Dysfunction of dendritic cells in aged C57BL/6 mice leads to failure of natural killer cell activation and of tumor eradication

Significance

Immunosenescence is an important phenomenon that leads to enhanced susceptibility both to bacterial and virus infections and to tumorigenesis. The reciprocal activation of dendritic cells (DCs) and natural killer cells (NKs) is a critical point in the maturation of both the adaptive and innate immune systems. Its failure could be a key point in immunosenescence. In this article, we show that in aged C57BL/6 mice that were known to be susceptible to mousepox virus, DCs are dysfunctional and unable to activate NKs. This defect also results in failure to eliminate RMA-S lymphoma mutant tumor cells in an NK-sensitive tumor model. A more complex situation regarding DC dysfunction is also described in a small sample of the outbred human population.

Abstract

The reciprocal activation of dendritic cells (DCs) and natural killer cells (NKs) plays a key role in both innate and adaptive immunity. The effect of aging on this cross-talk, a critical step in virus disease control and tumor immunology, has not been reported. Splenic DCs and NKs were purified from both young and old C57BL/6 mice and cocultured in the presence of polyinosinic:polycytidylic acid (poly I:C). The resulting activation of NKs was measured as expression of CD69 and secretion of IFN-γ. However, DCs from old mice could not activate NKs from either young or old mice in vitro or in vivo. In contrast, DCs from young mice efficiently activated NKs from both young and old mice. DCs from old mice were deficient in poly I:C-stimulated secretion of IL-15, IL-18, and IFN-α. Gene expression analysis revealed many other differences between DCs of old and young mice. Young mice strongly eradicated MHC class I-negative NK-sensitive RMA-S lymphoma mutant tumor cells, but old mice did not, in concert with the previous report that mousepox kills aged, but not young, C57BL/6 mice. Furthermore, a similar dysfunction of DC and its key role in NK activation was found in 27 out of 55 healthy human donors.

Natural killer cells (NKs) and dendritic cells (DCs) are key players in the immune system. Previous studies have shown that their cross-talk plays a central role in the NK functions in tumor immunity and viral diseases (reviewed by ref. 1). DCs were reported to prime NK-mediated antitumor responses in 1999 (2). A few years later, the reciprocal interaction of DCs and NKs during viral infection was also described (3). Cytokines from DCs as well as cell–cell contact were shown to enhance NK function. DCs secreted IL-12 and IL-18, which promoted cytokine production by NKs (4, 5). Further, membrane-bound IL-15/15Rα on DCs is important in promoting NK survival and proliferation (6–8), and type I IFNs are crucial in the induction of NK cytotoxicity. In addition, some surface molecules on DCs—for example, CD40, CD48, CD70, CD80, CD86, major histocompatibility complex class I-related chain A/B (MICA/B), leukocyte function-associated Ag-1 (LFA-1), and chemokine (C-X3-C motif) ligand 1 (CX3CL1)—have been reported to be involved in the effect of DCs on NKs (1).

It is well known that aged humans are more susceptible to malignancies and infectious diseases and less responsive to vaccination because of immunosenescence—that is, progressive deterioration in immune function with aging. Increasing evidence shows aging affects many cells in both innate immunity and adaptive immunity (9). For example, impairment of DC and macrophage functions, including phagocytic activity, cytokine secretion, wound repair, and antigen presentation, were observed (10, 11). In T cells, a shift from the naïve to memory phenotype, a decreased proliferative response, and impaired cytolytic activity were found (12). Some of these defects in cell functions are intrinsic, whereas others might be the consequence of the complicated interactions between immune cells (13). As key cells in innate immunity, functional changes in both NKs and DCs with aging were described. For example, increased numbers of NKs, decreased NK cytotoxicity on a per-cell basis, and reduced levels of cytokines and chemokines upon activation were reported in humans (14). Importantly, aged C57BL/6 mice lost their resistance to mousepox due to a decrease in numbers of NKs and impaired trafficking, a phenomenon not found in several other mouse strains (15). For DCs, aging leads to decreased dendritic cell-specific CD209 (DC-SIGN) expression on immature bone marrow-derived cells (BMDCs) without affecting their function in T-cell priming and Toll-like receptor (TLR) response (16, 17). In a tumor antigen system, BMDCs from aged mice were less efficient in stimulating antigen-specific T cells and displayed defective trafficking (18). However, to our knowledge, the effect of aging on DC-induced NK activation has not been studied. DC/NK cross-talk is a critical factor in maturation of the immune system, affecting both innate and adaptive immunity. In the present work, CD11c+ DCs and DX5+ (CD49b) NKs purified from the spleens of young and aged C57BL/6 mice were cocultured in the presence of polysinosinic:polycytidylic acid (poly I:C), together with corresponding in vivo experiments, to study the effect of aging on this system.

