Ag presentation by dendritic cells (DC) in vivo is essential to the initiation of primary and secondary T cell responses. We have reported that DC presenting Ag in the context of MHC I molecules also become targets of specific CTL and are rapidly killed in mice. However, activated DC up-regulate expression of serine protease inhibitor (SPI)-6, a specific blocker of the cytotoxic granule protein granzyme B, which modulates their susceptibility to CTL-mediated killing in vitro. We wanted to determine whether susceptibility to CTL-mediated killing in vivo is also modulated by DC activation. As was previously reported by others, DC treated with different doses of LPS expressed higher levels of SPI-6 mRNA than did untreated DC. The increased expression of SPI-6 was functionally relevant, as LPS-treated DC became less susceptible to CTL-mediated killing in vitro. However, when these LPS-treated DC were injected in vivo, they remained sensitive to CTL-mediated killing regardless of whether the CTL activity was elicited in host mice via active immunization or was passively transferred via injection of in vitro-activated CTL. LPS-treated DC were also sensitive to killing in lymph node during the reactivation of memory CTL. We conclude that increased SPI-6 expression is not sufficient to confer DC with resistance to direct killing in vivo. However, SPI-6 expression may provide DC with a survival advantage in some conditions, such as those modeled by in vitro cytotoxicity assays.

Ag presentation by dendritic cells (DC)5 in vivo is critical to the induction of immune responses and to the expansion and differentiation of effector and memory T cells (1). DC are sensitive to many signals mediated by infectious agents as well as to signals from other immune cells, which affect DC activation status, migration, cytokine secretion, and survival. The balance of these multiple variables ultimately determines the magnitude and phenotype of the resulting immune response.

We have reported that DC presenting cognate Ag in the context of their MHC I molecules become sensitive to CTL-mediated killing in vivo (2, 3). DC killing by CTL has a clear impact on the ability of Ag-presenting DC to prime naive CD8+ T cells (2) and to restimulate existing immune responses (4), and it appears to act as a feedback mechanism that prevents the chronic or excessive activation of T cells in vivo (4). Similar observations have been reported by other authors (5, 6).

Given the critical role of DC in immune responses, it has been proposed that they must be able to resist CTL killing in at least some circumstances. DC activated by pretreatment with LPS or CD40L, or by interaction with memory CD8+ T cells, have been reported to up-regulate expression of the serine protease inhibitor (SPI)-6 and to become resistant to CTL-mediated killing in vitro (7, 8). Similarly, high numbers of cells presenting a viral Ag in the context of MHC I were recovered from the lymph nodes of mice harboring a specific CD4+ T cell response (9), suggesting that interaction with CD4+ T cells may also make DC resistant to CTL-mediated killing in vivo.

SPI-6 belongs to a family of murine and human serine proteases that appear to protect cells from the toxic effects of cytotoxic granules. Expression of SPI-6 by activated CD8+ T cells is critical in ensuring cytotoxic granule integrity and in allowing the development of memory CD8+ T cell responses after viral infection (10). Additionally, SPI-6 is also expressed by DC and other cell types (10, 11, 12). SPI-6 mediates its antiapoptotic function by inhibiting the activity of granzyme B, to which it irreversibly binds (13). However, studies in mice that are deficient in granzyme B have demonstrated that this mediator is not critical to the cytotoxic function of CTL (14), and it is unclear how blocking granzyme B activity could by itself be sufficient to prevent cell-mediated killing. Additional serine protease inhibitors, such as SPI-CI (serine protease inhibitor involved in cytotoxicity inhibition), are reported to inhibit the activity of other granzymes, and they may cooperate with SPI-6 in conferring protection from cytotoxic function (15).

Killing by CTL interferes with the ability of DC to elicit cytokine production and optimally boost CD8+ T cell responses (4, 16), and it may represent a barrier to the optimal delivery of immunotherapy in vivo. Making DC resistant to CTL-mediated killing in vivo would have important applications in vaccines against infectious diseases and cancer. We therefore examined whether stimuli that induce the up-regulation of SPI-6 in DC also make them resistant to CTL-mediated killing in vivo.

All mice were maintained at the Malaghan Institute Biomedical Research Unit, Victoria University of Wellington. Experimental procedures were approved by the relevant Animal Ethics Committee and conducted in accordance with institutional guidelines.

C57BL/6 mice were from breeding pairs originally obtained from The Jackson Laboratory. Breeding pairs for the TCR transgenic line 318 (17) were gifted by Prof. H. Pircher (University of Freiburg, Freiburg, Germany) and were maintained by brother × sister mating.

All cultures were in complete IMDM (cIMDM), which consisted of IMDM supplemented with 5% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 55 μM 2-ME (all from Invitrogen).

The gp33 peptide (LCMV33–41 KAVYNFATM) was purchased from Mimotopes.

DC were prepared from the bone marrow of C57BL/6 mice by culturing in GM-CSF and IL-4 as previously described (18). After 7 days of culture, loosely adherent cells were harvested by gentle pipetting. In some cases, DC were activated by adding 0.1, 1.0, or 10 μg/ml of LPS (Sigma-Aldrich) during the last 24–48 h of culture as indicated. Alternatively, DC were activated by culturing on a monolayer of CD40L-expressing NIH-3T3 cells (19) and separated from contaminating fibroblasts using anti-CD11c magnetic beads (Miltenyi Biotec) and magnetic selection. DC were routinely checked for purity and maturation status by assessing expression of CD11c, MHC II, CD40, CD80, and CD86 by flow cytometry.

For immunizations, DC were harvested and resuspended at 1 × 106 cells/ml. Peptide was added to the cells at a final concentration of 10 μM and cells were incubated at 37°C for 2 h. Cells were pelleted, washed three times in IMDM to remove excess peptide, and injected s.c. in the flank of mice in 100 μl.

