Differential Production of IL-12, IFN-α, and IFN-γ by Mouse Dendritic Cell Subsets

Mice

All mice were bred under specific pathogen-free conditions in the animal facility of The Walter and Eliza Hall Institute. For most of the experiments, C57BL/6J Wehi mice 6–10 wk of age were used. The CD8^−/− C57BL/6 mice and the CD4^−/− C57BL/6 mice were originally obtained from T. Mak (Ontario Cancer Institute, Toronto, Canada) (26, 27). The CD3ε^−/− mice were originally obtained from B. Malissen (Centre d’Immunologie Luminy, Marseille, France) (28).

Cytokines, Abs, and reagents

Murine rGM-CSF and murine rIL-4 were gifts from Immunex (Seattle, WA). Rat rIFN-γ (bioactive in mouse) and murine rIL-18 were purchased from PeproTech (Rocky Hill, NJ). Murine rIL-12 p70 was purchased from R&D Systems (Minneapolis, MN). LPS and polyinosinic-polycytidylic acid (poly(I:C)) were purchased from Sigma-Aldrich (Castle Hill, Australia). Flt-3 ligand (Flt-3L) was produced in this laboratory from the Chinese hamster ovary (CHO)-flk2 cell line provided by N. Nicola (The Walter and Eliza Hall Institute). Pansorbin (fixed and heat-killed Staphylococcus aureus (SAC)) was purchased from Calbiochem-Novabiochem (Alexandria, Australia). Oligonucleotides containing a CpG motif (CpG) were synthesized by GeneWorks (Adelaide, Australia) according to a published sequence (CpG1668; Ref. 29) either as diester or fully phosphorothioated (-ph). A preparation of CD40-L (a membrane extract prepared from Sf9 insect cells transfected with murine CD40L, a dilution of 1/100 had the highest bioactivity) was provided by D. Tarlinton (The Walter and Eliza Hall Institute). The hybridoma producing the agonistic mAb against mouse CD40 (FGK45.5) was provided by A. Rolink (Basel Institute for Immunology, Basel, Switzerland). The fluorescent Ab used for selecting DC was FITC-conjugated anti-CD11c (N418). The Abs used for segregating DC subtypes were PE-conjugated anti-CD4 and Cy5-conjugated anti-CD8 and were purified and labeled as published elsewhere (3, 4, 5).

Mouse DC isolation and sorting

Pools of spleens, thymii, or s.c. or mesenteric LN were used for DC extraction as described in detail elsewhere (4, 5). Briefly, organs were chopped, digested with collagenase, and treated with EDTA. Light density cells were collected by a density centrifugation procedure. Non-DC lineage cells were depleted by coating them with a mixture of mAbs and then depleting the coated cells with magnetic beads coupled to anti-rat-IgG (3). The DC-enriched preparations were then immunofluorescent stained with anti-CD11c-FITC, anti-CD4-PE, and anti-CD8-Cy5 mAb. For the experiments with the CD8^−/− mice an anti-CD205-PE (anti-DEC-205) was used to select the equivalent of the CD8⁺ population in combination with anti-CD11c-Cy5 and anti-CD4-FITC. Propidium iodide was added in the final wash to label dead cells. For cytometric sorting, the cells were gated for DC characteristics, namely, high forward and side scatter and bright staining for CD11c, with propidium iodide-staining cells and autofluorescent cells excluded. These selected DC were then segregated based on the staining for CD8 (excluding cells showing intermediate levels of fluorescence, which included autofluorescent cells), staining for CD4, or absence of staining for both. The purity of the sorted DC was >98%. For some experiments the contaminating cells in the initial DC preparation with strong autofluorescence were also collected. For some of the experiments using T cell mutant mice and IL-12 as readout, an in vivo treatment with Flt-3L for 10 consecutive days (10 μg/mice/day administered s.c.) was performed to reduce the number of mice needed; Flt-3L treatment did not affect the IL-12-producing capacity of the CD8⁺ DC.

