The Lipopolysaccharide Adjuvant Effect on T Cells Relies on Nonoverlapping Contributions from the MyD88 Pathway and CD11c⁺ Cells

Mice

C57BL/6 mice were purchased from The Jackson Laboratory. MyD88^−/− mice (11, 20) were obtained from Dr. R. Medzhitov (Yale University, New Haven, CT) and were backcrossed 9 times to C57BL/6. CD11c-DT receptor (DTR) Tg mice (27) were a gift from Drs. S. Jung and D. Littman (Skirball Institute, New York, NY), and were backcrossed 10 times to C57BL/6. MyD88^−/− and CD11c-DTR Tg mice were screened by PCR from tail DNA, and SM1 TCR Tg mice (36) (CD4⁺ Thy1.1⁺ V_β2⁺ RAG^−/−) were bred by our laboratory and were also crossed with MyD88^−/− mice. All mice were maintained in the animal facility at the University of Connecticut Health Center under specific pathogen-free conditions and handled in accordance to National Institutes of Health federal guidelines.

Generation of bone marrow chimeras

Bone marrow cells were flushed from femurs and tibias taken from DTR Tg or C57BL/6 mice. RBC were lysed with ammonium chloride, and the remaining cells were resuspended at 1–5 × 10⁷ cells/ml in balanced salt solution (BSS) supplemented with HEPES, l-glutamine, penicillin, streptomycin, and gentamicin sulfate. For bone marrow reconstitution, C57BL/6 mice were gamma-irradiated twice with 550 rad at a 3-h interval, followed by i.v. injection of 1–5 × 10⁶ bone marrow cells. Bone marrow chimeras were allowed to rest for ≥8 wk before use. For experiments, DT (Sigma-Aldrich) was i.p. injected into bone marrow chimeras at 100 ng/mouse or a minimum of 4 ng/g bodyweight at 48–72 h intervals, beginning 3–4 days before immunization.

Immunization schedule

All reagents were diluted in BSS or PBS and injected in a total volume of 0.2 ml. For superantigen studies, mice were injected i.p. with 1 μg of staphylococcal enterotoxin A (SEA; Toxin Tech), followed by Salmonella typhimurium LPS (Sigma-Aldrich) 18 h later. LPS doses were determined from titration studies on individual batches to determine the amount providing maximal T cell survival. For experiments involving MyD88^−/− mice, 20–60 μg of LPS was used for superantigen studies while 75–250 μg of LPS was used for adoptive transfer studies. For experiments involving bone marrow chimeras, 15–100 μg of LPS was used. For adoptive transfer studies, ∼5 × 10⁵ bulk cells from lymph nodes and spleens of SM1 TCR Tg mice were injected i.v., corresponding to ∼1 × 10⁵ SM1 T cells identified as CD4⁺Vβ2⁺Thy1.1⁺. The next day, 100 μg of FL-pep (residues 427–441; Invitrogen Life Technologies) was injected, followed by LPS 18 h later. For experiments involving exogenous OX40 (CD134) stimulation, 50 μg of anti-OX40 (OX86 clone) was injected 5–10 min after FL-pep, and rat IgG (Sigma-Aldrich) was used in the negative control group.

Cell isolation and processing

Spleens, peripheral lymph nodes (PLN; inguinal, axillary, brachial) and mesenteric lymph nodes were crushed through nylon mesh cell strainers (Falcon; BD Biosciences), and spleens were treated with ammonium chloride to lyse RBC. For experiments with bone marrow chimeras, one-third of each spleen was digested with collagenase D (Roche) for ≥30 min at 37°C before being passed through cell strainers and subsequently analyzed for CD11c expression. Liver and lung lymphocytes were obtained as described (37). Briefly, following perfusion livers were crushed through cell strainers, and the cells partitioned on a 35% Percoll (Sigma-Aldrich) gradient. Lungs were cut into pieces and incubated in 1.3 mM EDTA at 37°C for 30 min, followed by collagenase (Invitrogen Life Technologies) for 1 h. Lung cells were then fractionated on a 44 and 67% Percoll gradient (Amersham Biosciences), with lymphocytes partitioning at the interface.

For experiments involving purification of SM1 T cells, liver lymphocytes were combined from two identically treated mice and incubated with biotinylated anti-Thy1.1 mAb (100 μg/ml; eBioscience) on ice for 15 min. Thy1.1-labeled cells were purified with anti-biotin microbeads (Miltenyi Biotec), according to the manufacturer’s protocol. CD11c⁺ cells were obtained from spleens of naive C57BL/6 mice using CD11c microbeads (Miltenyi Biotec).

Cell culturing, staining, and flow cytometry

For superantigen studies, 1 × 10⁶ splenocytes were cultured at 37°C in 0.2 ml of complete tumor medium, consisting of MEM with FBS, amino acids, salts, and antibiotics. Cultures were stimulated with SEA (0.1 μg/ml) for 24 h, and supernatants were analyzed for production of IFN-γ and IL-2 using ELISA kits from BD Biosciences.