DC were washed, pelleted, and resuspended at 5 × 106 cells/ml in TRIzol reagent (Invitrogen) for RNA preparation; contaminating DNA was removed using the DNA-free kit (Ambion). cDNA was synthesized from 5 μg RNA using the SuperScript First Strand kit (Invitrogen) and random hexamer primers. For real-time PCR, the following primers were used (Sigma-Aldrich): SPI-6 forward, TTC CAC CTT GCT GAG GTC CA, SPI-6 reverse, CAG TGC AGA TGA TGT GTC GTG; 18S forward, GTA ACC CGT TGA ACC CCA TT; 18S reverse, CCA TCC AAT CGG TAG TAG CG. For amplification reactions, cDNA, primers and SYBR Green PCR MasterMix (ABgene) were mixed in triplicate and run on an ABI Prism 7700 sequence detection system (Applied Biosystems) for 40 cycles of 15 s at 95°C and 1 min at 62°C with a hot start of 15 min at 95°C. The amplified product was homogeneous by melting temperature, and its identity as SPI-6 was confirmed by sequencing. For each sample the number of cycles required to reach a predetermined signal magnitude was calculated, and the expression of SPI-6 in treated DC (Exp) relative to untreated control DC (Cont) was calculated as 2−ΔΔCt, where ΔΔCt = (No. of cycles for serpinExp − No. of cycles for 18SExp) − (No. of cycles for serpinCont − No. of cycles for 18SCont).

DC were purified using anti-CD11c magnetic beads (Miltenyi Biotec), washed, pelleted, and resuspended at 1 × 107 cells/ml in cell lysis buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.9% Triton X-100, 1 mM PMSF, and 1 mM EDTA), and the equivalent of 25 μg protein was applied to a 4–12% SDS-acrylamide gel (Invitrogen) for protein separation. Proteins were transferred to a nitrocellulose membrane (Bio-Rad) and probed with monoclonal anti-human PI-9 (Alexis Biochemicals) or anti-actin 20-33 (Sigma-Aldrich) followed by goat anti-mouse Ig HRP (Dako) or goat anti-rabbit IgG HRP (Santa Cruz Biotechnology), respectively. Bound Abs were revealed using SuperSignal West Pico chemiluminescent substrate (Pierce Biotechnology) and quantified using Kodak digital science 1D software (Eastman Kodak).

Total lymph node cells from L318 mice were cultured with LPS-treated DC and 1 μM gp33 at a ratio of five lymph node cells to one DC for 4 days, and then in 100 U/ml rhIL-2 for 2–3 days. At the end of culture activated T cells were harvested, washed in IMDM, and used for in vivo or in vitro cytotoxicity assays. This protocol routinely yielded cell populations that were 90–95% Vα2+Vβ8+CD62Llowgranzyme B+ by FACS analysis.

DC were harvested, resuspended at 1 × 106/ml, and incubated in the presence or absence of 1 μM gp33 for 1 h at 37°C. DC were plated at 5 × 104/well in round-bottom 96-well plates (BD Biosciences) and in vitro-activated CTL were added to the well. Plates were gently spun to allow contact of effector and target cells and incubated for 4 h at 37°C. At the end of incubation plates were spun, washed in FACS buffer, and anti-CD8-FITC and anti-CD11c-allophycocyanin or anti-CD11c-Alexa 647 Abs were added to each well. Cell suspensions were analyzed on a FACSCalibur (BD Biosciences) and analyzed using FlowJo software (Tree Star). For analysis, the CD11c+CD8 population in each well was identified and gated, and the percentage of propidium idodide-positive cells within this populations was assessed as percentage DC killing.

DC were labeled with fluorescent dyes so that their survival in vivo could be examined (3). DC were left untreated or treated with different doses of LPS as indicated and injected into mice that had been immunized with DC and Ag 6–8 days or 7 wk earlier. One population of DC was incubated with 1 μM gp33 peptide and labeled with the green fluorescent dye CFSE (Molecular Probes) while the second population of DC was labeled with the orange dye Cell Tracker Orange (CTO; Molecular Probes) as described (3). Each mouse received 1 × 106 CFSE-labeled DC and 1 × 106 CTO-labeled DC mixed in a volume of 50 μl and injected intradermally into the distal forelimb. At various times after injection, draining axillary and brachial lymph nodes were harvested, incubated for 1 h at 37°C in an enzyme cocktail containing 2.4 mg/ml collagenase II (Invitrogen) and 0.1 mg/ml DNase I (Sigma-Aldrich), and passed several times through a 21-gauge needle and then gauze before FACS analysis. The percentage of surviving DC in individual lymph nodes was calculated according to the formula (No. Ag-loaded DC/No. non-Ag-loaded DC in immune lymph node) × 100/(No. Ag-loaded/No. non-Ag-loaded DC in control lymph nodes).

Statistical analyses were conducted using the GraphPad Prism software using the tests indicated in the individual figures. Student’s t test was used for data sets that passed the normality test; the Mann-Whitney nonparametric test was used in all other cases.

To determine whether activation of DC from bone marrow cultures induces the up-regulation of SPI-6 as described (7), we treated DC with different doses of LPS in vitro. Alternatively, DC were activated by culture on monolayers of mouse fibroblasts expressing high levels of CD40L, while control DC were cocultured on mock-transfected fibroblasts. Both of these methods induced DC to markedly up-regulate MHC II, CD80, and CD86 molecules and to secrete cytokines such as TNF-α, IL-6, and IL-12 (data not shown).