Stimulation of isolated DC for cytokine production

Sorted splenic mouse DC (1 to 2 × 10⁵) were cultured in 200 μl modified RPMI 1640 medium containing 10% FCS, in 96-well round-bottom plates at 37°C, in an atmosphere of 10% CO₂ in air. After culture the supernatant was collected, separated from cells by centrifugation, and stored until analysis at −70°C. For IL-12 production the stimulation mixture consisted of cytokines and a microbial or T cell stimulus. The cytokines were GM-CSF (200 U/ml) and IFN-γ (20 ng/ml) in the presence or absence of IL-4 (100 U/ml) as indicated (see Figs. 1⇓ and 2⇓). The stimuli used were CD40-L (1/100 or 1/200 dilution) or CpG-ph (0.5 μM) or anti-CD40 mAb (25 μg/ml) and/or LPS (100 ng/ml) or poly(I:C) (100 μg/ml). The incubation time was 18–34 h. For IFN-α production the stimulation used was poly(I:C) (100 μg/ml) and CpG (10 μM) or CpG-ph (0.5 μM). The incubation time was 2 days. For IFN-γ production the stimulation used was IL-12 or IL-18 (at the levels specified in the figures), IL-4 (100 U/ml), GM-CSF (200 U/ml), rat-IFN-γ (20 ng/ml), and CpG-ph (0.5 μM). The incubation time was 3 or 5 days.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

IL-12 production by subtypes of splenic DC in response to T cell or microbial stimuli in the presence or absence of IL-4. Sorted subtypes of splenic DC (5 × 10⁵/ml for A, B, and E; 1 × 10⁶/ml for C and D) were stimulated in the presence of GM-CSF and IFN-γ with (A–C) or without (D, E) the presence of IL-4. In each of A–D the stimulus used was as indicated for 18 (A and B) or 23 h (C and D). SAC was used for 18 h of stimulation in E. A, C, and D, IL-12 p70; B, IL-12 p40 as detected by specific ELISA on the culture supernatants. E, All three secreted forms of IL-12 (p40 monomer, p70 heterodimer, (p40)2 homodimer) as detected by Western blotting. Molecular masses (kDa) are indicated. One experiment representative of three to seven individual experiments is shown.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

IL-12 production by DC subtypes from spleen, thymus, and LN. Sorted DC (10⁶/ml) were stimulated by a combination of LPS and anti-CD40 mAb in the presence of GM-CSF and IFN-γ. Culture supernatants were analyzed by ELISA. The data for spleen DC is a pooled average from five separate experiments. The data for thymus and mesenteric (m) and s.c. (s) LN are from one experiment representative of three.

Analysis of IL-12 and IFN-γ by ELISA

Aliquots of DC culture supernatants were assayed by two-site ELISA. Briefly, 96-well polyvinyl chloride microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with appropriate purified capture mAb, namely, R2-9A5 (anti-IL-12 p70, the hybridoma obtained from American Type Culture Collection (ATCC, Manassas, VA)), C15.6 (anti-IL-12 p40, PharMingen), R4-6A4 (anti-IFN-γ, ATCC). Cytokine binding was then detected with appropriate biotinylated detection mAb, namely, R1-5D9 (anti-IL-12 p40, ATCC), C17.8 (anti-IL-12 p40, hybridoma provided by L. Schofield, The Walter and Eliza Hall Institute), XMG1.2 (anti-IFN-γ, ATCC). The readout was then obtained by incubating with streptavidin-HRP conjugate (Amersham Pharmacia Biotech, Buckinghamshire, U.K.), followed by a substrate solution containing 548 μg/ml ABTS (Sigma-Aldrich) and 0.001% hydrogen peroxide (Ajax Chemicals, Auburn, Australia) in 0.1 M citric acid, pH 4.2. The OD of each of the samples was scanned at 405–490 nm using an ELISA plate reader.

IL-12 polypeptide analysis by Western transfer and immunoblotting

Aliquots of DC culture supernatants were subjected to SDS-PAGE (9% acrylamide) under nonreducing conditions. Aliquots of SeeBlue Pre-Stained Standards (Novex, San Diego, CA) were included on each gel for the estimation of molecular mass. The electrophoresed proteins were transferred onto Immobilon-P membrane (Millipore, North Ryde, Australia) according to the manufacturer’s instructions. Membranes were blocked with 5% BSA in PBS overnight at 4°C. IL-12 polypeptides were detected by incubation with biotinylated C17.8 (anti-IL-12 p40) mAb (0.5 μg/ml in 1% BSA, 0.05% Tween 20 in PBS) for 1 h at 4°C, followed by incubation with streptavidin-HRP conjugate (Amersham Pharmacia Biotech) dilution in 1% BSA, 0.05% Tween 20 in PBS for 1 h at 4°C. The membranes were then developed with Supersignal West Pico Chemiluminescent Substrate (Pierce, Rockford, IL), according to the manufacturer’s instructions.

