Requirement for Type 2 NO Synthase for IL-12 Signaling in Innate Immunity

See allHide authors and affiliations

Science  07 May 1999:
Vol. 284, Issue 5416, pp. 951-955
DOI: 10.1126/science.284.5416.951


Interleukin-12 (IL-12) and type 2 NO synthase (NOS2) are crucial for defense against bacterial and parasitic pathogens, but their relationship in innate immunity is unknown. In the absence of NOS2 activity, IL-12 was unable to prevent spreading ofLeishmania parasites, did not stimulate natural killer (NK) cells for cytotoxicity or interferon-γ (IFN-γ) release, and failed to activate Tyk2 kinase and to tyrosine phosphorylate Stat4 (the central signal transducer of IL-12) in NK cells. Activation of Tyk2 in NK cells by IFN-α/β also required NOS2. Thus, NOS2-derived NO is a prerequisite for cytokine signaling and function in innate immunity.

The innate immune response to bacteria and protozoan parasites is characterized by the rapid recognition of microbial antigens, after which activated inflammatory cells release soluble mediators and antimicrobial effector molecules (1). The early production of IL-12 by granulocytes, macrophages, or dendritic cells stimulates the cytotoxic activity of NK cells and enhances their release of IFN-γ. Along with IL-12, IFN-γ then facilitates the later development of type 1 T helper cells (TH1 cells) that ultimately activate macrophages for the destruction of intracellular pathogens through cognate interactions and further secretion of IFN-γ (2). A key antimicrobial agent implicated in this killing process is nitric oxide (NO) generated from the amino acid l-arginine by the inducible isoform of NO synthase (called iNOS or NOS2) (3). NOS2-derived NO can also regulate T cell proliferation, cytokine production, apoptosis, and signaling activity in vitro (4–6) and in vivo (7–9). In genetically resistant mice cutaneously infected with the protozoan parasite Leishmania major, the inhibition of early parasite spreading, the up-regulation of IFN-γ, and the induction of NK cell cytotoxicity at day 1 of infection were abolished after genetic deletion or functional inactivation of NOS2 (9), a response that is similar to that of mice after neutralization of IL-12 (10). This suggests that NOS2 deficiency might affect the availability of IL-12. Although the baseline expression of IL-12 subunit p40 mRNA in naı̈ve NOS2−/− mice was lower as compared to that of NOS2+/+ mice, the production of active IL-12 p70 heterodimers by inflammatory macrophages was not reduced in the absence of NOS2 (9). Thus, both NO/NOS2 and IL-12 regulate the innate response to L. major, but whether and how they might interact in vivo remained unknown. We show here that NO/NOS2 is an integral part of the IL-12 signaling cascade in NK cells and therefore constitutes a prerequisite for the function of IL-12 in innate immunity.

We initially analyzed whether responsiveness to exogenous IL-12 in vivo required NOS2 activity. Wild-type NOS2+/+mice (129/SvEv × C57BL/6) were cutaneously infected with L. major promastigotes (11). At 24 hours, the parasites were found at the site of inoculation and in the draining popliteal lymph node (pLN). In NOS2-deficient mice (9, 12), in contrast, the parasites were disseminated to visceral organs, even after treatment with IL-12 (13) (Fig. 1A). In genetically susceptible BALB/c mice, which lack functional IL-12 early during infection (10, 14), treatment with IL-12 prevented parasite spreading at day 1 of infection (Fig. 1A). However, simultaneous application ofl-N 6-iminoethyl-lysine (L-NIL) (13), a potent competitive inhibitor of NOS2 (9), reversed the protective effect of IL-12 (Fig. 1A). Treatment of NOS2+/+ mice with IL-12 increased expression of IFN-γ mRNA and NK cell cytotoxic activity in the pLN after infection, whereas no such effect was observed in NOS2−/− mice or in NOS2+/+ mice injected with L-NIL (Fig. 1, B and C). Similar amounts of IL-12 receptor β1 and β2 mRNA (15) were detected in the pLN of NOS2+/+ and NOS2−/−mice at day 1 of infection (Fig. 1D), indicating that the impaired parasite control, IFN-γ expression, and NK cell activity in NOS2−/− mice at day 1 of infection (9) resulted from an inability to respond to IL-12 but not from a lack of IL-12 receptor expression.

Figure 1

NOS2 is required for IL-12 responsiveness in vivo. NOS2+/+ or NOS2–/– mice (129/SvEv × C57BL/6) (9,12) or BALB/c mice were infected with L. major(11) with or without application of PBS, rmIL-12, or L-NIL (13). (A) Parasite spreading in L. major–infected mice at 24 hours of infection. DNA from given organs was analyzed for the presence of a 120-bp fragment ofL. major kinetoplast DNA by PCR (9). One of four experiments is shown. (B) Cytotoxic activity of pLN cells from day 1–infected mice. NK cell activity was tested against51Cr-labeled YAC-1 cells as targets (9). One of three experiments is shown. (C) Expression of IFN-γ mRNA in the pLN from uninfected or day 1–infected mice. Quantitation was by competitive PCR (9). One of three experiments is shown. (D) mRNA expression of IL-12 receptor β1 and β2 chain in uninfected or day 1–infected mice as quantitated by competitive PCR (15). One of two experiments is shown. ND, not detectable.