In initial experiments we determined the levels of SPI-6 mRNA in DC using real-time RT-PCR. As shown in Fig. 1,A, treatment with LPS at 100 ng/ml induced a 5–10-fold up-regulation of SPI-6 mRNA. Increased mRNA levels were already detectable 6 h after addition of LPS (data not shown), remained elevated at 24 h, and were declining by 48 h but were still higher than in untreated DC. Treatment with 1 μg/ml LPS (Fig. 1,A) or 10 μg/ml LPS (not shown) did not induce any further up-regulation of SPI-6 mRNA compared with 100 ng/ml LPS. Coculture with CD40L-expressing fibroblasts also induced up-regulation of SPI-6 mRNA in bone marrow-derived DC (BM-DC) (Fig. 1 B); again, this was comparable to the up-regulation induced by treatment with 100 ng/ml LPS. Similar results were obtained when SPI-6 and SPI-CI mRNA levels were examined in BM-DC using the premade TaqMan gene expression assay (not shown).

FIGURE 1.

Increased expression of SPI-6 mRNA in DC activated by LPS or by coculture with CD40L-expressing fibroblasts. DC were cultured from BM precursors using GM-CSF and IL-4. On day 6 DC were activated by adding the indicated final concentrations of LPS to the cultures, or by coculturing DC with CD40L-expressing fibroblasts, or control fibroblasts, for the indicated times. At the end of culture, cells were harvested, DC were positively selected using anti-CD11c magnetic beads, lysed in TRIzol, and cDNA was prepared for quantitative real-time PCR. Expression of SPI-6 in each sample was normalized to 18S RNA as an internal control and is shown as fold increase relative to untreated DC. Averages ± SD for three to six samples are shown. *, 0.01< p < 0.05; **, 0.001 < p < 0.01 by the nonparametric Mann-Whitney U test. Data are from one of three separate experiments that gave similar results.

FIGURE 1.

Increased expression of SPI-6 mRNA in DC activated by LPS or by coculture with CD40L-expressing fibroblasts. DC were cultured from BM precursors using GM-CSF and IL-4. On day 6 DC were activated by adding the indicated final concentrations of LPS to the cultures, or by coculturing DC with CD40L-expressing fibroblasts, or control fibroblasts, for the indicated times. At the end of culture, cells were harvested, DC were positively selected using anti-CD11c magnetic beads, lysed in TRIzol, and cDNA was prepared for quantitative real-time PCR. Expression of SPI-6 in each sample was normalized to 18S RNA as an internal control and is shown as fold increase relative to untreated DC. Averages ± SD for three to six samples are shown. *, 0.01< p < 0.05; **, 0.001 < p < 0.01 by the nonparametric Mann-Whitney U test. Data are from one of three separate experiments that gave similar results.

Close modal

Expression of SPI-6 protein in LPS-treated and control DC was determined using Western blotting. As shown in Fig. 2,A, some SPI-6 expression was already detected in untreated DC. Treatment with 100 ng/ml LPS induced a further increase in SPI-6 expression, which was clearly apparent at 24 h and remained similarly high at 48 h. Treatment with higher doses of LPS did not cause further increases in SPI-6 protein levels (Fig. 2 B). We conclude that activation by LPS or CD40L induces increased expression of SPI-6 in DC.

FIGURE 2.

Increased expression of SPI-6 protein in DC treated with LPS. DC were cultured from BM precursors using GM-CSF and IL-4, and on day 6 LPS was added to the cultures at the indicated final concentrations. At the indicated times cells were harvested and DC were positively selected using anti-CD11c magnetic beads and lysed. Proteins were separated by SDS-PAGE, transferred to a membrane, and immunoblotted using the indicated Abs. Loading was normalized using an actin-specific Ab. A, SPI-6 expression in DC treated with 100 ng/ml LPS. B, SPI-6 expression in DC treated with 100 ng or 10 μg LPS/ml. Bars represent normalized SPI-6 expression in individual samples (arbitrary units). Data are from one of two separate experiments that gave the same result.

FIGURE 2.

Increased expression of SPI-6 protein in DC treated with LPS. DC were cultured from BM precursors using GM-CSF and IL-4, and on day 6 LPS was added to the cultures at the indicated final concentrations. At the indicated times cells were harvested and DC were positively selected using anti-CD11c magnetic beads and lysed. Proteins were separated by SDS-PAGE, transferred to a membrane, and immunoblotted using the indicated Abs. Loading was normalized using an actin-specific Ab. A, SPI-6 expression in DC treated with 100 ng/ml LPS. B, SPI-6 expression in DC treated with 100 ng or 10 μg LPS/ml. Bars represent normalized SPI-6 expression in individual samples (arbitrary units). Data are from one of two separate experiments that gave the same result.

Close modal

To determine whether the increased expression of SPI-6 was sufficient to confer resistance to CTL-mediated killing, we used DC that had been left untreated, or treated with different doses of LPS, as targets of CTL assays in vitro. As effector cells we used L318 TCR transgenic T cells that had been activated by coculturing with LPS-activated DC and gp33 peptide. These T cells were >95% CD62Llow and granzyme B+ (Fig. 3), and expressed high specific cytotoxic activity against EL-4 targets (data not shown). CTL and DC were mixed together in 96-well plates for 4 h, and DC killing was quantified by flow cytometry by determining the percentage of propidium iodide-positive cells in the CD11c+ population (Fig. 4,A). As shown in Fig. 4,B, DC that had been pretreated with LPS for 24 h were clearly being killed by CTL in vitro, but were markedly less sensitive to killing than DC not treated with LPS. At least 3-fold as many CTL were required in order for LPS-treated DC to be killed to a level comparable to untreated DC. Treatment with higher doses of LPS (1 μg/ml vs 100 ng/ml) did not appear to increase the resistance to killing. This result is consistent with the results in Figs. 1 and 2, showing that DC treated with 100 ng/ml to 10 μg/ml LPS expressed similar amounts of SPI-6 mRNA and protein.