Analysis of IFN-α by bioassay

Culture supernatants were tested for IFN antiviral activity using a cytopathic effect reduction bioassay as previously described (30). Briefly, 3 × 10⁴ L cells were seeded into 96-well plates and left for 4 h at 37°C to adhere. Duplicates of IFN standards and test supernatants were added to the plate in semi-log10 serial dilutions. The dilutions of standards were such to give a range of IFN activity from 1 to 10⁴ IU/ml. The plate was incubated overnight for 15 h, the medium was removed, and Semliki Forest virus was added to the wells at a titer 100 times that of the tissue culture ID₅₀. Plates were incubated for 3 days at 37°C and then scored for cell viability. Every plate contained a duplicate row of cells without either IFN or test supernatant to serve as a control for maximal cell death. IFN titers were determined as the concentration of IFN to provide protection to 50% of the cells relative to National Institutes of Health standard Ga02901511.

Splenic CD8⁺ DC are the major IL-12 producers in response to microbiological and T cell-derived stimuli

Recent studies have indicated that the splenic CD8⁺CD205⁺CD11b⁻ DC are the major producers of IL-12 in response to microbiological stimuli, both in vivo and in vitro (15, 16, 17, 18, 19). To obtain functional information on the newly described, most numerous splenic DC population expressing CD4 (4) and to investigate whether T cell-derived stimuli gave the same results as the microbial stimuli, we analyzed all three spleen DC populations for their IL-12 p70- and IL-12 p40-producing capacities. The CD8⁺CD4⁻ splenic DC were identified as the major producers of IL-12 p70 and p40 in response to either T cell stimuli (CD40-L or anti-CD40 mAb) or to a range of microbiological stimuli (CpG, LPS, poly(I:C), SAC) (Fig. 1⇑ and data not shown). In comparison, both the CD8⁻ DC populations produced much less IL-12, with the CD8⁻CD4⁻ DC producing relatively more than the CD4⁺ DC.

Recently we and others described a new aspect of the control of DC IL-12 production, namely, that IL-4 drastically up-regulates IL-12 p70 but down-regulates IL-12 p40 (31, 32). Apart from IL-4, we also found that IFN-γ and GM-CSF were necessary for optimal IL-12 p70 production (31). To determine whether all DC populations were subject to the same control, the effect of IL-4, GM-CSF, and IFN-γ on the IL-12 production of all three splenic DC populations was investigated. The relative effects of each cytokine on the DC subsets were found to be as described for unseparated splenic DC with IL-4 promoting IL-12p70 production in each case (Fig. 1⇑), and the presence of GM-CSF and IFN-γ optimizing this IL-4 effect (data not shown). Although IL-4 served to maximize IL-12 p70 production, the relative differences in IL-12 p70 production between the DC subsets were unchanged in the absence of IL-4, with CD8⁺ DC producing the most, and CD4⁺ DC the least IL-12 p70. The presence of IL-4, GM-CSF, and IFN-γ did not change the surface phenotype of the DC as judged by CD4 and CD8 expression. Because the IL-12 production of the CD8⁻ DC was very low, and often under the detection limit of the assay in the absence of IL-4 (Fig. 1⇑D), we wished to ensure that the poor response was not simply due to absence of a receptor for one particular stimulus. However, the lower production of IL-12 by splenic CD8⁻ DC could be observed with a large number of stimuli and with all cytokines tested (Fig. 1⇑ and data not shown).