Results

The Effect of Aging on Murine DC-Induced NK Activation.

Coculture of purified CD11c+ DCs and DX5+ NKs from C57BL/6 mice in the presence of poly I:C increased the fraction as well as the level of expression of CD69 on NKs and induced IFN-γ secretion (Fig. 1A). To investigate the effect of aging on this DC-induced NK activation, NKs and DCs were purified from young and aged mice, defined as less than 3 mo versus more than 18 mo of age, respectively, and coincubated in various combinations. Young DCs induced only slightly higher CD69 expression on young NKs compared with coculture of old DCs with old NKs (Fig. 1B). However, although DCs from young mice induced robust IFN-γ secretion from young NKs, DCs from aged mice lost nearly all of their ability to induce IFN-γ secretion by aged NKs (Fig. 1C). Moreover, coculture of DCs from aged mice with NKs from young mice failed to induce secretion of IFN-γ. However, if DCs from young mice were cocultured with NKs from old mice, the old NKs secreted a high level of IFN-γ. No significant difference was found in IL-2–induced IFN-γ secretion from young and old NKs. Thus, NKs from young and old mice have a similar capability for IFN-γ secretion. The impaired activation of old NKs induced by old DCs is not due to an intrinsic defect in old NKs but to dysfunction of old DCs.

Fig. 1.

The effect of aging on murine DCs induced NK activation in vitro. (A, Left) Purified CD11c+ DCs and DX5+ NKs from C57BL/6 mice were cocultured at a ratio of 1:1 in the presence of 10 ng/mL poly I:C for 24 h and separated as shown. (A, Right) CD69 expression by NKs from both young and aged mice increased after coculture with DCs and poly I:C but not after coculture with either DCs or poly I:C alone. (B) NKs from young mice or old mice were cocultured with autologous DCs from aged mice or young mice, unless otherwise noted, in the presence of poly I:C. As a positive control for NK function, NKs were stimulated with IL-2 (1 mg/mL). The percentage CD69+ NKs was determined. (C) Supernatants were collected for measurement of IFN-γ secretion. Data represent mean ± SD of triplicate samples of 3–6 mice in each group and are representative of at least three independent experiments. *P < 0.05.

In Vivo Activation of NK.

To further confirm the effect of aging on DC-induced NK activation, poly I:C was injected into young and aged mice; 24 h later, the expression of CD69 and Granzyme B on splenic NKs was examined. In old mice, a significantly lower percentage of NKs became CD69- or Granzyme B-positive after poly I:C injection (Fig. 2), the former, at least, due to the defect in DC activation of NKs (8). Previous research showed that in vivo the activation of NKs is mainly induced by DCs. All of these results suggest that, in aged mice, poly I:C-stimulated old DCs failed to activate NKs.

Fig. 2.

In vivo activation of NKs: CD69 and granzyme expression. Poly I:C was injected into young and aged mice i.p. After 24 h, the expression of (A) CD69 and (B) Granzyme B on splenic NKs was analyzed. Data represent mean ± SD of triplicate samples and are representative of at least three independent experiments of 3–6 mice in each group. *P < 0.05.

