Invariant NKT Cells Amplify the Innate Immune Response to Lipopolysaccharide

Mice

C57BL/6J, BALB/cJ, B6.129S1-Il12b^tm1Jm/J (IL-12p40^−/−), and B6.129P2-Il18^tm1Aki/J (IL-18^−/−) mice were obtained from The Jackson Laboratory and maintained in the vivarium at the La Jolla Institute for Allergy and Immunology (La Jolla, CA). CD1d^−/− mice on the C57BL/6J background were a gift from Dr. L. Van Kaer (Vanderbilt University, Nashville, TN). B6.129-Tcra-J^tm1Tgi (Jα18^−/−) mice on the C57BL/6J background were a gift from Dr. M. Taniguchi (RIKEN Institute, Yokohama, Japan). Vα14-Jα18 transgenic mice on the C57BL/6J.Cα^−/− background were a gift from Dr. A. Bendelac (University of Chicago, Chicago, IL). Unless otherwise mentioned, all mice were used between 8 and 16 wk of age, and were age and gender matched.

Reagents and Abs

Escherichia coli O111:B4 LPS and Salmonella abortus equi LPS were purchased from Sigma-Aldrich, either purified by phenol extraction (used for injections) or phenol extracted and further purified by ion-exchange chromatography (used for cell culture).

Abs to the following mouse Ags were purchased from BD Biosciences, as conjugates to FITC, Alexa 488, PE, PerCP-Cy5.5, PE-Cy7, allophycocyanin, or biotin: CD1d, CD4, CD8α, CD11b, CD11c, CD18, CD69, Gr-1, NK1.1, TCRβ, IFN-γ, and TNF. Allophycocyanin-conjugated anti-F4/80 was purchased from eBioscience. Anti-mouse IL-4-Alexa 488 was obtained from Caltag Laboratories. Anti-mouse macrosialin (CD68) was purchased from Serotec. Anti-mouse CD16/32 used for FcR blocking was isolated in our laboratory. Purified anti-mouse CD1d (1B1) for blocking experiments was obtained from BD Biosciences. PE-conjugated αGalCer-CD1d tetramers were synthesized in our laboratory as described (17), and used to stain cell suspensions. Recombinant mouse GM-GSF (rmGM-CSF) was purchased from PeproTech. Recombinant mouse IL-12 and IL-18 (rmIL-12 and rmIL-18) were purchased from R&D Systems. α-Galactosylceramide (αGalCer; (2S,3S,4R)-1-O-(α-d-galactopyranosyl)-N-hexacosanoyl-2-amino-1,3,4-octadecanetriol) was synthesized at Kirin Pharmaceuticals (18, 19).

Treatments

All treatment protocols were conducted according to Institutional animal care guidelines. Mice were injected i.v. in the tail vein in a final volume of 200 μl, and blood was harvested by cardiac puncture after sacrifice. Mice were injected with 40 μg of LPS diluted in PBS, and 2 μg of αGalCer in vehicle, unless otherwise mentioned.

Cell preparation

All cell preparations were conducted using complete RPMI (RPMI 1640; Invitrogen Life Technologies) supplemented with 10% FCS, penicillin-streptomycin-glutamine, and 2-ME). Livers were perfused with ice-cold PBS before isolation of myeloid cells. Livers were mashed through a 70-mm cell strainer (Fisher Scientific) and washed. The suspension was centrifuged over a 40–70% Percoll (Amersham Biosciences) gradient and mononuclear cells were isolated by harvesting the interphase. Liver mononuclear cells were resuspended in the appropriate amount of medium and stained before FACS or sorting. Spleens were mashed and RBC were lysed using red cell lysing buffer from Sigma-Aldrich, after which the cells were resuspended in complete RPMI 1640. The cell suspension was stained and used for FACS, sorting, or for cell culture.