CD8⁺ DC of thymus and LN also have the highest capacity for IL-12 production

CD8α-expressing DC cannot only be isolated from spleen but also from thymus and LN. In thymus the DC expressing high levels of CD8αα account for the majority of all DC; pickup of CD8αβ from T lineage thymocytes causes even the DC that do not produce CD8α to stain at moderate levels (4). In LN, DC vary with respect to the level of CD8α on their surface, and at least the high and medium levels of staining represent genuine expression by the DC themselves (3, 9, 10). In contrast to the spleen, where high expression of CD8 correlates with high expression of CD205, LN contain additional CD205⁺ DC showing low to moderate CD8 staining (33). Therefore, we separated DC from thymus or LN into a population staining brightly and a population showing low to intermediate fluorescence after staining for CD8α. Assays for IL-12 production after stimulation with T cell-derived or microbiological stimuli, in the presence or absence of IL-4, revealed that high expression of CD8 correlated with the highest IL-12-producing capacity of the DC from all lymphoid organs, not just from spleen, although the levels of IL-12 produced from LN and thymus were much lower (data not shown). To increase the levels of IL-12p70 produced from these other lymphoid organs (as well as from spleen), a combination of a T cell-derived stimulus (anti-CD40) and a microbiological stimulus (LPS) was used (Fig. 2⇑), resulting in an additive increase in IL-12 production. Because the difference in IL-12 production capacity between CD8^high and CD8^low DC from thymus and LN was smaller than that found for splenic DC, it is possible that the DC population from those organs showing the lower levels of CD8 staining might contain a small subpopulation of CD8^int DC with IL-12-producing capacity equal to that of the CD8^high DC. The absolute level of IL-12 produced by the CD8^high DC varied in the different lymphoid organs, with the DC from spleen consistently showing the highest levels and the DC from the mesenteric LN having the lowest level of production, regardless of the stimulus used and whether or not IL-4 was included (Fig. 2⇑ and data not shown). The lower production of IL-12 by mesenteric LN DC might be due to the high levels of IL-10 in intestinal associated lymphoid organs, because IL-10 has been reported to have a negative priming effect on the IL-12-producing capacity of DC (34, 35).

CD8α itself does not determine the IL-12-producing capacity of DC

Because we found that IL-12-producing capacity correlates with high expression of CD8 on DC in all lymphoid organs, we asked whether the expression of the CD8 molecule itself in some way accounts for that difference. The CD8⁺ DC from spleen are also positive for the expression of CD205, and within the spleen these are the only DC that express CD205; accordingly, CD205 can be used to select for the CD8⁺ lymphoid-related DC even in the absence of CD8 (36). Therefore, we isolated DC from CD8α null mice on the basis of their CD205 expression, and determined whether these DC still produced IL-12. As shown in Fig. 3⇓A, CD205⁺ DC from CD8 null mice produced as much IL-12 as DC isolated from wild-type mice. Thus, although CD8α serves as a useful marker of IL-12-producing DC, it is not a component that determines or regulates IL-12 production.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

The use of mutant mice to determine the influence of CD8 and of subsets of lymphocytes on the capacity of DC to produce IL-12. Lymphoid-related DC (5 × 10⁵/ml) were sorted based on their expression of CD205 (A) or CD8α (B) from normal (wt) or mutant mice and stimulated with CD40-L (1/100 dilution) in the presence of GM-CSF, IFN-γ, and IL-4 for 34 h. Culture supernatants were analyzed for IL-12 p70 by ELISA. One experiment representative of two (B) or three (A) individual experiments is shown.

T cell priming of DC in vivo is not required for IL-12-producing capacity

The DC expressing different markers are localized in different areas of the lymphoid organs. It is well established that the CD8⁺ DC are mainly found in the T cell areas of the spleen, whereas the CD8⁻ are found in the marginal zones (2). Therefore, we asked whether it was the resultant close contact with T cells that determined the high IL-12-producing capacity of CD8⁺ splenic DC. A previous report indicated that the IL-12 production of total spleen cells in CD40-L null mice and SCID mice in response to soluble Toxoplasma gondii tachozoite extract was not decreased, suggesting that T cell function was not involved (15). However, because in this study only total spleen cells were used and the authors were not able to detect bioactive IL-12 p70 but only IL-12 p40, we examined the issue in more detail.