Characterization of Young and Old Splenic DCs.

The difference between DCs from young and old mice was studied next. Several cytokines and surface markers have been reported to be involved in DC-induced NK activation (19). To examine this, splenic CD11c+ DCs were purified and treated with poly I:C. Young DCs secreted significantly higher levels of IL-15/IL-15R, IL-18, and IFN-α after poly I:C treatment compared with old DCs (Fig. 3A). However, neither IL-12p70 nor TNF-α were produced by young or old splenic DCs upon poly I:C stimulation. BMDCs stimulated with LPS were used as a positive control and did secrete IL-12p70 and TNF-α.

Fig. 3.

Cytokine secretion and surface phenotype of young and old splenic DCs. Splenic CD11c+ DCs were purified from young and aged mice and treated with poly I:C. After 24 h, the secretion of cytokines including IL15/IL-15R, IL-18, and IFNα (A) and expression of cell surface markers including CD40, CD48, CD80, CD86, Ia, GITRL, and RAE-1 (B) were analyzed. Data represent mean ± SD of triplicate samples and are representative of at least three independent experiments of 3–6 mice in each group. *P < 0.05.

The surface markers CD40, CD48, CD80, CD86, Ia, glucocorticoid-induced TNF receptor ligand (GITRL), and ribonucleic acid export-1 (RAE-1) were all expressed by DCs (Fig. 3B). Significantly higher levels of CD48, CD86, Ia, GITRL, and RAE-1 were expressed on young DCs compared with aged DCs in the presence and absence of poly I:C treatment. Interestingly, poly I:C treatment did not induce any change in the analyzed cell surface markers. Because DCs can only activate NKs in the presence of poly I:C (Fig. 1), the difference in cell surface markers is not likely to be the cause of the differential activation of NKs by young and old DCs.

To understand in more detail the genomic differences that distinguish young and old DCs, microarray profiling was conducted on highly purified CD11c+ splenic DC populations from young and aged mice treated with poly I:C for 4 h. Hierarchical cluster analysis using a Pearson correlation demonstrated unique gene differences between DCs from young and aged mice and clearly separated young and aged DCs into two groups (Fig. 4A).

Fig. 4.

DCs from young and aged mice have distinct transcriptional profiles. (A) Hierarchical cluster analysis clearly separates transcriptional profiles of young DCs from aged DCs. (B) Volcano plot based on fold change and P value of all transcripts in young DCs (left side) and old DCs (right side). Genes that are similarly expressed between the two populations are shown as black dots. Unique gene signatures based on a Rank <5,000, an absolute expression value difference of >20, and a >1.8-fold change between the two populations are depicted in blue dots (56 probes, young DCs) and red dots (251 probes, aged DCs).

To identify the functional differences that reflect the transcriptional differences between young and old DCs, a gene set enrichment analysis (GSEA) was performed. GSEA is a computational method developed at the Broad institute of Harvard University and MIT that uses a defined set of genes and determines which “functional” sets of genes are up-regulated. In old DCs, 76 functional gene sets were significantly enriched at a nominal P value < 1%, whereas in young DCs 13 gene sets were significantly enriched (Table S1). In young DCs, five gene sets out of the 13 were associated with tissue remodeling and development, whereas in aged DCs 15 out of the 76 significantly enriched gene sets were associated with immune response and lymphocyte activation and eight gene sets were associated with DNA repair. Interestingly core genes of the immune response-related gene sets found in old DCs include cytokines such as IL-4, IL-10, and colony stimulating factor-1 (CSF-1) that are involved in Th2 cell differentiation, immune suppression, and macrophage differentiation, respectively. In contrast, core genes associated with tissue remodeling-related pathways as found in young DCs include two cytokines directly involved in NK activation—namely, IL-18 and IL-7—as well as growth factors such as TGF-β1, TGF-β2, and EGF. Unique gene signatures for young and old DCs were generated and revealed 56 and 251 differentially expressed probes that encode 50 and 210 unique proteins in young and aged DCs, respectively (Fig. 4B and complete gene list included in Table S2). In addition, functional classification of the unique gene signatures confirms an increased representation of immune-related genes and genes involved in DNA repair, apoptosis, and cell-cycle regulators in aged DCs (Table S3), whereas young DCs have an increased representation of genes involved in cell adhesion and immune synapse formation that may represent an increased ability to directly interact and activate other immune cells such as NKs (Table S2). Overall, the gene expression profiles of young and aged DCs suggest that DC intrinsic differences such as cytokine secretion profiles and capacity to form immune synapses are responsible for the difference in the ability of old and young DCs to activate NKs.