Staining for flow cytometry and intracellular cytokine staining

Cells were resuspended in staining buffer (PBS with 10% FCS) and anti-CD16/32 was added to block nonspecific staining. Abs to surface markers were added and cells were stained for 20 min on ice. Cells were washed and resuspended in staining buffer and analyzed on a FACSCalibur (BD Biosciences).

Cell suspensions obtained from treated mice were stained for surface markers as described above, then fixed and permeabilized using CytoFix/CytoPerm buffer from BD Biosciences, according to the manufacturer’s protocol. Abs against intracellular cytokines or CD68 were added and the cells were stained on ice for 30 min. Cells were washed and resuspended in staining buffer and analyzed on a FACSCalibur.

FACS sorting and cell purification

Vα14i NKT cells were sorted from livers or spleens, using labeling with αGalCer-CD1d tetramers and anti-TCRβ, and NKT cells were sorted with anti-NK1.1 and either anti-TCRβ or anti-CD5. Prior to sorting with CD1d-αGalCer tetramers, spleen cells were depleted of B cells and CD8 T cells using anti-CD19 and anti-CD8α conjugated magnetic beads and MACS columns (Miltenyi Biotec), according to the manufacturer’s protocol. The CD19⁻CD8⁻ cell suspension was labeled with αGalCer-CD1d tetramers and anti-TCRβ mAb, washed, and sorted. For sorting NK1.1⁺TCRβ⁺ or NK1.1⁺CD5⁺ cells, spleen suspensions were labeled with anti-NK1.1-PE and anti-TCRβ or anti-CD5 Abs and NK1.1⁺ cells were enriched using anti-PE-conjugated magnetic MACS beads (Miltenyi Biotec), washed, and sorted. CD8α⁺ and CD11b⁺ cells were rigorously excluded while sorting for NK1.1⁺TCRβ⁺ or NK1.1⁺CD5⁺ cells. Cells were routinely sorted to >97% purity on a FACSDiva (BD Biosciences). For the enrichment of Vα14i NKT cells from Vα14-transgenic (Tg) mice, spleen cells from these mice were first depleted of CD8α⁺ and CD19⁺ cells using MACS beads as described above. The depleted fraction was then further depleted of residual CD8α⁺, CD19⁺, B220⁺, CD11c⁺, CD11b⁺, and TCRγδ⁺ cells by labeling with biotinylated Abs to the listed Ags (eBioscience) followed by addition of anti-biotin conjugated MACS beads. The percentage of Vα14i NKT cells in the enriched fraction was determined and the equivalent of 5 × 10⁴ Vα14i NKT cells was cultured with bone marrow (BM) dendritic cells (DCs) as described below.

BM-derived DCs and cell culture

BM DCs were prepared as described (20). Briefly, BM was harvested from femurs of mice and cultured in 100-mm bacterial dishes in 10 ml of complete RPMI with 200 U/ml rmGM-CSF. On day 3 of the culture, 10 ml more of complete RPMI with 200 U/ml rmGM-CSF were added. BM DCs were harvested on day 6 and cultured with sorted NKT cells, with the addition of LPS or recombinant cytokines. A total of 2.5 × 10⁵ BM DCs were cultured with 5 × 10⁴ sorted NKT cells in a round-bottom 96-well plate at 37°C and 5% CO2 for 48 h, as described (4). The culture supernatants were harvested, and IFN-γ and IL-4 levels measured by ELISA.

Cytokine detection

Abs and recombinant cytokine standards for IFN-γ, IL-4, and IL-12 ELISAs were purchased from BD Biosciences. IL-18 and TNF ELISA kits were purchased from R&D Systems. All ELISAs were conducted according to the manufacturer’s instructions. Cytokine estimation by cytometric bead array was conducted using the BioPlex system from Bio-Rad and results were read using a Luminex plate reader.