Because CD8 and the CD4 molecules are involved in the function and the development of CTLs and Th cells, respectively (27, 37), we tested the effects of elimination of these components separately. As shown in Fig. 3⇑, the absence of most of the CTLs in the CD8 null mice had no influence on the IL-12-producing capacity of the CD205⁺, lymphoid-related DC. In a similar approach, the absence of most Th cells in CD4 null mice had no effect on the IL-12-producing capacity of the DC, ruling out “priming” effects of Th cells (Fig. 3⇑). To cover the possibility that one T cell type might substitute for the other, we also used CD3 null mice that lack simultaneously CD4 and CD8 T cells (28), again without effect on DC IL-12 production. In all experiments the IL-12 p70 production of the CD8⁺ DC or CD205⁺ equivalent DC isolated from the different T cell-deficient mice was at least as high (and often higher) than with the control mice (Fig. 3⇑ and data not shown). Therefore, it is not likely that the proximity of T cells to the CD8⁺ DC in vivo is responsible for their higher capacity for IL-12 production. Enumeration and phenotypic analysis of DC from the mutant mice showed that all DC populations were apparently present, although they lacked the expression of CD4 or CD8 in the relevant knockout mice. All DC subtypes were present even in the absence of all peripheral T cells in the CD3⁻/⁻ mice. The DC in these mutant mice expressed normal levels of MHC class II and of the costimulation molecules CD80, CD86, and CD40.

Splenic CD4⁻CD8⁺ DC are the major producers of IFN-α

Type I IFNs have direct antiviral effects on cells and also have various indirect effects leading to immune stimulation. In human blood and peripheral lymphoid organs a population of cells with a CD11c^lowCD4⁺CD3⁻Ig⁻ surface phenotype and a plasmacytoid appearance in electron microscopy (plasmacytoid monocytes or plasmacytoid T cells) has been shown to have an extreme capacity to produce IFN-α (38). These same IFN-producing cells have also been found to act as precursors of DC if cultured with IL-3 and CD40L or virus (39, 40). It has been suggested that these DC precursors that produce IFN-α and their product DC correspond to the murine lymphoid-related CD8⁺ DC lineage because it was shown that they also have the potential to develop under appropriate conditions into T cells (41). Type 1 IFN production could be a marker of the murine DC corresponding to these human DC lineages. To help resolve this issue we used a bioassay to determine the capacity of the three subtypes of mouse spleen DC to produce IFN-α. Poly(I:C) (a double-stranded RNA analog) alone or together with CpG and a variety of cytokines were tested for their ability to induce mouse spleen DC populations to produce IFN-α (data not shown). It was found that a combination of poly(I:C) and CpG induced the highest levels of IFN-α production. This was not effected by the presence of GM-CSF and/or IL-4 (data not shown). Thus stimulation conditions that would lead to the production of IL-12 were amenable to IFN-α production, but the regulatory control for each cytokine differed. As shown in Fig. 4⇓, the CD4⁻8⁺ lymphoid-related DC produced much more IFN-α than the CD8⁻ myeloid-related DC. This supports the original concept that the murine CD8⁺ DC correspond most closely to the human plasmacytoid precursors and their DC progeny. The poly(I:C)/CpG combination led to an activation of all DC populations in culture (as judged by an increase in surface MHC class II expression, an increase in forward light scatter, and a more dendritic appearance). However, the production of IFN-α was not simply a consequence of this activation because other microbiological reagents, such as LPS, which also led to DC activation in vitro, did not induce DC to produce IFN-α.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

IFN-α production by subtypes of splenic DC. Sorted splenic DC (1 × 10⁶/ml) were stimulated with a combination of CpG and poly(I:C) for 48 h. Frozen supernatants were tested for IFN-α by bioassay. One of three experiments with similar results is shown.