FIGURE 3.

In vitro-activated CTL express granzyme B. Lymph node cell suspensions from L318 TCR transgenic mice were cultured in vitro in the presence of DC and the specific gp33 peptide for 4 days, harvested, and cultured for 3 more days in IL-2-containing medium. Cultured cells were harvested and examined for expression of the surface activation markers CD44 and CD62L by flow cytometry. Additional cells were fixed, permeabilized, and examined for expression of the intracellular granule protein granzyme B, Cells were >95% Vα2+Vβ8+ at the time of analysis. Filled gray histograms represent unstained/isotype control; black lines, Ab staining. Data are from one of three separate experiments that gave similar results.

FIGURE 3.

In vitro-activated CTL express granzyme B. Lymph node cell suspensions from L318 TCR transgenic mice were cultured in vitro in the presence of DC and the specific gp33 peptide for 4 days, harvested, and cultured for 3 more days in IL-2-containing medium. Cultured cells were harvested and examined for expression of the surface activation markers CD44 and CD62L by flow cytometry. Additional cells were fixed, permeabilized, and examined for expression of the intracellular granule protein granzyme B, Cells were >95% Vα2+Vβ8+ at the time of analysis. Filled gray histograms represent unstained/isotype control; black lines, Ab staining. Data are from one of three separate experiments that gave similar results.

Close modal
FIGURE 4.

LPS-treated DC are partially resistant to CTL-mediated killing in vitro. DC were generated as described in Fig. 1 and treated with the indicated concentrations of LPS for 18–20 h, or left untreated. At the end of culture DC were harvested and incubated for 4 h with different numbers of in vitro-activated CTL, in the presence or absence of gp33 peptide Ag, to assess susceptibility to killing. A, Representative dot plots of CD11c+CD8 gated events from an experiment using DC that were pretreated with 100 ng/ml LPS or left untreated, and a ratio of 30 effectors to 1 target. B, Averages + SD for triplicate wells are shown. In some cases error bars are hidden by the symbol. The two panels in B refer to separate experiments. **, p < 0.01 by a nonparametric Mann-Whitney U test. Each of the panels in B is from one of two separate experiments that gave the same result.

FIGURE 4.

LPS-treated DC are partially resistant to CTL-mediated killing in vitro. DC were generated as described in Fig. 1 and treated with the indicated concentrations of LPS for 18–20 h, or left untreated. At the end of culture DC were harvested and incubated for 4 h with different numbers of in vitro-activated CTL, in the presence or absence of gp33 peptide Ag, to assess susceptibility to killing. A, Representative dot plots of CD11c+CD8 gated events from an experiment using DC that were pretreated with 100 ng/ml LPS or left untreated, and a ratio of 30 effectors to 1 target. B, Averages + SD for triplicate wells are shown. In some cases error bars are hidden by the symbol. The two panels in B refer to separate experiments. **, p < 0.01 by a nonparametric Mann-Whitney U test. Each of the panels in B is from one of two separate experiments that gave the same result.

Close modal

To determine whether the partial resistance to CTL-mediated killing in vitro was reproduced in vivo, we used DC that had been activated with LPS as targets of CTL in vivo. DC were labeled with different fluorochromes, loaded with gp33 peptide or left untreated, and injected into naive C57BL/6 mice or into C57BL/6 mice that had been immunized with DC + gp33 peptide 1 wk earlier to induce a specific CTL response. Different DC activation conditions were tested, with LPS concentrations ranging from 100 ng/ml to 10 μg/ml and treatment times from 18 to 48 h; results from some of these experiments are shown in Fig. 5 A–C. In all cases, both populations of DC survived well after injection into nonimmune recipients, and, as previously reported (20, 21), LPS-treated DC could be recovered in higher numbers than untreated DC from lymph nodes. In contrast, when injected into gp33-immune hosts, gp33-loaded DC failed to survive and could not be recovered from the draining lymph node. This was the case whether the DC had been treated for 24 or 48 h with 100 ng–10 μg/ml LPS, or left untreated. Thus, LPS-treated DC did not become resistant to CTL-mediated killing in vivo.

FIGURE 5.

LPS-treated DC remain sensitive to CTL-mediated killing in vivo. DC were generated and treated with LPS as described in Fig. 1, or left untreated. After 24–48 h DC were harvested, labeled with CFSE or CTO, loaded with gp33 peptide or left untreated, and mixed in equal numbers before s.c. injection into immune or nonimmune mice. Draining lymph nodes were recovered 48 h later, and DC were quantified by flow cytometry. A and B, DC treated with 10 μg/ml LPS for 24 h were injected into mice that had been immunized 1 wk earlier with gp33-loaded DC (Imm) or DC only (NI). DC numbers are shown in A; relative percentages of gp33-loaded DC for the same experiment are shown in B. C, DC treated with 100 ng or 1 μg/ml LPS for 48 h were injected into mice that had been immunized with gp33-loaded DC, or DC only, 1 wk earlier. Relative percentages of gp33-loaded DC are shown. D and E, DC treated with 1 μg/ml LPS for 48 h were injected into mice that had been adoptively transferred with 5 × 106 in vitro-activated CTL 1 day earlier. DC numbers are shown in D; relative percentages of gp33-loaded DC for the same experiment are shown in E. F, As in E, but mice were adoptively transferred with different numbers of CTL as indicated. In all panels, bars represent the average + SD of six samples per group. Ns, not significant; *, 0.01 < p < 0.05; ***, p < 0.001 in a two-tailed Student’s t test.