RNA preparation and real-time PCR

Cells were resuspended in TRIzol (Invitrogen Life Technologies) for RNA isolation. Proteins and DNA removed by precipitation with chloroform and RNA was extracted by precipitation with isopropanol. The resulting total RNA was used for cDNA synthesis by the Superscript II system (Invitrogen Life Technologies). cDNA was used for real-time PCR using SYBR green (Bio-Rad), with the following primers: IFN-γ forward, 5′-GGATGCATTCATGAGTATTGC, reverse, 5′-CCTTTTCCGCTTCCTGAGG; TNF forward, 5′-AACTAGTGGTGCCAGCCGAT, reverse, 5′-CTTCACAGAGCAATGACTCC. Gene expression was normalized to L32, forward, 5′-GAAACTGGCGGAAACCCA and reverse, 5′-GGATCTGGCCCTTGAACCTT. Data was collected and analyzed on an ABI GeneAmp5700 (Applied Biosystems).

Statistical analysis

Statistical analysis was conducted using the GraphPad Prism software package from GraphPad Software.

Vα14i NKT cells produce exclusively IFN-γ after stimulation with LPS

It has been shown that Vα14i NKT cells produce IFN-γ after exposure to LPS (11, 16). To determine whether they produce other cytokines, mice were injected i.v. with 40 μg of E. coli LPS and cytokine production by Vα14i NKT cells was assayed by intracellular cytokine staining (ICCS) of gated αGalCer-CD1d tetramer⁺ cells, analyzed directly ex vivo, without further stimulation or addition of brefeldin A. Vα14i NKT cells showed increased expression of the early activation Ag CD69 in response to LPS as early as 2 h after injections (data not shown). However, production of IFN-γ by Vα14i NKT cells from the liver and the spleen was detected by ICCS only 3 h after LPS injection (data not shown) and peaked at 6 h (Fig. 1⇓A), with ∼30% of the cells positive. IFN-γ production did not correlate with CD4 expression or any other cell surface marker tested (data not shown). At this time point after LPS injection, Vα14i NKT cells comprised approximately half of the total IFN-γ-producing cells in the liver and one-fifth of those in the spleen (Fig. 1⇓B). The other IFN-γ-producing cells at this time were mostly NK cells (data not shown). Strikingly, Vα14i NKT cells did not produce detectable IL-4 (Fig. 1⇓C), in contrast to the outcome following stimulation with αGalCer (Fig. 1⇓C). Vα14i NKT cells also did not produce detectable TNF between 1 and 6 h after LPS injection, although this cytokine also is induced by αGalCer-mediated TCR stimulation (data not shown). These data show that Vα14i NKT cells produce a more limited set of cytokines after LPS stimulation compared with GSL Ag stimulation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Vα14i NKT cells produce IFN-γ but no IL-4 in response to LPS. A and B, Vα14i NKT cells in both the livers and spleens of C57BL/6J mice produced IFN-γ as measured by ICCS 6 h after LPS injection. A, IFN-γ production specifically by Vα14i NKT cells. The numbers indicate the percentage of gated Vα14i NKT cells producing IFN-γ. Top panels, The gating for Vα14i NKT cells (arrows). B, IFN-γ production compared between Vα14i NKT cells and other cells in the spleen and liver. Numbers indicate percentage of total live cells in the quadrant. Data shown are from one mouse of a group of two, each experiment was repeated at least three times. C, Vα14i NKT cells do not produce IL-4 as measured by ICCS 6 h after LPS injection, but do produce IL-4 1.5 h after αGalCer injection. Uninjected (shaded histograms) and LPS injected (dotted lines) are essentially superimposable. One representative experiment of at least three.

Vα14i NKT cells do not respond directly to LPS

To establish the mechanism by which Vα14i NKT cells responded to LPS, αGalCer-CD1d tetramer⁺ cells were stained with Abs to the TLR4-MD2 complex to determine whether these cells could directly recognize LPS. We did not detect the presence of this complex on the surface of Vα14i NKT cells (data not shown). Furthermore, real-time PCR performed on Vα14i NKT cells FACS sorted from spleen (data not shown) did not detect TLR4 mRNA.