The CD4⁻8⁻ DC but not the CD4⁻8⁺ or CD4⁺8⁻ DC produce high amounts of IFN-γ

Recently it was published that after stimulation with IL-12 splenic CD8⁺ DC produce more IFN-γ than do CD8⁻ DC (24). The same group extended their analysis to DC stimulated with a combination of IL-12 and IL-18 or IL-4, reporting that under these conditions CD8⁺ DC and CD8⁻ DC produced equal amounts of IFN-γ (25). We determined the capacity of all three populations of splenic DC, freshly isolated and extensively purified, to produce IFN-γ. In contrast to the published work, we found that under optimal conditions of stimulation, as well as under all conditions tested, it was the CD4⁻8⁻ DC and not the CD4⁻8⁺ DC or the new CD4⁺8⁻ DC that produced the highest amount of IFN-γ (Fig. 5⇓). Furthermore, stimulation with IL-12 alone did not produce any detectable IFN-γ by CD4⁻8⁺ DC. In the presence of both IL-12 and IL-18 all three populations had enhanced IFN-γ production, but the CD4⁻8⁻ DC remained the superior producers.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

IFN-γ production by subtypes of splenic DC. Sorted subpopulations of splenic DC (5 × 10⁵/ml) were stimulated in culture as specified below, then supernatants were harvested after 3 (A-D) or 5 days (E), frozen, then analyzed by ELISA for IFN- γ. A, Titration of IL-12 as only exogenous factor; one of three individual experiments is shown. B, Effects of various combinations of IL-12 with IL-4, IFN-γ, and GM-CSF, on the IFN-γ production of splenic DC. Note that the rat IFN-γ used is bioactive in mouse but does not cross-react in the ELISA for mouse IFN-γ; one of two individual experiments is shown. C, Titration of IL-12 in the presence of IL-18; one of three experiments with similar results is shown. D, Titration of IL-18 in the presence of IL-12; one of two experiments with similar results is shown. E, DC were stimulated with factors known to induce IL-12 production, but analyzed for IFN-γ production; one of two experiments with similar results is shown.

We considered the possible sources of the discrepancies between our data and that of Ohteki et al. (24) and Fukao et al. (25). First, the methods for DC isolation varied. In particular, our procedure involved direct isolation of DC without any adherence or culture steps, whereas these steps may have primed, activated, or changed the DC in the procedures used by the other groups. The protocol used by Fukao et al. (25) for isolating CD8⁻ DC involved positive selection of the CD8⁺ DC and a second positive selection with CD11c. This procedure must have included a high proportion of CD4⁺8⁻ DC, which are the most numerous DC population in spleen, and which are poor producers of IFN-γ, so diluting the CD4⁻8⁻ DC population. Second, the presence of autofluorescent macrophages may have confused the functional studies. We recently have described how easily such cells can contaminate a DC preparation and be enriched with a positively sorted fraction such as CD8⁺ DC (4); in this study we excluded such contaminants. Despite the care of previous workers to exclude T cells or NK cells, the presence of such contaminating autofluorescent macrophages could have given misleading results. We have determined the endogenous IL-18 production by such isolated autofluorescent cells and found that after 18 h they produce 10 times more than the purified DC themselves (data not shown). The simultaneous presence of CD8⁺ DC, together with such autofluorescent contaminants that produce endogenous IL-18, is the equivalent to the exogenous addition of IL-18, which would then synergize with IL-12 to promote IFN-γ production (Fig. 5⇑). However, this cannot account for the previous results where saturating amounts of IL-18 were used.

An interesting point is that in our hands the CD4⁻8⁻ DC showed the highest IFN-γ production even when the source of stimulating IL-12 was endogenous induction (Fig. 5⇑E), despite the fact that these DC have a relatively low capacity to produce IL-12. One explanation is that the CD4⁻8⁻ DC are able to respond to minimal amounts of IL-12. As shown in Fig. 5⇑, A and C, the CD4⁻8⁻ DC produce IFN-γ with as little as 5 pg/ml IL-12. The reason for this excellent response to low IL-12 levels could be high expression of a high affinity IL-12 receptor. Consistent with this hypothesis, CD8⁻ DC have been shown to express higher levels of the IL-12 receptor mRNA than CD8⁺ DC (42). Alternatively, the cytokines used in the stimulation of IL-12 production (IFN-γ, IL4, and GM-CSF) may synergize to enhance the survival of CD4⁻CD8⁻ DC in culture.

Concluding remarks

DC are well known for their important role in stimulating the adaptive immune system, due to their capacity for efficient presentation of Ag to naive T cells. In response to many microbiological challenges and other danger signals, or to T cell-derived stimuli like CD40L, DC are able to produce a variety of factors with important regulatory functions. Mouse DC are heterogeneous by surface phenotype, but the correlation of surface marker expression with functional differences is not as yet completely elucidated. In this study we extend this functional information and show that the differential expression of the molecules CD8 and CD4 correlates with major differences in the capacity to produce IL-12, IFN-α, and IFN-γ.