For adoptive transfer studies, 1 × 10⁶ splenocytes or liver lymphocytes were cultured in complete tumor medium with or without FL-pep (5 μg/ml) and brefeldin A (5 μg/ml; Calbiochem) for 5 h. PMA (50 ng/ml; Calbiochem) and ionomycin (1 μg/ml; Invitrogen Life Technologies) was added to some cultures for the final 3 h. These cells were stained intracellularly for cytokines.

The following mAbs were purchased from eBioscience: Thy1.1-allophycocyanin, TNF-allophycocyanin, rat IgG1-allophycocyanin (isotype control), CD44-FITC, MHC class II-PE, CD25-PE, rat IgG1-PE. The following mAbs were purchased from BD Biosciences: CD4-PerCP, Thy1.1-PerCP, Thy1.2-PE, TCR Vβ3-PE, IFN-γ-PE, CD8a-allophycocyanin, Thy1.2-allophycocyanin, CD11c-allophycocyanin, and TCR Vβ2-FITC. Surface and intracellular staining was performed as previously described (38). Briefly, cells were resuspended in staining buffer consisting of BSS, 3% FBS, and 0.1% sodium azide. Nonspecific binding was blocked by a solution containing mouse serum, human IgG, and the anti-Fc mAb 2.4G2 (39), followed by incubation with a fluorescently conjugated mAb(s) on ice for 30 min. For intracellular staining, surface staining was performed, followed by fixation with 2% paraformaldehyde, permeabilization with 0.25% saponin, and incubation at room temperature with the anti-cytokine mAb. Flow cytometry was conducted on a FACSCalibur flow cytometer and data was analyzed using CellQuest (BD Biosciences) or FlowJo software (Tree Star).

Statistical analysis

Two-tailed Student’s t tests were performed for either equal variances between the groups or unequal variances. The type of variance was determined by performing F tests, in which F > 0.05 corresponds to equal variances and F < 0.05 corresponds to unequal variances.

MyD88 is required for LPS to promote optimal survival of endogenously activated T cells

MyD88 is important for induction of T cell stimulation (11), but its contribution to precise responses, such as clonal expansion, long-term survival, and effector function, has not been dissected. Because the generation of T cell survival by LPS involves proinflammatory cytokines, is independent of CD28 costimulation (7, 8), and regulates a multifaceted T cell survival program (40), we hypothesized that MyD88 signaling is paramount. We tested this idea in several ways, but first by using a superantigen model to examine endogenous T cell responses. Injection of SEA into mice causes expansion of Vβ3⁺ T cells followed by their deletion through programmed activation-induced cell death. SEA activates both CD4 and CD8 T cells expressing Vβ1, Vβ10, Vβ11, and Vβ17 in addition to Vβ3 (41). Administration of LPS within 24 h after SEA rescues Vβ3⁺ T cells from peripheral deletion by promoting their long-term survival (7, 38, 42).

To study the role of MyD88 during T cell activation, C57BL/6 (wild-type (WT)) and MyD88^−/− mice were immunized with SEA and LPS. Immunization with SEA alone resulted in expansion of CD4 Vβ3 T cells in the spleen and PLN on day 3 followed by deletion (Fig. 1⇓A). MyD88^−/− mice responded normally to SEA alone, demonstrating that T cell priming was unaffected by the absence of MyD88. SEA-specific CD8 T cells also followed this trend, showing normal expansion in both the spleen and PLN of MyD88^−/− mice (Fig. 1⇓B). Injection of LPS 18 h after SEA increased the percentage of CD4 and CD8 T cells expressing Vβ3 in WT mice. Consistent with previous studies (7, 42), the size of the Vβ3 population remained high following initial expansion such as that on day 12, frequencies were approximately five times higher in mice treated with LPS (Fig. 1⇓). In contrast, LPS only had a marginal effect on CD4 or CD8 T cell accumulation in MyD88^−/− mice (Fig. 1⇓, right panels). Notably, following expansion their Vβ3 population underwent significant contraction, and LPS-induced survival was impaired in the MyD88^−/− mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

LPS enhanced T cell survival in a MyD88-dependent manner. C57BL/6 (WT) and MyD88^−/− mice were injected with SEA at time 0 and LPS 18 h later. The Vβ3 T cell population was assessed in PLN and spleens on days 3, 5, and 12 after SEA injection. Unimmunized mice were used for time 0. Shown are the percentages of CD4 cells (A) and CD8 cells (B) expressing Vβ3 for mice immunized with SEA alone (▵) or SEA plus LPS (▴). Frequencies of nonspecific Vβ14 T cells were not increased by treatments (data not shown). Data are pooled from three experiments with a total of six to eight mice per group and shown as mean ± SEM. A two-tailed Student t test revealed that treatment with LPS significantly increased Vβ3 T cell frequencies in WT mice on day 12 with p < 0.001, whereas the corresponding p value for MyD88^−/− mice was larger.