Effect of DCs on Eradication of an NK-Sensitive Tumor.

Because old DCs cannot activate NKs efficiently, the question of whether the eradication of NK-sensitive tumor cells was affected in aged mice [in parallel to the inability of NKs from aged mice to eliminate the mousepox virus (19)] was studied. RMA-S tumor cells, which express a very low level of MHC class I and are sensitive to NK killing, were mixed with RMA cells, which express a high level of MHC class I and are not sensitive to NK killing, in a ratio of 1:1 and transferred to young and aged mice i.p., respectively (20). To activate DCs, both groups of mice were injected with poly I:C 24 h before administration of RMA-S and RMA cells. The mice were euthanized at 12–14 h after i.p. injection. Tumor cells were harvested and stained with Ly49A, which is expressed on both RMA and RMA-S cells and with anti–H-2, to specifically identify RMA and RMA-S that have high H-2 and low H-2 expression, respectively (Fig. 5A). NKp46 was used to exclude Ly49A+ NKs. In old-1 mice (16 mo old) and old-2 mice (21 mo old), significantly less RMA-S cells were eradicated compared with young mice (Fig. 5B).

Fig. 5.

Effect of young and old DCs on eradication of an NK-sensitive tumor. (A) RMA-S cells mixed with RMA cells at the ratio of 1:1 were transferred i.p. to young and aged mice. After 12–14 h, the mice were euthanized and the tumor cells were harvested from the peritoneal cavity and stained with Ly49A and H-2 to identify RMA cells (LY49A H-2^HI) and RMA-S cells (LY49A H-2^LO). NKp46 was used to exclude Ly49A+ NKs. Equal numbers of peritoneal cells obtained from young and aged mice were used. The absolute number of tumor cells in the old mice samples appears to be smaller than in the young mice samples because the peritoneal cells from the old mice included many nonlymphocytic cells, probably macrophages, as well as RMA-S and RMA tumor cells. (B) The percentage of RMA-S cells relative to RMA cells was calculated and compared between young and aged mice. (C) Schematic of the procedure for adoptive transfer of young and old splenic DCs. DCs in young CD11c⁻ DTR mice were ablated and then separated into two groups. Equal numbers of purified young or old CD11c+ splenic DCs were adoptively transferred. (D) After injection of poly I:C and inoculation of RMA and RMA-S cells, the mice were euthanized and the percentage of RMA-S cells relative to RMA cells was analyzed. (E) The expression of CD69 and Granzyme B in DCs from young mice was significantly higher than in old mice. Data represent mean ± SD of triplicate samples and are representative of at least three independent experiments (except for C and D, which were carried out only twice due to the paucity of old CD11c-DTR mice). *P < 0.05.