DC from different lymphoid organs with high surface expression of CD8 have the highest capacity for IL-12 production. This major production of the CD8-positive population could be shown in response to microbial products as well as to T cell-derived stimuli. Therefore, the CD8-positive DC might be considered as professional IL-12 producers for an immediate response to microbial products, targeting the innate immune system as well as priming the adaptive immune system in the encounter with T cells. The very large amounts of IL-12 produced by the CD8⁺ DC might enable long range and even systemic IL-12 responses to occur, whereas the low level of IL-12 produced by the CD8⁻ DC might be more for short range interactions or for an autocrine mechanism. One example of the latter may be the superior IFN-γ production of the CD8⁻4⁻ DC in response to endogenously induced IL-12 (Fig. 5⇑E). We have not elucidated the reasons for the differences in IL-12-producing capacity of the major spleen DC populations, but by using genetically modified mice we were able to rule out a direct role of the CD8 molecule itself or a priming through to close contact with T cells as the reason for the high IL-12 production by CD8⁺ DC. Recently the chemokine receptor CCR5 has been linked to the IL-12-producing capacities of the CD8⁺ splenic DC (19). A source of CCR5 ligands (i.e., macrophage-inflammatory protein (MIP)-1α or MIP-1β) might be endothelial cells; therefore, the migration of the lymphoid-related DC through CCR5 ligand-expressing endothelia might ultimately be the reason for the higher capacity of the CD8⁺ DC to produce IL-12.

It is of interest that the expression of CD8 also correlates with the major capacity for IFN-α production by DC. IFN-α is a cytokine that like IL-12 has a variety of immune stimulatory properties, including increasing the cytotoxicity of NK and T cells. In humans IFN-α also has Th1 directing properties. Because both cytokines are important for both the innate and the adaptive immune systems, the major function of the CD8⁺ DC in defense against virus, bacteria, and tumors could be secretion of such cytokines. IFN-α production in human blood and lymphoid organs has been linked to a special cell type with plasmacytoid appearance (20, 21). This cell type has precursor potential for T cells as well as for DC and depends on IL-3 but not GM-CSF for survival and proliferation in vitro, features related to the CD8⁺ DC in the mouse (43). We have shown that these mouse CD8⁺ DC can be induced to produce IFN-α using the same types of stimuli (analogs of bacterial and viral nucleic acids) that induce IFN-α production from plasmacytoid cells in vitro. A high IFN-α-producing capacity could be interpreted as an additional functional link between murine CD8⁺ DC and the human plasmacytoid-derived DC.

In contrast to the superior capacities of the CD4⁻8⁺ DC for IL-12 and IFN-α production, we identified the CD8⁻4⁻ DC as major producers of IFN-γ in response to IL-12, or a combination of IL-12 and IL-18. The in vivo relevance of this data is highlighted by Ohteki et al. (24), who have previously identified DC as an important source of IFN-γ production in vivo, even in the absence of lymphoid cells and NK cells. Even though the CD4⁻8⁻ DC and CD4⁻8⁺ DC exist in spleen in nearly equivalent numbers, these DC populations are located in different areas of the spleen, the red pulp or the T cell areas, respectively, and produce a different pattern of cytokines. To date the most numerous splenic DC subtype, CD4⁺8⁻, has not been found to be a major producer of cytokines.

This study examined the cytokine production of DC subsets in isolation. We have found that the mouse CD8⁺ and CD8⁻ DC subsets of the spleen differentially make cytokines. In theory, the two subsets could directly influence the cytokine production of each other in vivo. In an in vivo model system, Grohmann et al. (44) have shown that CD8⁺ DC can indeed directly effect the function of CD8⁻ DC, due to effects of IFN-γ on the CD8⁺ DC subset. However, in the normal dynamic in vivo situation the functional interdependence of the mouse DC subsets is less clear. Mice that lack one or more DC subset are now required to facilitate the study of the functional interplay between the DC populations.