Studies with Salmonella LPS demonstrated that Vα14i NKT cells require IL-12 production by APCs to respond to LPS (16). In accordance with the published results, Vα14i NKT cells from IL-12p40^−/− mice were impaired in their response to LPS, with a 5-fold reduction in the percentage of cells that produced IFN-γ, measured by ICCS (Fig. 2⇓A). IL-12p40^−/− mice are deficient for the p40 subunit of IL-12, which is shared by IL-23, but the earlier study using blocking IL-12 Abs (16) would suggest that it is IL-12 that is required.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

IFN-γ production by Vα14i NKT cells in response to LPS depends on IL-12 and IL-18. Mice deficient in the p40 subunit of IL-12 (A) or IL-18 were injected with LPS (B) and IFN-γ production by spleen Vα14i NKT cells was assayed using ICCS 6 h later. Panels on the left show the αGalCer-CD1d tetramer⁺ population that was gated (arrows). Data shown are from one mouse of a group of two, each experiment was repeated three times.

Because the absence of IL-12p40 did not completely eliminate IFN-γ production by Vα14i NKT cells after LPS injection, we examined the effect of IL-18 deficiency, since earlier studies have shown that Vα14i NKT cells produce cytokines when exposed to IL-18 in vitro (21). Also, IL-18 synthesis is strongly induced by LPS, and the combination of IL-12 and IL-18 is known to have potent stimulatory effects on IFN-γ production (22, 23). We therefore measured IFN-γ production by Vα14i NKT cells in response to LPS in IL-18^−/− mice (Fig. 2⇑B). Vα14i NKT cells from IL-18^−/− mice had an even more pronounced impairment in IFN-γ production after LPS (Fig. 2⇑B). No enhancement in IL-4 production in response to LPS by Vα14i NKT cells from IL-12p40^−/− and IL-18^−/− mice was observed (data not shown). This defect in IFN-γ production by Vα14i NKT cells in IL-12p40^−/− and IL-18^−/− mice was not due to an intrinsic defect in the capacity of Vα14i NKT cells to produce cytokines, as they responded normally to αGalCer injection (data not shown).

Because type I IFNs are involved in the production of IFN-γ by memory CD8⁺ T cells in response to LPS (24), the role of type I IFNs in the Vα14i NKT cell response to LPS was tested using mice that are unresponsive to type I IFNs (IFNAR1^−/−). Percentages of IFN-γ-producing Vα14i NKT cells observed by ICCS 6 h after injection in IFNAR1^−/− mice were slightly reduced but still comparable to wild-type mice (data not shown), suggesting that type I IFN responsiveness was not required for Vα14i NKT cells to respond to LPS. These findings indicate that IFN-γ production by Vα14i NKT cells in response to LPS requires APC activation and the consequent production of cytokines such as IL-12 or IL-23 and IL-18.