To demonstrate that diminished RMA-S clearance was due to impaired NK activation by old DCs, DCs in young CD11c-diptheria toxin receptor (DTR) mice were ablated by treatment with diptheria toxin (Fig. 5C). Subsequently, DC-ablated young mice were reconstituted by adoptive transfer of equal numbers of young or old DCs. After injection of poly I:C and inoculation of RMA and RMA-S cells, as described above, the mice were euthanized and the eradication of RMA-S cells was analyzed. Young mice that had been reconstituted with old DCs could not eradicate RMA-S cells as efficiently as those reconstituted with young DCs (Fig. 5D). Furthermore, significantly lower expression of CD69 and Granzyme B on NKs in young DTR mice that received old DCs compared with mice that received young DCs was observed (Fig. 5E). Thus, DCs are the key factor for NK activation and determined the killing efficiency of NKs. Aging leads to dysfunction of DCs in these mice and results in a lower efficiency of tumor cell eradication.

Heterogeneity in the Ability of Mature DCs to Induce IFN-γ Secretion from Autologous NKs in the Outbred Human Population.

The question of whether these findings in an inbred mouse strain could be extended to DCs and NKs from the outbred human population was examined. The source of these cells was blood bank leukopaks from which platelets had been removed (Materials and Methods). Fifty-five samples were analyzed. Mature DCs and NKs were prepared and cocultured for 24 h, as described. NKs were separated and analyzed (Fig. 6A). Expression of CD69, one measure of NK activation, was increased in all 55 donors.

Fig. 6.

DC/NK cross-talk with human cells. (A) Enhanced CD69 expression on NKs resulting from DC/NK coculture as described in Materials and Methods using autologous cells. Both immature and mature DCs were used. Unstimulated NKs and NKs stimulated with IL-2 were used as controls. (B) IFN-γ secretion by NKs resulting from incubation of a functional autologous DC/NK pair (Left, donor A) and absence of IFN-γ secretion in a nonfunctional pair (Right, donor B). Note that the addition of a low level of IL-2 or IL-15 that does not induce IFN-γ secretion is required during the coincubation. (C) Coincubation of heterologous NK/DC pairs. DCs from a nonfunctional pair (donor B) are not activated by NKs from functional donor A (Left), whereas DCs from donor A can activate NKs from the nonfunctional donor B. (D) Induction of CD69 expression and IFN-γ secretion by NKs are both age independent in this study, in which young donors were 20–35 y of age and old donors were 55–70 y of age. (E) IL-12, IL-18, and TNFα secretion by nonfunctional (as from donor B) and functional (as from donor A) DCs. Both immature DCs (without LPS stimulation) and mature DCs (after LPS stimulation) were used.

Supernatants were analyzed for secretion of IFN-γ as a marker of functional NK activation. DCs from 24/55 donors promoted the secretion of IFN-γ from autologous NKs in the presence of very low-dose IL-2 (10 IU/mL) or IL-15 (1 ng/mL), which, by themselves, had no effect (Fig. 6B, Left). In contrast to studies with murine cells, no IFN-γ secretion was observed in most of the cocultures in the absence of these minimal amounts of IL-2 or IL-15. Mature DCs promoted the secretion of IFN-γ from NKs in the absence of IL-2 or IL-15 in only 4/55 donors. DCs from 27/55 donors could not induce IFN-γ from NKs, even in the presence of low-dosage IL-2 or IL-15 (Fig. 6B, Right), indicating, again, that IFN-γ secretion and CD69 expression are separate phenomena, even though both are activation markers of NKs.

Importantly, moreover, functional DCs from a donor (“donor A”) with potent ability to stimulate autologous NKs to produce IFN-γ could also promote the secretion of IFN-γ from allogeneic NKs of donor B. However, dysfunctional DCs from a donor (“donor B”) with no ability to stimulate autologous NKs to produce IFN-γ could also not promote the secretion of IFN-γ from allogeneic NKs from donor A (Fig. 6C). These findings parallel those seen in young and aged C57BL/6 mice. No significant increase of apoptotic DCs was detected in the allogeneic coculture system.