FIGURE 5.

LPS-treated DC remain sensitive to CTL-mediated killing in vivo. DC were generated and treated with LPS as described in Fig. 1, or left untreated. After 24–48 h DC were harvested, labeled with CFSE or CTO, loaded with gp33 peptide or left untreated, and mixed in equal numbers before s.c. injection into immune or nonimmune mice. Draining lymph nodes were recovered 48 h later, and DC were quantified by flow cytometry. A and B, DC treated with 10 μg/ml LPS for 24 h were injected into mice that had been immunized 1 wk earlier with gp33-loaded DC (Imm) or DC only (NI). DC numbers are shown in A; relative percentages of gp33-loaded DC for the same experiment are shown in B. C, DC treated with 100 ng or 1 μg/ml LPS for 48 h were injected into mice that had been immunized with gp33-loaded DC, or DC only, 1 wk earlier. Relative percentages of gp33-loaded DC are shown. D and E, DC treated with 1 μg/ml LPS for 48 h were injected into mice that had been adoptively transferred with 5 × 106 in vitro-activated CTL 1 day earlier. DC numbers are shown in D; relative percentages of gp33-loaded DC for the same experiment are shown in E. F, As in E, but mice were adoptively transferred with different numbers of CTL as indicated. In all panels, bars represent the average + SD of six samples per group. Ns, not significant; *, 0.01 < p < 0.05; ***, p < 0.001 in a two-tailed Student’s t test.

Close modal

It was possible that CTL elicited by in vivo immunization may express different killing mediators compared with in vitro-activated cells, thus explaining the difference between the in vivo and in vitro DC killing results. To test this possibility we used in vitro-activated CTL, which express granzyme B, and transferred them i.v. into naive mice. As shown in Fig. 3, these CTL populations appeared essentially homogeneous in terms of granzyme B expression, suggesting that differential distribution in vivo should have little effect on killing in different tissues. One day after CTL transfer, the same mice were injected with target DC that had been treated with 1 μg/ml LPS, or left untreated. As shown in Fig. 5, D and E, LPS-activated DC were susceptible to CTL-mediated killing, and in several experiments they were significantly more so than were untreated DC. This was the case even when DC were injected into mice that had received limiting numbers of CTL, sufficient to kill only a proportion of the injected DC (Fig. 5 F). We conclude that the differential sensitivity of LPS-treated DC to CTL-mediated killing in vivo and in vitro could not be explained by different characteristics of the effector CTL.

We have reported that memory CTL are also able to kill DC in vivo (2). Killing by memory CTL is delayed compared with killing by effector CTL, presumably to allow the reactivation of memory cells to effector function. We wanted to determine whether LPS-activated DC might become resistant to killing by memory CTL. DC were activated by treatment with 1 μg/ml LPS, or left untreated, and injected into mice that had been immunized with DC + gp33 peptide 7 wk earlier to induce a memory CTL response. As shown in Fig. 6, 24 h after injection all DC populations could be recovered from the lymph nodes of injected mice, regardless of LPS or gp33 treatment. At 72 h after injection, the proportion of gp33-loaded DC in the lymph nodes of immunized mice had declined compared with controls (Fig. 6). Both LPS-treated and untreated DC were decreased, but the decrease was statistically significant only in the case of LPS-treated DC. Therefore, LPS-treated DC were not resistant to killing by reactivated memory CD8+ T cells.

FIGURE 6.

LPS-treated DC remain sensitive to killing by memory CTL in vivo. DC were generated and treated with 1 μg/ml LPS as described in Fig. 1, or left untreated. After 24 h DC were labeled as described in Fig. 5 and injected into mice that had been immunized 7 wk earlier with gp33-loaded DC (Imm) or DC only (NI). Draining lymph nodes were recovered 24 (A) or 72 h (B) later, and DC were quantified by flow cytometry. Relative percentages of gp33-loaded DC are shown. Bars represent the average + SD of six samples per group. Ns, not significant; ***, p < 0.001 in a two-tailed Student’s t test.

FIGURE 6.

LPS-treated DC remain sensitive to killing by memory CTL in vivo. DC were generated and treated with 1 μg/ml LPS as described in Fig. 1, or left untreated. After 24 h DC were labeled as described in Fig. 5 and injected into mice that had been immunized 7 wk earlier with gp33-loaded DC (Imm) or DC only (NI). Draining lymph nodes were recovered 24 (A) or 72 h (B) later, and DC were quantified by flow cytometry. Relative percentages of gp33-loaded DC are shown. Bars represent the average + SD of six samples per group. Ns, not significant; ***, p < 0.001 in a two-tailed Student’s t test.

Close modal

In this study we investigated whether activation status affects the susceptibility of DC to CTL-mediated killing in vivo. We have previously reported that Ag-presenting DC are sensitive to CTL-mediated killing in vivo (2), and that this killing plays an important role in regulating the expansion of Ag-specific CD8+ T cells (4). However, DC treated with LPS or anti-CD40 have been described to become resistant to CTL-induced apoptosis in vitro (7). We wanted to establish whether activation-induced resistance to killing also occurs in vivo.

Initial experiments were aimed at establishing whether the described resistance of activated DC to CTL-mediated killing in vitro could be reproduced in our system. These experiments confirmed that, as reported by Medema et al. using a DC line (7), BM-DC treated with LPS or CD40L up-regulates expression of the granzyme B inhibitor SPI-6. The same treatments also induced up-regulation of at least another serpin, SPI-CI (data not shown), which is reported to block the function of granzyme M (15), suggesting that activated DC express several factors that may enable them to resist killing. Besides showing SPI-6 up-regulation, our experiments also confirmed that activated DC were less sensitive than untreated DC to CTL-mediated killing in vitro. While Medema et al. used a DNA fragmentation assay, which is a readout of granzyme B activity but does not detect target cell lysis (22), we used a propidium iodide uptake assay that reveals loss of membrane integrity and detects both granzyme A+B-dependent and independent killing (14). Using this assay, we were able to show that LPS-treated DC were killed to a lower extent than untreated DC. This observation appears consistent with reports that CTL lacking expression of granzyme A+B do not exhibit an absolute defect in target cell killing, although their killing is slower than the killing of normal CTL (14, 23, 24).