CD1d Ag presentation is not required to induce Vα14i NKT cell IFN-γ

We tested whether CD1d recognition in addition to cytokines was essential for the activation of mouse Vα14i NKT cells by LPS. To do this, we cultured purified Vα14i NKT cells, obtained by positive selection, with BM DCs in the presence or absence of LPS. Cytokine production was measured in the supernatants after 48 h (Fig. 3⇓, A and B). In one protocol, Vα14i NKT cells were sorted from mouse spleens using αGalCer-CD1d tetramers and Abs to TCRβ. Neither BM DCs nor Vα14i NKT cells cultured alone with LPS produced IFN-γ, nor did Vα14i NKT cells cultured together with BM DCs without LPS (data not shown). IFN-γ was produced when Vα14i NKT cell-DC cocultures contained LPS (Fig. 3⇓A). Surprisingly, addition of the anti-CD1d blocking Ab 1B1 had only a slight effect on IFN-γ production in this culture, and DCs from CD1d^−/− mice were almost as effective at stimulating IFN-γ in the presence of LPS as wild-type DCs (Fig. 3⇓A). However, DCs from IL-12p40^−/− mice were poor activators in the Vα14i NKT cell cocultures with LPS (Fig. 3⇓A), consistent with the reduced cytokine production by Vα14i NKT cell observed in the ex vivo analysis of cells from IL-12p40^−/− mice (Fig. 2⇑A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Ag presentation by CD1d is not required for IFN-γ production by Vα14i NKT cells in response to LPS. A and B, Vα14i NKT cells (NKT) were cultured with LPS and either wild-type (DC), CD1d^−/− (CD1^−/− DC) or IL-12p40^−/− (IL-12^−/−) BM-derived DCs, in the presence or absence of anti-CD1d blocking Abs (1B1). Vα14i NKT cells were sorted based on staining with one of two schemes: αGalCer-CD1d tetramer (A) and anti-TCRβ Abs or anti-NK1.1 and anti-CD5 Abs (B). The dose of LPS used in B is 10 μg/ml. No IFN-γ production was detected in cultures that contained either BM-derived DCs or Vα14i NKT cells alone. C, Vα14i NKT cells were enriched from the spleens of Vα14 Tg mice (Tg Spl) and cultured with LPS from either E. coli or S. abortus equi (S. abortus) and BM-derived DCs from either CD1d^−/− (CD1^−/− DC) or IL-12p40^−/− (IL-12^−/−). One representative of two experiments is shown. D, Sorted Vα14i NKT cells were cultured, without BM-derived DCs, with two different doses of rIL-12 and IL-18 (high dose: 10 ng/ml IL-12 and 100 ng/ml IL-18. Low dose: 100 pg/ml IL-12 and 1 ng/ml IL-18). IFN-γ levels after 48 h of culture were measured in the supernatants by ELISA. Each panel represents one of at least three independent experiments. ∗, Not done.

Vα14i NKT cells sorted with αGalCer-CD1d tetramers are partially activated, as judged by a modest amount of IL-4 secretion (ranging from 100 to 200 pg/ml in five independent experiments, data not shown) when cultured in vivo without further stimulation. It is therefore possible that the enriched Vα14i NKT cells (>97% purity) received a TCR signal during their enrichment. We used two additional purification schemes to address this issue. In one set of experiments, NKT cells were positively selected from the spleen using Abs to NK1.1 and TCRβ. Although this also could cause some degree of TCR engagement, unlike cells positively selected with the tetramer, these NKT cells did not produce IL-4 in vitro (data not shown). NKT cells purified in this way therefore were cultured with CD1d^−/− BM DCs and LPS (data not shown), and the absence of CD1d expression by the DC had little effect on IFN-γ production. Because Vα14i NKT cells express CD1d, although much lower amounts than BM DC, it is possible that Vα14i NKT cells present glycolipid self-Ag to one another. Therefore, we added anti-CD1d Ab to the LPS-containing cocultures in which CD1d^−/− BM DCs were used. Even under these conditions, when CD1d-mediated Ag presentation was most effectively blocked, similar levels of IFN-γ were elicited (data not shown).

In a second scheme, we used positive selection with Abs for NK1.1 and the inhibitory receptor CD5 (Fig. 3⇑B). This procedure completely avoids any TCR engagement and the resulting population consists of 75% tetramer binding cells (data not shown). Coculture of this highly enriched population with CD1d^−/− BM DCs and LPS in the presence of anti-CD1d blocking Abs also led to IFN-γ secretion (Fig. 3⇑B). To further ensure that Vα14i NKT cells were not activated during purification, these cells were enriched by negative selection from the spleens of mice transgenic for the invariant Vα14 TCR (Vα14 Tg). B cells, CD8 T cells, macrophages, DCs, and γδ T cells were depleted using lineage specific markers and NK cells were depleted by virtue of their CD11b expression (data not shown). The resulting population was ∼15% Vα14i NKT cells (called Tg Spl), and was cultured with BM-derived DCs from CD1d^−/− and IL-12p40^−/− mice (Fig. 3⇑C). Tg Spl cells produced IFNγ in response to LPS from either E. coli or S. abortus equi, even when cultured with CD1d^−/− BM derived DCs in the presence of blocking CD1d Abs (Fig. 3⇑C), but the levels of IFNγ were strongly reduced when the BM derived DCs were from IL-12p40^−/− mice (Fig. 3⇑C). These data indicate that CD1d-mediated Ag presentation is not required for IFN-γ production in cocultures of Vα14i NKT cells and BM DCs with LPS.