Analysis of T cell numbers revealed a similar trend showing that Vβ3 T cell accumulation was MyD88 dependent (Table I⇓). SEA alone caused Vβ3 T cell numbers to decline by ∼10-fold between days 3 and 12. LPS treatment in WT mice, however, sustained this population so that numbers of CD4 and CD8 T cells expressing Vβ3 decreased by only 2.8- and 1.9-fold, respectively, during this interval (Table I⇓). In contrast, Vβ3 T cell numbers in MyD88^−/− mice treated with SEA and LPS decreased by over 6-fold between days 3 and 12. Further analysis showed that compared with the unimmunized group (Table I⇓, none), SEA with LPS immunization in WT mice boosted CD4 and CD8 Vβ3 T cell numbers by 2.8- and 3.6-fold, respectively, compared with very little change using the same comparison in MyD88^−/− mice. Taken together, these data demonstrated that MyD88 contributed to LPS adjuvanticity by promoting T cell survival following initial expansion.

MyD88 expression by non-T cells contributes to T cell accumulation, not effector differentiation

When examining the endogenous response in MyD88^−/− mice (Fig. 1⇑), both T cells and non-T cells were MyD88 deficient. To separately examine the contribution of MyD88 expression in these compartments, we used an adoptive transfer system that allowed us to track monoclonal naive CD4 T cells. TCR Tg SM1 CD4 T cells, specific for peptide 427–441 from Salmonella flagellin (36), referred to as FL-pep, were transferred into WT or MyD88^−/− mice. The day after, mice were immunized with FL-pep followed by LPS 18 h later. SM1 CD4 T cells were tracked in the blood by combined staining for CD4 and Thy1.1. Interestingly, SM1 T cell accumulation was ≥5-fold lower in MyD88^−/− mice as compared with control mice from day 5 on (Fig. 2⇓A, upper left). Reduced T cell frequencies in the blood may reflect impaired trafficking, however, on day 10, SM1 CD4 T cell frequencies (Fig. 2⇓A, middle left) and numbers (lower left) in the spleen and liver were 4- to 6-fold lower in MyD88^−/− mice. These data demonstrated that LPS drives T cell accumulation through expression of MyD88 in non-T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Requirement of MyD88 expression by non-T cells for optimal T cell accumulation but not for activation or effector function. A, Left, WT (▪) and MyD88^−/− (□) mice received WT SM1 cells and were immunized with FL-pep and LPS. SM1 CD4 T cell frequency in blood (upper left) was determined at various time points. On day 10, mice were sacrificed and the frequencies (middle left) and total numbers (lower left) of SM1 CD4 T cells in the spleen and liver were determined. Data are pooled from three experiments with 9 mice per group. A, Right, MyD88-deficient SM1 cells (gray diamonds) were mixed with WT SM1 cells (dotted diamonds) at a 1:1 ratio and transferred into WT mice. SM1 CD4 T cell frequency in the blood and tissues was determined in a similar manner as the left panels. Data are from two experiments with 11 mice per group. B, CD44 expression on SM1 CD4 T cells. Representative histograms are gated on CD4⁺Thy1.1⁺ cells from blood samples obtained on day 0 (dotted line) and day 5 (filled histograms or solid line) following immunization. Bar graph shows mean fluorescent intensity (MFI) of CD44 staining on day 0 (striped bars) vs day 5 (solid). Data are derived from experiments performed in A and are matched by color of the symbol to the bar. CD44 up-regulation was statistically significant with p < 0.005. C, IFN-γ production by SM1 T cells following 5 h in vitro restimulation with FL-pep using day 10 splenocytes. Representative histograms are gated on Thy1.1⁺ cells. Bar graphs display the percent of SM1 T cells expressing IFN-γ (left) and total number of IFN-γ⁺ SM1 T cells (right). The bracket between the closed and open bars labeled 4.6-fold represents the difference between WT and MyD88^−/− hosts. Data are derived from experiments performed in A, with data matched by color. All data are shown as mean ± SEM. Asterisks (∗) next to open symbols represent statistical significance compared with closed symbols with p = 0.02–6 × 10⁻⁴.

In addition to signaling through TLRs, MyD88 is involved in signal transduction through receptors for IL-1, IL-18, IL-33, and IFN-γ (20, 22, 23). Because T cells may express receptors for these cytokines, it was possible that MyD88 expression by T cells was also important for LPS adjuvanticity. To test this notion, SM1 mice were crossed to MyD88^−/− mice. MyD88^−/− SM1 T cells were mixed with WT SM1 T cells at 1:1 and transferred into WT hosts, which were immunized with FL-pep and LPS. The SM1 T cell populations were independently tracked by their pattern of Thy1 expression. Both WT and MyD88^−/− T cells expressed Thy1.1 while the WT SM1 T cells also expressed Thy1.2 in the first experiment, allowing us to track them independently from MyD88^−/− SM1 T cells. In the second experiment, it was the MyD88^−/− SM1 T cells that expressed Thy1.2, and expression of Thy1.2 by SM1 T cells did not affect the results. The MyD88-deficient T cells underwent normal clonal expansion and long-term survival following immunization (Fig. 2⇑A, right panels), demonstrating that MyD88 expression by Ag-specific T cells was not required for clonal expansion nor survival.