In the first leukopaks used, ages of the donors were not available. However, ages were obtained for the donors of the last 16 leukopaks analyzed. DCs and NKs from seven young donors (20–35 y of age) and nine older donors (55–70 y of age) for whom ages were available were cocultured, and the percent of CD69+ NKs and secretion of IFN-γ were analyzed (Fig. 6D). No age-related difference in IFN-γ secretion or CD69 expression in autologous DC–NK coculture was evident. However, the “old” blood bank donors (ages 55–70) are not really aged humans, as the lifespan of humans extends into the 90s. Finally, secretion of soluble cytokines reported to be involved in DC-induced NK activation was examined. DCs with a higher capability to induce IFN-γ secretion from NKs (“functional DC”) produced IL-12, IL-18, and TNF-α, whereas dysfunctional DCs produced very little (Fig. 6E), similar to the data obtained for the young and aged C57BL/6 mice.

Thus, a dichotomy has been found in the outbred human population in the ability of DCs to activate autologous NKs. A large population study will be required to establish whether this variation is related to race, ethnicity, sex, age, or other factors and whether it has any relation to susceptibility or resistance to viral disease or tumors in humans. Clearly, however, reduced cytokine secretion by DCs is associated with reduced NK function, irrespective of the age of the donor.

Discussion

Coupled with previous work demonstrating enhanced susceptibility to virus infection in old mice (15), the data presented here suggest that failure of DC/NK cross-talk is an important component of immunosenescence in C57BL/6 mice. In elderly humans, immunosenescence is a complex process resulting in enhanced susceptibility to infection and tumors. For mice, age-related loss of resistance to mousepox virus (a virus related to human smallpox virus) has been shown to be due to a decrease in number and function of NKs (15). Here, defective NK activation in aged C57BL/6 mice is shown to result from impaired DC function. Adoptive transfer of old DC- to young DC-ablated mice resulted in diminished NK activation and defective eradication of NK susceptible tumor cells, compared with mice reconstituted with young DCs. Moreover, gene expression profiling of old and young DCs showed many distinct differences in these cells. Gene expression profiling and functional GSEA of old and young DCs demonstrated that pathways associated with Th2 and Treg response and lymphocyte activation are more prevalent in old DCs. However, core genes associated with NK activation in these pathways included IL-18, which was present in the young DC subset but lacking in old DCs (19). In vitro experiments confirmed the decrease in IL-18 in old DCs as well as a decrease in IL-15 and IFN-α.

A number of studies of the effects of aging on both murine and human DCs have been published, including effects on tumor immunity (reviewed by refs. 10, 18), but the effect(s) on reciprocal DC/NK activation has apparently not been investigated. Activated NKs mediate MHC-independent cytotoxicity in the innate immune defense against viral infections and some malignancies (21). This function results from deficiency of expression in these cells of MHC class I proteins that are ligands for inhibitory killer cell immunoglobulin-like receptor (KIR) on human NK or inhibitory Ly49 molecules on murine NKs as well as expression of NK-activating ligands induced by viruses or tumors. In the human system, contradictory data on the effect of aging on NKs exist due mainly to different selection criteria of the elderly population studied (22, 23). In mice, different reports on the cytotoxicity of NKs upon cytokine stimulation have appeared (24, 25). In the present work, no difference in IFN-γ secretion from young and aged NKs upon IL-2 stimulation was evident. However, a lower number of NKs in the spleens of aged mice was found, consistent with previous results on the decreased number and trafficking of NKs in aged mice (15).

DCs act as a bridge linking innate immunity and adaptive immunity. They activate naïve T cells to initiate adaptive immune responses. As a key component of innate immunity, DCs also interact with NKs and enhance their functions. However, in the present work, DCs from old mice could not activate NKs from either young or old mice. The impairment of NK activation was due not to any deficiency of NKs itself but to the dysfunction of DCs. This result is consistent with previous research on the functional alterations of liver innate immunity (26). In that research, young lymphocytes secreted a lower level of IFN-γ when they were cocultured with old Kupffer cells in the presence of CpG oligodeoxynucleotides, whereas old lymphocytes produced a substantial amount of IFN-γ when cocultured with young Kupffer cells, indicating that the Kupffer cells were more important in the activation of lymphocytes.