APC-derived cytokines are sufficient to drive Vα14i NKT cell activation

Because CD1d expression on DCs appeared to be unnecessary to induce IFN-γ in the DC-T cell cocultures containing LPS (Fig. 3⇑, A–C), we tested whether cytokines alone were sufficient to activate Vα14i NKT cells. Vα14i NKT cells were sorted based on staining for the αGalCer-CD1d tetramer and TCRβ, and the cells cultured with one of two concentrations of rIL-12 (100 pg/ml or 10 ng/ml), IL-18 (1 or 100 ng/ml), or with both cytokines (Fig. 3⇑D), and production of IFN-γ by the cultured Vα14i NKT cells was measured 48 h later. The lower concentrations are physiologically relevant, because similar concentrations can be measured in the serum of mice after LPS injection (see Fig. 4⇓ below). Addition of IL-12 or IL-18 individually led to the synthesis of measurable amounts of IFN-γ from Vα14i NKT cells, even at the lower dose of recombinant cytokines (Fig. 3⇑D). Addition of both cytokines together resulted in strongly enhanced IFN-γ production (Fig. 3⇑D), even at the lower doses of recombinant cytokines (Fig. 3⇑D). Vα14i NKT cells cultured with rIL-12 or IL-18 or both cytokines in the presence of BM-derived DC showed comparable levels of IFN-γ production, indicating that any other LPS-independent APC-derived factors were not important (data not shown). NKT cells sorted based on expression of NK1.1 and TCRβ were cultured with IL-12 and IL-18 with similar results (data not shown), excluding an activating role for the αGalCer-CD1d tetramer. In conclusion, the data are consistent with the hypothesis that the stimulation provided by DC for Vα14i NKT cells in vitro is entirely due to the IL-12 and IL-18 they produce, and that the two cytokines together, even in small quantities, are both necessary and sufficient for maximal IFN-γ production by Vα14i NKT cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Vα14i NKT cell-deficient Jα18^−/− mice have defects in the production of proinflammatory cytokines in response to LPS. A–D, C57BL/6J or Jα18^−/− mice were injected with LPS and bled at the indicated time points. Cytokines levels in the serum were measured by ELISA (A and C) or using the Bioplex system (B and D), as described in detail in Materials and Methods. A, Data shown are averages from two mice, from one experiment of three. B–D, Data shown are the averages of measurements from three mice per group.

Reduced innate immune response in mice lacking Vα14i NKT cells

Having characterized the Vα14i NKT cell response to LPS, we investigated how Vα14i NKT cell deficiency influenced the overall innate immune response to LPS. It has been reported that CD1d expression is necessary for the progression of the generalized Shwartzman reaction (11), an animal model of septic shock in which an initial priming dose (in this model, 40 μg) of LPS is followed 24 h later by a higher (400 μg) challenge dose, resulting in mortality. One other study, however, did not come to a similar conclusion (12). In agreement with the second study, we also found that CD1d^−/− mice on the C57BL/6 background did not have decreased mortality following LPS challenge, with 4 of 11 (36.6%) CD1d^−/− mice surviving compared with 4 of 15 (26.6%) wild-type control animals. In contrast, when Jα18^−/− mice on the C57BL/6 background were analyzed, they were protected, with 11 of 15 (76.6%) surviving. Kaplan-Meier survival curve analysis showed that the survival of Jα18^−/− mice was statistically significant when compared with survival of wild-type mice (p = 0.0120), but the survival of CD1d^−/− mice was not (p = 0.6038). A dichotomy in the results obtained comparing Jα18^−/− and CD1d^−/− mice has been interpreted in other experimental systems as reflecting a difference between the effects of lacking only Vα14i NKT cells (Jα18^−/− mice), and the additional effects of lacking other CD1d-restricted T cells with more diverse TCRs, sometimes called type II NKT cells, in the CD1d^−/− mice (25, 26). In several experimental systems, such as the recurrence of lymphoma growth (25), these type II NKT cells have been shown to have functional capabilities different from Vα14i NKT cells. Regardless, using survival as the endpoint clearly has variable results in different studies, perhaps because of differences in the LPS used, mouse genetic background, or conditions in different vivaria, and does not provide much quantitative insight into the in vivo response to LPS. We therefore focused our analysis on the response to the priming dose of LPS in the Jα18^−/− strain and control mice, because this strain provides the most precise ablation of Vα14i NKT cells.