Thus, the effect of reduced T cell accumulation in MyD88^−/− mice was wholly derived from non-T cells. One possible explanation for these results was that MyD88^−/− APCs were unable to transmit appropriate signals for T cell activation (43). We found that T cells up-regulated CD44 normally in MyD88^−/− mice in response to FL-pep and LPS (Fig. 2⇑B), demonstrating their activation was MyD88 independent. An additional experiment revealed LPS caused peptide-stimulated SM1 T cells to down-regulate CD62L in a MyD88-dependent manner, which is interesting considering the reduced T cell accumulation in MyD88^−/− mice (data not shown). Nevertheless, a more stringent test for T cell activation is the capacity to undergo effector differentiation as measured by IFN-γ synthesis. We show direct evidence that T cells primed in MyD88^−/− mice produced IFN-γ when restimulated in vitro on day 10 (Fig. 2⇑C). Approximately 50% of SM1 CD4 T cells primed in MyD88^−/− mice produced IFN-γ following in vitro restimulation compared with 30% of SM1 CD4 T cells recovered from WT mice. Optimal IFN-γ production on a per cell basis required LPS, as immunization with FL-pep plus LPS increased the frequency of SM1 T cells producing IFN-γ by over 3-fold on average for individual experiments when compared with immunization with FL-pep alone (data not shown). This, coupled with the effect of LPS on T cell accumulation, caused this adjuvant to have quite a dramatic effect on total numbers of effector cells. It was interesting that the percentage of T cells producing IFN-γ was highest in MyD88^−/− mice; however, the total number of T cells capable of producing IFN-γ was much lower in MyD88^−/− mice because their accumulation was significantly reduced. Lastly, we show that IFN-γ production did not require expression of MyD88 in SM1 CD4 T cells (Fig. 2⇑C, gray). Taken together, these data showed that MyD88 expression by non-T cells was central for T cell accumulation but not for their activation or effector differentiation.

Although MyD88^−/− APCs supported SM1 CD4 T cell effector differentiation, the reduced T cell accumulation in MyD88^−/− mice may suggest their possible involvement in proliferation. To address this, we performed in vitro proliferation assays from three experiments by culturing splenocytes with Ag for 72 h and measuring thymidine incorporation during the last 8 h. The first experiment involved mice that were immunized with SEA plus LPS whereas the second experiment involved mice that received SM1 T cells and were immunized with peptide plus LPS. The third experiment examined proliferation of naive SM1 T cells. In all cases, MyD88^−/− APCs were able to support Ag-specific T cell proliferation as measured by thymidine incorporation. When the data were corrected for the numbers of Ag-specific T cells, MyD88^−/− APCs were equivalent to WT APCs at supporting proliferation (data not shown). Therefore, we suggest the reduced T cell accumulation in MyD88^−/− mice is not due to an inability of T cells to undergo proliferation, but rather their inability to survive following proliferation. Nevertheless, in vitro proliferation data may differ from in vivo results, suggesting that in vivo studies such as CFSE dilution or BrdU labeling will be necessary for a complete understanding of this important issue.

MyD88 is required for LPS to generate long-term T cell survival

Because MyD88 affected SM1 CD4 T cell clonal expansion (Fig. 2⇑A), we were unable to determine its role in generating long-term survival of this population. To circumvent this issue, we attempted to restore clonal expansion in MyD88^−/− mice by using an anti-OX40 agonist mAb. Activated CD4 T cells express OX40 (44), and the combination of anti-OX40 mAb plus LPS has a synergistic effect on T cell survival (45). Because SM1 CD4 T cells primed in MyD88^−/− mice became activated (Fig. 2⇑, B and C), we reasoned they should be responsive to the anti-OX40 mAb. If MyD88 was only involved in clonal expansion, the anti-OX40 mAb should restore SM1 T cell accumulation in MyD88^−/− mice. If MyD88 is required for survival, the anti-OX40 mAb should not be able to induce T cell survival following a boosted clonal expansion.

WT SM1 CD4 T cells were transferred into MyD88^−/− mice, followed by immunization with FL-pep and either anti-OX40 mAb alone, LPS, or anti-OX40 mAb with LPS, or control mice given rat IgG Ab in place of anti-OX40 mAb. WT mice were immunized similarly for comparison; however, because WT mice are far more sensitive to LPS, their dose was ∼2.5 times lower than that of MyD88^−/− mice to equalize responsiveness.