When CTL-mediated killing was examined in vivo, we were surprised to find that both untreated and LPS-treated DC were killed to a similar extent in mice that had either been recently immunized or that had immunological memory from distant immunization. Similarly, both DC populations were similarly sensitive to killing when exposed to limiting numbers of effector CTL in vivo. The discrepancy between in vitro and in vivo results was not due to the use of CTL populations with different characteristics, for example, differentially expressing the SPI-6 target molecule granzyme B. When the same CTL populations were used both in vivo and in vitro, reduced killing of LPS-treated DC was observed in vitro, but not in vivo. Thus, the differential sensitivity of LPS-treated DC to CTL-mediated killing in vitro and in vivo must be explained by mechanisms other than the characteristics of the effector CTL.

One possible explanation for the differential resistance to killing in vivo and in vitro is the duration of contacts between effector CTL and target DC. As resistance to killing in vitro was only relative, one would expect all target DC, regardless of their activation status, to be eventually eliminated if exposed to CTL for a sufficiently long time. This situation may be occurring in vivo, where contacts between CTL and target DC cannot be artificially manipulated as in the in vitro situation. Unfortunately, we do not have sufficient information on the duration of DC-CTL contacts in vivo vs in vitro to evaluate the likelihood of this possibility.

A second possibility is that other differences between in vivo and in vitro conditions may result in the relative resistance of activated DC to CTL-mediated killing in vitro, but not in vivo. In vitro, high cell concentrations and high E:T ratios may result in target cells being killed in a bystander fashion, due to the effects of cytotoxic mediators released by neighboring CTL during the killing of specific targets (25, 26, 27). In contrast, the interaction of CTL and target DC in vivo is likely to involve lower E:T ratios, as well as conditions that are less compatible with bystander killing. It is interesting to note that protection from bystander killing is one of the roles that were originally proposed for SPI-6 when its expression in different cell types was first described (12, 28).

Rapid down-regulation of SPI-6 expression in DC injected in vivo might be another potential reason why DC are resistant to CTL-mediated killing in vitro, but not in vivo. Due to the low numbers of DC that accumulate in the draining lymph node, we were not able to examine expression of SPI-6 in DC after injection. However, in vivo killing by effector CTL appears to take place relatively rapidly during the first 6–20 h after injection. Even if SPI-6 expression declined over time, we would expect that at least some resistance to killing should still be apparent in our in vivo experiments. In contrast, no decrease in killing could be observed, and LPS-treated DC often appeared slightly more, rather than less, sensitive to CTL-mediated killing. Other groups have also reported that DC may become resistant to CTL-mediated killing in some circumstances (8, 9). Mueller at al. have reported that CD4+ T cells, and activation by LPS or CD40L, can support the survival of DC in vivo in the presence of a preexisting CTL response (9). It is unclear why in our hands LPS was not sufficient to prevent DC killing in vivo. Clearly, SPI-6 was up-regulated in our LPS-treated DC, to a level that was sufficient to decrease susceptibility to CTL killing in vitro. It is possible that additional molecules may be required to induce resistance of DC to killing in vivo, and these molecules may not be fully induced in our DC. In support of this possibility, we used BM-DC, while Mueller et al. used DC generated from spleen. These cells did differ in their response to LPS in additional respects besides the resistance to killing; for example, LPS treatment dramatically increased the migration of our BM-DC to the lymph node (Ref. 20, 21 and this paper), but did not affect the migration of spleen DC (9). It then appears possible that the DC used in each case may be representative of different subpopulations of DC, such as inflammatory, tissue-derived DC vs lymph node-resident DC, and that these DC may also differ in their response to activating signals such as LPS.

The question of whether DC that present specific Ag in the context of their MHC-I should be sensitive or resistant to CTL-mediated killing has been debated in the literature. DC are critical to the generation of immune responses and as such their survival must be preserved. For example, DC resident in tissues where active immune responses are taking place, such as the respiratory epithelium during influenza infection, are likely to be exposed to cytotoxic mediators released by NK cells and CTL attacking neighboring infected cells. The bystander elimination of these DC would compromise the immune status of the tissue, leaving it especially vulnerable to subsequent infections. Mechanisms that support DC survival would clearly be beneficial in this case. On the other hand, DC that become infected, or are cross-presenting Ag to which a strong effector CTL response already exists, are likely to be redundant or even deleterious to the immune response, and protecting their survival would most likely be counterproductive.

Our data indicating that DC that have up-regulated expression of SPI-6, and possibly of other serpins, become relatively resistant to CTL in vitro but not in vivo might suggest that SPI-6 can protect DC from CTL activity, although not from direct killing. Immature, tissue-resident DC have been shown to express low but detectable levels of SPI-6 (11), and our data with untreated DC (Figs. 1 and 2) and DC freshly isolated from untreated spleen (Ref. 11 and K. A. Andrew, unpublished) also suggest a similar low level of expression. SPI-6 expression can be further up-regulated in DC upon exposure to infectious stimuli, and this up-regulation is likely to be functionally important given the preferential localization of effector CTL to nonlymphoid tissues (29). Once DC migrate from tissues to lymph nodes, they may be relatively protected from the effects of CTL, as CTL are often excluded from the lymph node environment. However, continued expression of SPI-6 is likely to remain important to preserve lymph node integrity in circumstances when CTL enter reactive lymph nodes (6). In this regard, it would be interesting to determine whether SPI-6 may be differentially expressed in subpopulations of DC in lymph nodes and spleen, given their differential ability to cross-present Ag in the context of MHC-I (30).