Jα18^−/− and wild-type control mice were injected with LPS and serum cytokines were measured at the indicated time points after injection, by either cytometric bead array (Fig. 4⇑, B and D) or by ELISA (Fig. 4⇑, A and C). The amounts of serum IL-12p70 (Fig. 4⇑B) and IL-18 (Fig. 4⇑C) in Jα18^−/− mice were comparable to those in wild-type mice at the early hours following LPS injection, while IFN-γ levels 6 h after LPS (Fig. 4⇑A) were strongly reduced in Jα18^−/− mice. The most striking difference, however, was in the amount of serum TNF (Fig. 4⇑D), with a nearly 10-fold reduction in Jα18^−/− mice at 1.5 h after LPS. A similar reduction in serum TNF levels in Jα18^−/− mice was observed in additional experiments, where cytokine levels were measured by ELISA rather than cytometric bead array (data not shown). Because IL-12p70 and IL-18 levels are similar in wild-type and Jα18^−/− mice, the impaired production of these cytokines cannot explain the reduction in serum IFN-γ and TNF seen in Jα18^−/− mice, and the data are consistent with the hypothesis that Vα14i NKT cells act downstream of TLR4-mediated activation of APC, and respond to the IL-12 and IL-18 produced by those APC to stimulate IFN-γ and TNF synthesis.

Impaired LPS activation of innate immune cells in Jα18^−/− mice

We considered it unlikely that Vα14i NKT cells were directly responsible for the high levels of IFN-γ and TNF in the serum of mice injected with LPS. NK cells have been shown to be the principal source of serum IFN-γ after LPS (27), and by ICCS, Vα14i NKT cells did not produce TNF in response to LPS (data not shown). This suggests that activated Vα14i NKT cells rapidly induce cytokine production by other cell types, by analogy to their rapid activation of NK cells after αGalCer injection (28, 29, 30).

IFN-γ production by NK cells from wild-type and Jα18^−/− mice was examined by ICCS 6 h after LPS injection (Fig. 5⇓). As was done for the analysis of Vα14i NKT cells, the gated (NK1.1⁺, TCRβ⁻) NK cells were analyzed immediately ex vivo without further stimulation or the use of brefeldin A. The majority of NK cells, expressed intracellular IFN-γ 6 h after LPS, but this percentage was strongly reduced in Jα18^−/− mice, particularly in the liver (Fig. 5⇓). By ICCS analysis, however, NK cells were not major producers of TNF (data not shown). Previous studies have identified neutrophils and macrophages as some of the primary sources of TNF produced in response to LPS (31), therefore these cell types were tested for TNF production after LPS in Jα18^−/− mice. Because of the very early appearance of TNF in the serum, ICCS analysis for TNF synthesis was conducted 45 min after LPS injection (Table I⇓). Although the data varied between experiments, the percentage of splenic neutrophils (CD11b^high Gr-1^high) and macrophages (CD11b^lowGr-1^lowF4/80⁺, also verified to be CD68^high) from Jα18^−/− mice that produced TNF was consistently lower than that seen in wild-type mice (Table I⇓), suggesting that a defect in myeloid cell-derived TNF was the cause of the reduction seen in systemic TNF in Jα18^−/− mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

