the endosomes (35). The two other known nucleic acid sensing TLRs, TLRs 7 and 9 are not expressed in human monocyte derived DCs and are also present in the endosomes (36).

Furthermore, we did not find any difference in the expression of DAI (Fig. 4, B and C), one of the cytosolic DNA sensors identified recently. However, we did find activation of IRF-3 and phosphorylation of p65 unit of NF-κB signaling pathway, both of which are reported to be activated downstream of DAI (13). IRF-3 activity was significantly higher in the DNA-treated DCs from aged subjects compared with young subjects (Fig. 5). Because IRF-3 is involved in the production of IFN-α (11), it may explain the increased secretion of IFN-α in DCs from aged subjects. The phosphorylation of p65 was also significantly enhanced in the DCs from aged subjects (Fig. 6, A–C). However, in contrast to IRF-3, the basal level of p65 activation was significantly higher in DCs from aged subjects suggesting that DCs in aged subjects are in an activated state compared with young counterparts. Abrogation of cytokine secretion by NF-κB inhibitor (Fig. 6 E) lent further support to this hypothesis. The activated state of DCs in aged subjects could be due to factors in the aged microenvironment such as increased proinflammatory cytokines that may cause the activation of DCs resulting in their increased reactivity to human DNA. Our previous studies had revealed no significant difference in the phenotype of DCs from aged and young subjects (8, 10). Nevertheless, it was possible that DCs were not fully activated to manifest a mature phenotype but may still be primed or semiactivated. The phagocytic capacity of DCs decreases as they become activated and we have reported that DCs from aged subjects are impaired in their capacity to phagocytose Ags (10). Earlier studies from our laboratory have also shown that the expression of PTEN protein, which negatively regulates the PI3K signaling pathway, is increased in DCs from aged subjects. This leads to increased activation of p38 MAPK and enhanced secretion of proinflammatory cytokines (10) in response to TLR ligands. PI3K also regulates the activation of NF-κB pathway (37). Increased PTEN expression in aged subjects may thus account for the increased level of NF-κB activation in DCs. This semiactivated state of DCs alters the response of DCs to Ags resulting in the secretion of increased level of cytokines.

In summary, we show that DCs from aged subjects are more reactive to human DNA and secrete increased level of proinflammatory cytokines. Increased reactivity to self may result in impaired tolerance and thus be one of the mechanisms of age-associated chronic inflammation and autoimmunity. Recent reports indicate that there is enhanced basal level of activation in of NF-κB in DCs from rheumatoid arthritis patients (38). However, it remains to be determined whether similar mechanisms may be involved in other autoimmune diseases associated with increased proinflammatory cytokine production such as rheumatoid arthritis and systemic lupus erythromatous.

We thank GCRC for providing the samples of young control subjects.

The authors have no conflicting financial interests.