In conclusion, we show that DC treated with LPS express increased levels of SPI-6, but remain sensitive to CTL-mediated killing in vivo. Our results then suggest that SPI-6 expression may have functions that are important in preserving DC integrity, but are distinct from the protection of DC from direct CTL killing. Further studies using SPI-6-deficient DC (10) may be useful in resolving this question.

The authors thank all staff at the Malaghan Institute of Medical Research for help and suggestions, Dr. Phil Bird (Monash University, Australia) for useful discussions, and all staff at the Biomedical Research Unit for excellent animal husbandry.

The authors have no financial conflicts of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was funded by a research grant from the Royal Society of New Zealand Marsden Fund and the Health Research Council of New Zealand to F.R. H.S. was the recipient of a Lottery Health Ph.D. Scholarship, and R.P. was the recipient of a University of Otago Ph.D. Scholarship.

5

Abbreviations used in this paper: DC, dendritic cells; BM-DC, bone marrow-derived DC; CTO, Cell Tracker Orange; gp33, fragment 33–41 of the lymphocytic choriomeningitis virus glycoprotein; SPI-6, serine protease inhibitor 6; SPI-CI, serine protease inhibitor involved in cytotoxicity inhibition.

1
Guermonprez, P., J. Valladeau, L. Zitvogel, C. Thery, S. Amigorena.
2002
. Antigen presentation and T cell stimulation by dendritic cells.
Annu. Rev. Immunol.
20
:
621
-667.
2
Hermans, I. F., D. S. Ritchie, J. Yang, J. M. Roberts, F. Ronchese.
2000
. CD8+ T cell-dependent elimination of dendritic cells in vivo limits the induction of antitumor immunity.
J. Immunol.
164
:
3095
-3101.
3
Ritchie, D. S., I. F. Hermans, J. M. Lumsden, C. B. Scanga, J. M. Roberts, J. Yang, R. A. Kemp, F. Ronchese.
2000
. Dendritic cell elimination as an assay of cytotoxic T lymphocyte activity in vivo.
J. Immunol. Methods
246
:
109
-117.
4
Yang, J., S. P. Huck, R. S. McHugh, I. F. Hermans, F. Ronchese.
2006
. Perforin-dependent elimination of dendritic cells regulates the expansion of antigen-specific CD8+ T cells in vivo.
Proc. Natl. Acad. Sci. USA
103
:
147
-152.
5
Belz, G. T., L. Zhang, M. D. Lay, F. Kupresanin, M. P. Davenport.
2007
. Killer T cells regulate antigen presentation for early expansion of memory, but not naive, CD8+ T cell.
Proc. Natl. Acad. Sci. USA
104
:
6341
-6346.
6
Guarda, G., M. Hons, S. F. Soriano, A. Y. Huang, R. Polley, A. Martin-Fontecha, J. V. Stein, R. N. Germain, A. Lanzavecchia, F. Sallusto.
2007
. L-selectin-negative CCR7 effector and memory CD8+ T cells enter reactive lymph nodes and kill dendritic cells.
Nat. Immunol.
8
:
743
-752.
7
Medema, J. P., D. H. Schuurhuis, D. Rea, J. van Tongeren, J. de Jong, S. A. Bres, S. Laban, R. E. Toes, M. Toebes, T. N. Schumacher, et al
2001
. Expression of the serpin serine protease inhibitor 6 protects dendritic cells from cytotoxic T lymphocyte-induced apoptosis: differential modulation by T helper type 1 and type 2 cells.
J. Exp. Med.
194
:
657
-667.
8
Watchmaker, P. B., J. A. Urban, E. Berk, Y. Nakamura, R. B. Mailliard, S. C. Watkins, S. M. van Ham, P. Kalinski.
2008
. Memory CD8+ T cells protect dendritic cells from CTL killing.
J. Immunol.
180
:
3857
-3865.
9
Mueller, S. N., C. M. Jones, A. T. Stock, M. Suter, W. R. Heath, F. R. Carbone.
2006
. CD4+ T cells can protect APC from CTL-mediated elimination.
J. Immunol.
176
:
7379
-7384.
10
Zhang, M., S. M. Park, Y. Wang, R. Shah, N. Liu, A. E. Murmann, C. R. Wang, M. E. Peter, P. G. Ashton-Rickardt.
2006
. Serine protease inhibitor 6 protects cytotoxic T cells from self-inflicted injury by ensuring the integrity of cytotoxic granules.
Immunity
24
:
451
-461.
11
Bladergroen, B. A., M. C. Strik, N. Bovenschen, O. van Berkum, G. L. Scheffer, C. J. Meijer, C. E. Hack, J. A. Kummer.
2001
. The granzyme B inhibitor, protease inhibitor 9, is mainly expressed by dendritic cells and at immune-privileged sites.
J. Immunol.
166
:
3218
-3225.
12
Buzza, M. S., C. E. Hirst, C. H. Bird, P. Hosking, J. McKendrick, P. I. Bird.
2001
. The granzyme B inhibitor, PI-9, is present in endothelial and mesothelial cells, suggesting that it protects bystander cells during immune responses.
Cell. Immunol.
210
:
21
-29.
13
Bird, C. H., V. R. Sutton, J. Sun, C. E. Hirst, A. Novak, S. Kumar, J. A. Trapani, P. I. Bird.
1998