Anti-OX40 mAb alone or rat IgG with LPS treatment caused expected levels of SM1 CD4 T cell clonal expansion and contraction in WT mice (Fig. 3⇓A, top). Immunization with both anti-OX40 mAb and LPS had a profound synergistic effect on clonal expansion (day 5) and survival (day 14), consistent with a previous study from our laboratory (45). Notably, the SM1 CD4 T cell population of this group did not undergo contraction, resulting in 10-fold higher frequencies in the blood on day 14 compared with mice immunized with rat IgG and LPS. This was associated with higher SM1 CD4 T cell levels in the tissues on day 10 or 14 (Fig. 3⇓B). Immunization with anti-OX40 mAb and LPS enhanced SM1 CD4 T cell frequencies in WT mice by 5.4- and 4.4-fold in the spleen and liver, respectively, while total SM1 CD4 T cell numbers were enhanced 6- to 7-fold on average.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Repairing clonal expansion did not rescue T cell survival in MyD88^−/− mice. WT SM1 CD4 T cells were transferred into WT and MyD88^−/− mice, followed by immunization with FL-pep and either anti-OX40 mAb (white circles), rat IgG plus LPS (gray circles), or anti-OX40 mAb plus LPS (black circles). A, SM1 T cell frequencies in blood at various times. Data were pooled from three experiments with a total of 8–12 mice per group. Asterisks (∗) represent statistically significant differences between anti-OX40 plus LPS and rat IgG plus LPS groups with p = 0.008–8 × 10⁻⁵. B, Left, SM1 CD4 T cell frequencies from individual mice in spleen and liver on day 10 (squares) or 14 (circles). B, Right, Total SM1 CD4 T cell numbers from individual mice on day 10 (squares) or 14 (circles). Data are from four experiments and bars indicate the mean. Asterisks (∗) represent statistically significant differences between rat IgG plus LPS and “Both” groups with p < 0.02. C, IFN-γ and TNF synthesis following 5 h in vitro restimulation of day 14 splenocytes with FL-pep. Representative dot plots (left) are gated from Thy1.1⁺ cells and analyzed for IFN-γ and TNF. Scatter plot (right) shows total numbers of splenic double-positive SM1 T cells for individual mice analyzed on days 10 (squares) or 14 (circles). Individual mice that did not generate enough total SM1 T cells to allow for analysis of cytokine production are represented as below the limit of detection. Data are from experiments shown in B. Bars for each group represent the mean for individual samples above the limit of detection. For WT recipients, the difference between rat IgG plus LPS and “Both” groups is statistically significant (p < 0.005) while for MyD88^−/− recipients the same comparison is not significant (p = 0.14).

In MyD88^−/− mice, anti-OX40 mAb alone induced relatively normal levels of clonal expansion and immunization with rat IgG and LPS elicited a similar response as anti-OX40 mAb alone (Fig. 3⇑A, bottom). Importantly, however, immunization with anti-OX40 mAb and LPS substantially boosted clonal expansion by 7-fold on day 5. In fact, clonal expansion in this group was comparable to WT mice, although slightly lower, suggesting a minor MyD88-dependent effect on clonal expansion with this immunization protocol. However, after day 5, the SM1 CD4 T cell population in MyD88^−/− mice underwent rapid and massive deletion. Thus, in comparison to the response in WT mice where no deletion was observed between days 5 and 14, the SM1 CD4 T cell population in MyD88^−/− mice immunized with anti-OX40 mAb plus LPS decreased by 9-fold between days 5 and 14 (Fig. 3⇑A). On day 10 or 14, SM1 CD4 T cell numbers in MyD88^−/− mice immunized with anti-OX40 mAb and LPS were reduced by 3- and 15-fold in the spleen and liver, respectively, when compared with similarly immunized WT mice (Fig. 3⇑B). Notably, only 3 of the 10 MyD88^−/− mice analyzed generated significant SM1 CD4 T cell numbers, and on day 21, OX40 stimulation had no effect on SM1 CD4 T cell numbers in the MyD88^−/− mice that were immunized with FL-pep and LPS (unpublished data). Therefore, even though exogenous OX40 costimulation restored clonal expansion in MyD88^−/− mice, it did not substitute for MyD88-dependent factors in promoting T cell survival through the action of LPS.

To determine whether OX40-stimulated T cells in MyD88^−/− mice were functional, splenocytes were restimulated in vitro with FL-pep. Consistent with Fig. 2⇑C, the percent of SM1 CD4 T cells producing both IFN-γ and TNF (double producers) following immunization with rat IgG and LPS was highest when they were primed in MyD88^−/− mice (Fig. 3⇑C). Nevertheless, maximal effector potential occurred following immunization with anti-OX40 mAb and LPS. Some mice did not generate enough SM1 T cell numbers for us to analyze cytokine production. In particular, following immunization with FL-pep and anti-OX40 mAb, 5 of 8 WT mice and 2 of 5 MyD88^−/− mice did not produce effector cell numbers above a reasonable limit of detection (Fig. 3⇑C, right). For MyD88^−/− mice treated with LPS, 7 of 20 mice did not produce effector cell numbers above our limit of detection, similar to no LPS. When comparing total numbers of splenic double producers in WT mice, the group immunized with anti-OX40 mAb and LPS had 16-fold more effector cells on average than the group immunized with rat IgG and LPS (Fig. 3⇑C, right). In the MyD88^−/− mice that generated detectable effector cell numbers, treatment with anti-OX40 mAb and LPS produced only 6.8 times more double producers than treatment with rat IgG and LPS. Furthermore, by day 21, MyD88^−/− mice that were treated with anti-OX40 mAb and LPS did not have higher effector cell numbers than the ones treated with rat IgG and LPS (unpublished data). Therefore, although effector differentiation was clearly MyD88 independent, LPS signaling through the MyD88-dependent pathway was required for the accumulation of effector SM1 CD4 T cells.