NK cells from Jα18^−/− mice display impaired IFN-γ production in response to LPS. C57BL/6J or Jα18^−/− mice were injected with LPS and spleens and livers harvested after 6 h. IFN-γ production was detected by ICCS performed directly ex vivo without any further treatments, and compared with intracellular staining with the appropriate isotype control Ab. The numbers indicate percentages of intracellular IFN-γ⁺ cells. The gates for intracellular IFN-γ levels are set based on intracellular IFN-γ staining in NK cells from control PBS-injected mice (data not shown). Panels on the left show the NK cell gate. Data shown are from one representative mouse, at least five mice were tested in each case.

IFN-γ augments the proinflammatory cascade

We hypothesized that IFN-γ production by Vα14i NKT cells was in part responsible for the augmentation of the innate immune response to LPS. IFN-γ is known to be important for the induction of TNF in the inflammatory response (32, 33, 34) and IFN-γ^−/− mice had a strong reduction in serum TNF 1.5 h after LPS injection (data not shown). Moreover, in the context of glycolipid Ag stimulation of their TCR, the immediate production of IFN-γ by Vα14i NKT cells contributes to the rapid activation of NK cells and other cell types (28, 29, 30, 35). Although expression of activation markers such as CD69 by Vα14i NKT cells was increased at early time points after LPS injection, at time points as early as 1 h, the cells were negative for intracellular IFN-γ when analyzed directly ex vivo (data not shown). Because TNF production by myeloid cells is already robust by 1 h, it therefore appeared that the bulk of the IFN-γ synthesis by Vα14i NKT cells occurs too late to be responsible for the early wave of myeloid cell TNF production. As noted above, our ICCS protocol should closely reflect the in vivo situation, but it is not a highly sensitive assay, and therefore it may not be able to detect lower amounts of early cytokine synthesis. To determine whether Vα14i NKT cells express detectable IFN-γ at very early time points after LPS injection, these cells were sorted from the livers and spleens of mice 1 h after LPS injection, and IFN-γ mRNA expression was measured by real-time PCR (Fig. 6⇓A). IFN-γ mRNA was enhanced in Vα14i NKT cells 1 h after LPS injection compared with similarly prepared cells from control mice, with a 5-fold level of enhancement seen in the liver (Fig. 6⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

IFN-γ from Vα14i NKT cells may augment the production of TNF in response to LPS. A, RNA was purified from Vα14i NKT cells sorted from the spleens or livers of C57BL/6J mice that had been injected 1 h earlier with either PBS or LPS. IFN-γ or TNF message levels were measured by real-time PCR. Data shown are from one of two independent experiments. B, C57BL/6J or Jα18^−/− mice were injected with LPS, and another cohort of Jα18^−/− mice was injected with LPS together with 100 ng of recombinant IFN-γ (Jα18^−/− + rIFNγ). The mice were bled 1.5 h after injection and TNF levels in the serum measured by ELISA. The numbers of mice per group are indicated beneath the x-axis. Data shown are from one of two independent experiments.

To test whether reduced early IFN-γ synthesis was responsible for the reduction in serum TNF seen in response to LPS in Jα18^−/− mice, these mice were injected with rIFN-γ together with LPS, and TNF production in the serum was measured after 1.5 h (Fig. 6⇑B). A total of 100 ng/mouse of rIFN-γ was sufficient to induce an ∼2-fold increase in serum TNF when coadministered with LPS (Fig. 6⇑B). These observations suggest that the reduction in serum TNF observed in Jα18^−/− mice in response to LPS may be due to reduced early production of IFN-γ, and taken together, they further suggest that the early synthesis of IFN-γ by Vα14i NKT cells could be a critical factor for the full activation of the proinflammatory cascade in response to LPS.