. Selective regulation of apoptosis: the cytotoxic lymphocyte serpin proteinase inhibitor 9 protects against granzyme B-mediated apoptosis without perturbing the Fas cell death pathway.
Mol. Cell. Biol.
18
:
6387
-6398.
14
Waterhouse, N. J., V. R. Sutton, K. A. Sedelies, A. Ciccone, M. Jenkins, S. J. Turner, P. I. Bird, J. A. Trapani.
2006
. Cytotoxic T lymphocyte-induced killing in the absence of granzymes A and B is unique and distinct from both apoptosis and perforin-dependent lysis.
J. Cell Biol.
173
:
133
-144.
15
Bots, M., I. G. Kolfschoten, S. A. Bres, M. T. Rademaker, G. M. de Roo, M. Kruse, K. L. Franken, M. Hahne, C. J. Froelich, C. J. Melief, R. Offringa, J. P. Medema.
2005
. SPI-CI and SPI-6 cooperate in the protection from effector cell-mediated cytotoxicity.
Blood
105
:
1153
-1161.
16
Sad, S., D. Kagi, T. R. Mosmann.
1996
. Perforin and Fas killing by CD8+ T cells limits their cytokine synthesis and proliferation.
J. Exp. Med.
184
:
1543
-1547.
17
Pircher, H., K. Burki, R. Lang, H. Hengartner, R. M. Zinkernagel.
1989
. Tolerance induction in double specific T-cell receptor transgenic mice varies with antigen.
Nature
342
:
559
-561.
18
Garrigan, K., P. Moroni-Rawson, C. McMurray, I. Hermans, N. Abernethy, J. Watson, F. Ronchese.
1996
. Functional comparison of spleen dendritic cells and dendritic cells cultured in vitro from bone marrow precursors.
Blood
88
:
3508
-3512.
19
Kershaw, M. H., C. Hsu, W. Mondesire, L. L. Parker, G. Wang, W. W. Overwijk, R. Lapointe, J. C. Yang, R. F. Wang, N. P. Restifo, P. Hwu.
2001
. Immunization against endogenous retroviral tumor-associated antigens.
Cancer Res.
61
:
7920
-7924.
20
Huck, S. P., S. C. Tang, K. A. Andrew, J. Yang, J. L. Harper, F. Ronchese.
2008
. Activation and route of administration both determine the ability of bone marrow-derived dendritic cells to accumulate in secondary lymphoid organs and prime CD8+ T cells against tumors.
Cancer Immunol. Immunother.
57
:
63
-71.
21
Matthews, K. E., J. S. Qin, J. Yang, I. F. Hermans, M. J. Palmowski, V. Cerundolo, F. Ronchese.
2007
. Increasing the survival of dendritic cells in vivo does not replace the requirement for CD4+ T cell help during primary CD8+ T cell responses.
J. Immunol.
179
:
5738
-5747.
22
Simon, M. M., M. Hausmann, T. Tran, K. Ebnet, J. Tschopp, R. ThaHla, A. Mullbacher.
1997
. In vitro- and ex vivo-derived cytolytic leukocytes from granzyme A × B double knockout mice are defective in granule-mediated apoptosis but not lysis of target cells.
J. Exp. Med.
186
:
1781
-1786.
23
Heusel, J. W., R. L. Wesselschmidt, S. Shresta, J. H. Russell, T. J. Ley.
1994
. Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA fragmentation and apoptosis in allogeneic target cells.
Cell
76
:
977
-987.
24
Smyth, M. J., S. E. Street, J. Trapani.
2003
. Cutting edge: granzymes A and B are not essential for perforin-mediated tumor rejection.
J. Immunol.
171
:
515
-518.
25
Isaaz, S., K. Baetz, K. Olsen, E. Podack, G. M. Griffiths.
1995
. Serial killing by cytotoxic T lymphocytes: T cell receptor triggers degranulation, re-filling of the lytic granules and secretion of lytic proteins via a non-granule pathway.
Eur. J. Immunol.
25
:
1071
-1079.
26
Ando, K., K. Hiroishi, T. Kaneko, T. Moriyama, Y. Muto, N. Kayagaki, H. Yagita, K. Okumura, M. Imawari.
1997
. Perforin, Fas/Fas ligand, and TNF-α pathways as specific and bystander killing mechanisms of hepatitis C virus-specific human CTL.
J. Immunol.
158
:
5283
-5291.
27
Kojima, H., N. Shinohara, S. Hanaoka, Y. Someya-Shirota, Y. Takagaki, H. Ohno, T. Saito, T. Katayama, H. Yagita, K. Okumura, et al
1994
. Two distinct pathways of specific killing revealed by perforin mutant cytotoxic T lymphocytes.
Immunity
1
:
357
-364.
28
Sun, J., C. H. Bird, V. Sutton, L. McDonald, P. B. Coughlin, T. A. De Jong, J. A. Trapani, P. I. Bird.
1996
. A cytosolic granzyme B inhibitor related to the viral apoptotic regulator cytokine response modifier A is present in cytotoxic lymphocytes.
J. Biol. Chem.
271
:
27802
-27809.
29
Masopust, D., V. Vezys, A. L. Marzo, L. Lefrançois.
2001
. Preferential localization of effector memory cells in nonlymphoid tissue.
Science
291
:
2413
-2417.
30
Heath, W. R., G. T. Belz, G. M. Behrens, C. M. Smith, S. P. Forehan, I. A. Parish, G. M. Davey, N. S. Wilson, F. R. Carbone, J. A. Villadangos.
2004
. Cross-presentation, dendritic cell subsets, and the generation of immunity to cellular antigens.
Immunol. Rev.
199
:
9
-26.