LPS enhances T cell accumulation independently of CD11c⁺ cells, however, CD11c⁺ cells are required for effector T cell differentiation

Thus far, we showed that LPS enhanced T cell survival and effector function through MyD88-dependent and -independent mechanisms, respectively. Exogenous OX40 stimulation was able to substitute for LPS in generating T cell effector function, suggesting that effector function may be controlled by costimulatory molecules, such as OX40 ligand, expressed on APCs. One intriguing possibility is that different types of innate immune cells are specialized for generating T cell survival vs effector differentiation.

To test this possibility, we used mice that express DTR under control of the CD11c promoter (27). Injection of DT into these mice induces transient depletion of CD11c⁺ cells, including DCs, metallophilic macrophages, marginal zone macrophages, and any other cell expressing CD11c (33). Because a single DT injection was reported to be lethal within 7 days (46), we generated bone marrow chimeras to restrict expression of the transgene to the hemopoietic compartment. This technique allows DTR-expressing mice to survive multiple injections of DT (28, 46). WT mice were irradiated and reconstituted with bone marrow from DTR mice, and for controls WT mice were reconstituted with WT bone marrow. During these experiments, all bone marrow chimeras, including WT controls, were injected with DT at 48- to 72-h intervals to maintain depletion of CD11c⁺ cells, with the first injection given 3–4 days before immunization. On the day of sacrifice, the percentage of splenic cells expressing CD11c was reduced by 5- to 10-fold in DTR chimeras (Fig. 4⇓A), indicating that DT treatment was efficacious.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

LPS enhanced T cell accumulation independently of CD11c⁺ cells, but CD11c⁺ cells were required for optimal IFN-γ production by T cells. SM1 cells were transferred into WT (left panel) and DTR (right panel) chimeras, followed by immunization with FL-pep and LPS. A, CD11c expression on splenocytes from day 10, gated on live cells. B, SM1 CD4 T cell frequencies in blood of bone marrow chimeras immunized with FL-pep alone (open symbols) and FL-pep plus LPS (closed symbols). Data were pooled from six experiments with a total of 17–26 mice per group and shown are the mean ± SEM. C, Total SM1 T cell numbers in spleens and livers on day 10 are given as mean ± SEM. D, IFN-γ production by SM1 CD4 T cells on day 5 following restimulation of liver lymphocytes and splenocytes with FL-pep. Representative histograms are gated on Thy1.1⁺ cells and scatter plots show the percentage of SM1 T cells expressing IFN-γ from individual mice. Data were pooled from five experiments and the horizontal bar represents the mean. Asterisks (∗) represent significant statistical differences with p = 0.02–3 × 10⁻⁹.

WT SM1 CD4 T cells were transferred into WT or DTR bone marrow chimeras, followed by immunization with FL-pep and LPS. In response to FL-pep alone, SM1 T cell clonal expansion was minimal in both chimeras (Fig. 4⇑B), despite normal up-regulation of CD44 (unpublished data). LPS treatment with FL-pep increased SM1 T cell frequencies to a similar extent in WT (7.7-fold) and DTR (5.6-fold) chimeras (Fig. 4⇑B). This effect was also clearly observed in tissues on day 5 (Table II⇓). LPS treatment in DTR chimeras enhanced absolute numbers of specific SM1 CD4 T cells in lymphoid (PLN, mesenteric lymph nodes, spleen) and nonlymphoid (liver, lung) tissues by an amount that was equivalent to, or better than, WT chimeras. Second, 10 days after immunization LPS treatment potently increased splenic SM1 CD4 T cell numbers in both WT and DTR chimeras (Fig. 4⇑C). We noticed the overall cell recovery from the DTR chimeras was low, possibly reflecting changes in the organization of lymphoid structure as reported earlier (33). On day 10, the effect of LPS on liver SM1 T cells was greater for WT chimeras because the DTR chimeras immunized with FL-pep alone had unusually high SM1 T cell frequencies, indicating that CD11c⁺ cells may participate in peripheral T cell deletion, as seen in FL-pep alone responses (38). Overall, this demonstrated that LPS enhanced SM1 T cell clonal expansion and survival independently of CD11c⁺ cells when measured by either frequency or numbers, and analyzed in lymphoid and nonlymphoid tissues.