In summary, we show that aging leads to the impaired function of DCs in C57BL/6 mice. NKs cannot be activated efficiently and have lost the ability to kill tumor cells, as well as the ability to control at least mousepox (15). Combined with previous research showing that aged DCs have a decreased ability to induce antigen-specific T cells, leading to deficient tumor immunity, these results strongly suggest that age-related DC malfunction that results in reduced NK activation is a major reason for impaired tumor immunity. Thus, the two important functions of NKs—clearance of virus infections and cytotoxicity against tumor cells—are both diminished due to age-related dysfunction of DCs manifested as the inability to activate NKs.

Materials and Methods

Mice.

C57BL/6 (B6) mice of different ages were purchased from Taconic. Young CD11c DTR tg B6 mice were purchased from Jackson Laboratory and aged in the Harvard Biology Research Infrastructure vertebrate animal facility. Mice were used at 8–12 wk as young mice or 80–120 wk of age as old mice. All protocols were approved by the Standing Committee on the Use of Animals in Research and Teaching of the Faculty of Arts and Sciences, Harvard University.

Ex Vivo Murine DC–NK Coculture and Detection of NK Activation.

CD11c+ DCs and DX5+ NKs were purified from mouse splenic cells using the respective antibody conjugated with magnetic beads (Miltenyi Biotec). Then DCs and NKs were cocultured at a ratio of 1:1 at the cell concentration of 1 × 10⁶/mL in the presence of 10 ng/mL poly I:C. Twenty-four hours later, supernatants were collected for measurement of IFN-γ secretion using an ELISA kit as instructed (eBioscience, Inc.), and the cells were harvested for staining using the CD11c/DX5/CD69 combination (Fig. 1A, Left).

Detection of NK Activation in Vivo.

Mice were injected i.p. with 0.5 μg/g body weight of poly I:C (a ligand for TLR3 in DCs) (Sigma-Aldrich). Control mice received injections of the same volume of PBS. Twelve hours later, the mice were euthanized and cells were stained with NK1.1 or DX5/CD49b and CD69 and then fixed and permeabilized with the Cytofix/Cytoperm reagent (BD Biosciences) followed by intracellular Granzyme B staining. Results represent percent of gated NK1.1+ or CD49b/DX5+ cells producing Granzyme B.

Microarray Hybridization and Data Analysis.

Total RNA was isolated from FACS-purified CD11c+ splenic DC populations isolated from young (8-wk-old) and aged (110-wk-old) mice that had been injected with poly I:C for 4 h previously, using RNeasy micro kit (Qiagne) following the manufacturer’s instructions, and checked for RNA integration by an Agilent Bioanalyzer 2100 (Agilent Technologies). Microarray analysis was performed with the Affymetrix GeneChip Mouse Genome 430 2.0 arrays. The arrays were hybridized, washed, and scanned according to the standard Affymetrix protocol. The gene chip tests were performed by the professional staff of Shanghai Biochip Company. Microarray data are available in the ArrayExpress database (www.ebi.ac.uk/arrayexpress) under accession no. E-MTAB-2863. Raw microarray expression data were normalized with the expression file creator module from the GenePattern software package (http://genepattern.broadinstitute.org) (27). Hierarchical clustering analysis was performed with GenePattern using a Pearson correlation, and all data were visualized with the multiplot module from GenePattern. Gene signatures of differentially expressed genes were generated using the comparative marker selection module in GenePattern and based on a Rank <5,000, an absolute expression value difference of >20, and a >1.8-fold change between the two populations. Gene ontology and functional group was determined using Resource: Gene_Ontology_Tools. GSEA was performed within GenePattern software (28). Signal2Noise was used for ranking genes. Gene sets derived from the Biological Process Ontology (BP:GO) in the Molecular Signatures Database (mSigDB) v4.0, which includes 825 gene sets, were used to identify differential regulated functional gene sets between young and aged DCs.