To examine T cell effector potential, day 5 cells from the liver and spleen were restimulated in vitro with FL-pep for 5 h and stained for intracellular IFN-γ. The percentage of SM1 T cells producing IFN-γ was ∼50% lower for DTR chimeras (Fig. 4⇑D), indicating that CD11c⁺ cells were involved in generating optimal T cell effector differentiation. Because the generation of T cell survival by LPS involved MyD88 while effector differentiation did not (Fig. 2⇑), this suggested that CD11c⁺ cells preferentially contributed to the MyD88-independent aspect of the T cell response.

To determine whether the involvement of CD11c⁺ cells in T cell effector differentiation was applicable to a mixed endogenous T cell population, WT and DTR chimeras were immunized with SEA followed by LPS 18 h later. The T cell response to SEA alone was normal in the DTR chimeras (Fig. 5⇓A). Second, by day 10, LPS treatment in WT chimeras had increased the percentage of CD4 T cells expressing Vβ3 by 5-fold, and in accordance with Table II⇑, LPS enhanced CD4 Vβ3 T cell survival in the DTR chimeras by an amount equivalent to WT. In contrast, survival of CD8⁺ cells expressing Vβ3 was reduced in DTR chimeras by 56 and 33% in the spleen and PLN, respectively (data not shown). We speculate, however, that this was due to activation of the CD11c promoter in Ag-stimulated CD8⁺ T cells (47), followed by their expression of DTR and eventual depletion by DT. Overall, LPS enhanced endogenous CD4 T cell survival independently of CD11c⁺ cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Enhancement of endogenous T cell survival by LPS was independent of CD11c⁺ cells, but rescued T cells displayed irregular cytokine synthesis upon recall. WT and DTR bone marrow chimeras were immunized with SEA and LPS as indicated in the Materials and Methods. A, Day 10 percentages of CD4 T cells expressing Vβ3 in spleens and PLN of chimeras immunized with SEA alone (□) and SEA plus LPS (▪). B, One million splenocytes were restimulated in vitro on day 10 with SEA overnight. The concentration of IFN-γ (left) and IL-2 (right) in culture supernatants was determined by ELISA. ▦, Data gathered from restimulation of cells from WT chimeras; ▨, data gathered from DTR chimeras. The labeled brackets between the ▦ and ▨ represent the fold difference between WT and DTR bone marrow chimeras. Data were pooled from six experiments with a total of 5–12 mice per group and the mean ± SEM is given. Asterisks (∗) represent significant statistical differences with p = 0.014–1.3 × 10⁻⁴.

To examine cytokine potential, day 10 splenocytes were restimulated in vitro with SEA for 24 h, and culture supernatants were analyzed for the presence of IFN-γ and IL-2 by ELISA. Following immunization with SEA alone, the level of IFN-γ was 4-fold lower in cultures derived from DTR bone marrow chimeras (Fig. 5⇑B, left). Immunization with LPS dramatically increased IFN-γ levels for WT chimeras but only marginally affected cells from the DTR chimeras. Ultimately, for mice immunized with SEA and LPS, the level of IFN-γ was 6.4-fold lower in the DTR cultures (Fig. 5⇑B). IL-2 production followed a different trend. For mice immunized with SEA alone, the level of IL-2 was 2.3-fold lower in DTR cultures (Fig. 5⇑B, right), but SEA and LPS immunization restored it to a level comparable with WT. This was the case even though DTR cultures contained fewer Vβ3 T cells (data not shown). Overall, restimulation of cultures from DTR bone marrow chimeras immunized with SEA and LPS resulted in normal IL-2 production but much lower IFN-γ production, suggesting that T cells primed in the absence of CD11c⁺ cells may have a selective defect in the ability to produce IFN-γ.

Priming in the absence of CD11c⁺ cells causes an intrinsic and selective defect in T cell IFN-γ potential

A smaller portion of T cells that were primed in DTR chimeras were able to produce IFN-γ (Fig. 4⇑C), indicating a possible intrinsic defect. Alternatively, differences of non-T cell populations between the WT and DTR cultures could have affected these results. For example, DTR cultures contained 5- to 10-fold fewer CD11c⁺ cells and it was possible that CD11c⁺ cells were functionally superior APCs for stimulating IFN-γ in this assay, but perhaps not IL-2.

To address this, SM1 CD4 T cells were transferred into WT and DTR bone marrow chimeras, followed by immunization with FL-pep and LPS. On day 5, SM1 T cells were purified from livers, a locale for effector T cell residence (37), and cultured for 5 h with CD11c⁺ cells purified from spleens of naive WT mice. Cells were stimulated with either peptide, PMA and ionomycin, or nothing. Production of IFN-γ and TNF by SM1 T cells was assessed by flow cytometry. SM1 T cells from cultures that did not contain peptide or CD11c⁺ cells did not produce significant levels of cytokines (Fig. 6⇓A, left column). Following restimulation, a large proportion of SM1 T cells that were primed in WT bone marrow chimeras produced TNF and IFN-γ. Two main populations were observed: one produced TNF only and another produced both TNF and IFN-γ (double producer). When averaged from five independent experiments, 37 and 44% of SM1 T cells primed in WT bone marrow chimeras were double producers when restimulated with peptide or PMA plus ionomycin, respectively (Fig. 6⇓A). In contrast, the corresponding double producer population was much smaller for T cells primed in DTR chimeras: 18 and 21% on average following restimulation with peptide or PMA plus ionomycin, respectively. The total percentage of IFN-γ⁺ SM1 T cells was reduced by ≥2-fold when they were primed in DTR chimeras. Importantly, the ratio of double producers to TNF-only producers was ∼3-fold lower for SM1 T cells primed in DTR chimeras (Fig. 6⇓B). This was associated with a higher percentage that did not produce either TNF or IFN-γ (Fig. 6⇓C). Nonetheless, SM1 T cells primed in DTR bone marrow chimeras did not seem to have a global defect in function because the population producing TNF alone was unaffected by the absence of CD11c⁺ cells (Fig. 6⇓A, see upper left quadrants). Therefore, T cells primed in the absence of CD11c⁺ cells had an intrinsic and selective defect in IFN-γ production. This defect was distal to the TCR because it was not rescued with PMA plus ionomycin stimulation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

T cells primed in the absence of CD11c⁺ cells possessed an intrinsic and selective defect in IFN-γ synthesis potential. SM1 CD4 T cells were transferred into WT and DTR bone marrow chimeras, followed by immunization with FL-pep and LPS as indicated in Materials and Methods. On day 5, SM1 T cells were purified from livers and cultured for 5 h with CD11c⁺ cells purified from spleens of naive WT mice. Cultures were stimulated with either nothing, FL-pep, or PMA plus ionomycin. A, Dot plots from five independent experiments were gated on Thy1.1⁺ cells to reveal IFN-γ and TNF levels. The asterisk (∗) in the “no peptide” column for experiment 5 indicates Thy1.1⁺ cells were cultured with peptide in the absence of added CD11c⁺ cells. B, The ratio of IFN-γ⁺ TNF⁺ (double producers) cells to IFN-γ⁻TNF⁺ (TNF alone) cells was calculated for each experiment, with mean ± SEM displayed. C, Percentages of IFN-γ⁻TNF⁻ (nonproducers) SM1 T cells were calculated as mean ± SEM.

Enforced OX40 stimulation elicits Th1 effector differentiation independently of CD11c⁺ cells

CD11c⁺ cells may be specifically required for LPS to induce CD4 T cell effector differentiation, or they may be required for a variety of stimuli to elicit Th1 responses. In Fig. 3⇑, we showed that immunization with an agonist anti-OX40 mAb caused IFN-γ production by specific SM1 CD4 T cells just as well as LPS on a per cell basis. We wanted to determine whether the ability of the anti-OX40 mAb to enhance effector differentiation also depends on the presence of CD11c⁺ cells.

To test this, SM1 CD4 T cells were transferred into WT and DTR bone marrow chimeras, followed by immunization with FL-pep and either anti-OX40 mAb or rat IgG as a control. On day 5, splenocytes and liver lymphocytes were recalled with FL-pep in vitro and stained for production of intracellular IFN-γ and TNF. Immunization with FL-pep and rat IgG resulted in ∼10% of splenic SM1 T cells having the capacity to produce IFN-γ in both the WT and DTR bone marrow chimeras, while 30–40% produced TNF (Fig. 7⇓A). Immunization with FL-pep and anti-OX40 mAb increased the amount of splenic SM1 T cells producing IFN-γ to 31% in cells from WT chimeras and 42% in cells from DTR chimeras. Similar results were also observed for the liver (Fig. 7⇓B), demonstrating that CD4 T cells can undergo effector differentiation in DTR bone marrow chimeras when they receive appropriate signals. Therefore, the lack of effector differentiation in DTR bone marrow chimeras immunized with Ag and LPS (Figs. 4–6⇑⇑⇑) was likely due to a specialized function for CD11c⁺ cells to receive signals from LPS in vivo, to interpret these signals, and ultimately to translate them to CD4 T cells with the effect of instructing effector differentiation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Enforced OX40 stimulation drives Th1 effector differentiation independently of CD11c⁺ cells. SM1 CD4 T cells were transferred into WT or DTR bone marrow chimeras, followed by immunization with FL-pep and either anti-OX40 mAb or rat IgG control, as indicated. On day 5, lymphocytes from the spleen (A) and liver (B) were recalled in vitro with FL-pep for 5 h, and stained for intracellular IFN-γ and TNF. Representative dot plots are gated on Thy1.1⁺ cells. Bar graphs show the percent of double-positive SM1 T cells as the mean ± SEM of five individual experiments with a total of four to nine mice per group. There were no significant statistical differences between similarly treated WT and DTR bone marrow chimeras.