DC Ablation.

CD11c DTR tg B6 mouse breeding pairs from Jackson Laboratory were used. Pups were reared to 80–90 d of age and then received a single injection of 4 ng DTs (Sigma)/g body weight. After 2 d, the ablation of diptheria toxin was confirmed by staining for CD11c^high MHC-II⁺ cells in splenocytes. The reduction in numbers of CD11c^high DCs varies between 90% and 95%.

DC Transfers and Tumor Eradication.

For the transfer of DCs, 2 × 10⁶ purified splenic DCs from young or aged mice were injected i.v. into the tail vein of young mice from which DCs had been deleted by DT injection 24 h before. For the tumor eradication assay (20), RMA-S and RMA cells were mixed at the ratio of 1:1 and transferred i.p. to young or aged mice that had been injected with 0.5 μg/g body weight of poly I:C 24 h before. Twelve to 14 h later, the mice were euthanized and the tumor cells were harvested and stained with a broadly reactive anti–H-2 mAb (clone m1/42), which is expressed in different levels on RMA and RMA-S; Ly49A, which is expressed on both RMA and RMA-S cells; and NKp46, which is used to exclude NKs. Then, the remaining RMA-S cells were calculated using the ratio of RMA-S to RMA.

Generation and Coculture of Human DCs and NKs.

Whole-blood leukopacks, 200 mL each, harvested from healthy donors between the ages of 25 and 65 without further identification were obtained from Massachusetts General Hospital. Peripheral blood mononuclear cells (PBMCs) were isolated through density-gradient centrifugation with Ficoll-Paque PLUS lymphocyte isolation solution (GE Healthcare) and filtered through BD Falcon 40 μm nylon cell strainers (Becton Dickinson). CD14+ monocytes were selected using MACS Human CD14+ Microbeads and MACS separator (Miltenyi Biotec). The 1 × 10⁶/mL CD14+ monocytes were plated in Falcon six-well plates and incubated for 5 d at 37 °C with RPMI-Complete Media (RPMI-1640) (Gibco Life Technologies) supplemented with 10% (0.1 mL/1 mL total) FBS (Atlanta Biologicals), 100 units per mL of GlutaMaxTM (Penicillin-Streptomycin), 10 mM L-Glutamine, 1% Minimum Essential Medium Non-Essential Amino Acids Solution (MEM NEAA) (Life Technologies, Grand Island, NY), and 100 mM Sodium Pyruvate (Gibco), plus 50 ng/mL GM-CSF (PeproTech) and 100 ng/mL IL-4 (PeproTech) to obtain immature DCs. Fresh media (RPMI-Complete with GM-CSF and IL-4) was added to CD14+ monocytes on days 3 and 5. We added 100 ng/mL LPS from Escherichia coli (Sigma-Aldrich) on day 6 with further incubation for ∼18 to ∼24 h to induce maturation of DCs.

To obtain NKs, frozen PBMCs were thawed and washed with RPMI-1640. CD56+ NKs were selected using MACS Human CD56+ Microbeads Separation (Miltenyi-Biotec), then adjusted to a concentration of 1 × 10⁶ cells/mL.

NK/DC cocultures at a ratio of 1:1 were performed in BD Falcon U-bottom 96-well plates in the absence or presence of 10 U/mL IL-2 or 1 ng/mL IL-15. We used 1,000 U/mL IL-2 or 10 ng/mL IL-15 known to activate NKs in the absence of DCs as the positive control. After 36 h of coculture, cells were collected and analyzed for CD69 expression, and supernatant was collected for analysis of IFN-γ secretion. IL-12, IL-18, and TNF-α secretion by DCs was measured using ELISA kits (eBioscience).