The primary source of IFN-γ during the innate response toLeishmania is the NK cell (16). We therefore tested the production of IFN-γ by various NK cell populations for its dependency on NOS2-derived NO upon stimulation with IL-12. By flow cytometry analysis, total pLN cells from day 1–infected NOS2+/+ and NOS2−/− mice always contained a similar percentage (1 to 3%) of NK1.1+ CD3cells (17). NOS2+/+ pLN cells produced up to 2 to 5 ng of IFN-γ per milliliter after stimulation with 0.1 to 10 ng of IL-12 per milliliter, whereas less than 200 pg/ml or no IFN-γ was detectable in cultures from NOS2−/− mice or in L-NIL–treated cultures from NOS2+/+ mice (Fig. 2A). The unresponsiveness of NOS2−/− cells was restricted to day 1 of infection and was not due to an up-regulation of endogenous transforming growth factor–β (17). At later time points, during the expansion of IFN-γ–producing TH1 cells, wild-type and NOS2-deficient pLN cells responded equally well to IL-12 (Fig. 2A). Also, when the pLN cells were stimulated with a T cell mitogen (concanavalin A) instead of IL-12, the production of IFN-γ by NOS2+/+ at day 1 to 21 of infection was indistinguishable from that of NOS2−/− cells (17). These data suggest NOS2-dependent IFN-γ production by early responding NK cells rather than T cells.

Figure 2

NOS2 is required for IL-12–induced IFN-γ production by NK cells. pLN cells, NK cells, or TH1 cells were stimulated with IL-12 (100 pg/ml). The NOS2 inhibitor L-NIL (1 mM) was added ≥2 hours before stimulation. After 24 hours, IFN-γ was determined in the culture supernatants by ELISA (sensitivity 19.5 to 39 pg/ml) (9). (A) Total pLN cells from NOS2+/+ or NOS2−/− mice (each with 1 to 3 % NK1.1+ CD3 cells) after 1, 5, or 15 days of infection with L. major (11). One of three experiments is shown. (B) NK cells (18) enriched from the pLN of NOS2+/+ (10.7 % NK1.1+) and NOS2−/− (10.4 % NK1.1+) mice after 1 day of infection with L.major (11). Culture was with or without IL-2 (100 U/ml). One of three experiments is shown. (C) NK cells from the spleen of naı̈ve C57BL/6 mice (NOS2+/+, NOS2−/−, or Rag1−/−), stimulated in the presence of IL-2 directly after purification (19) or after expansion with IL-2 for 5 days (20). One of two to five experiments is shown. (D) KY-1 NK cells (21,22). DETA-NO or DETA was added immediately before IL-12. For clarity, only some of the control values in the absence of L-NIL are shown. One of three experiments is shown. (E) Primary T cells (24) or CD4+ TH1 cells [line LNC-2 and clone B10BI (25)], stimulated in the absence of IL-2. One of three experiments is shown.

Next, we partially purified NK cells from pLN cell suspensions of day 1–infected mice (18). The NK cell–enriched population from NOS2+/+ mice showed high production of IFN-γ after stimulation with IL-12 (in the presence or absence of IL-2) that was completely blocked by L-NIL. The corresponding cell population from NOS2−/− mice was comparably composed but did not respond to IL-12 (Fig. 2B). Similar results were obtained (i) with NK1.1+ CD3 cells purified to near homogeneity from the spleen of noninfected C57BL/6 (NOS2+/+, NOS2−/−, or T and B cell–free Rag1−/−) mice (19) (Fig. 2C); (ii) with T cell–depleted and IL-2–expanded adherent splenic NK1.1+ CD3 NK cells derived from naı̈ve NOS2+/+ or NOS2−/− mice (20) (Fig. 2C); and (iii) with the mouse NK cell clone KY-1 (21, 22). IL-12 (≥0.1 pg/ml) strongly up-regulated the IFN-γ release by KY-1 cells, which was blocked by L-NIL [Fig. 2D and Web Fig. 1(23)] but not by l-nitroarginine-methylester (L-NAME), which primarily inhibits constitutive NO synthases, or by the inactive analog D-NAME (17). The addition of small amounts (≥1 μM) of NO donors (DETA/NO or S-nitrosoglutathione) completely reversed the inhibitory effect of L-NIL, indicating that exogenous NO can facilitate the activation of NK cells. The control compounds diethylene-triamine (DETA) and glutathione were inactive [(Fig. 2D) and (17)]. Thus, IL-12–stimulated NK cell production of IFN-γ is NOS2-dependent.

IL-2–propagated KY-1 cells expressed small amounts of NOS2 protein, which was partially dependent on endogenous IFN-γ (17) and was increased upon stimulation with IL-12 (Fig. 3A). Treatment of KY-1 cells with IL-12 led to the accumulation of 8.8 ± 0.9 μM NO2 within 24 hours, as compared to 2.6 ± 0.7 μM in the absence of IL-12 and 0.8 ± 0.1 μM in cultures without IL-2 and IL-12 (mean ± SEM of five experiments). L-NIL, which antagonized the IL-12–induced production of IFN-γ, did not prevent the stimulatory effect of IL-12 on NOS2 protein expression, which indicates that some functions of IL-12 are preserved in the absence of NOS2 activity (Fig. 3A). NOS2 protein was also found in wild-type (but not in NOS2−/−) splenic NK cells expanded in IL-2 and stimulated by IL-12 (Fig. 3A). The average production of NO2 by these cells was 12.2 ± 2.9 μM NO2 within 48 hours, as compared to 4.9 ± 2.1 μM NO2 in the absence of IL-12 (mean ± SEM of five experiments). In contrast to NK cells, the IL-12–induced production of IFN-γ by purified splenic T cells (24) and by two TH1 cell clones (B10BI and LNC-2) was refractory to inhibition by L-NIL (Fig. 2E). This is in line with the lack of NOS2 in these cells (25). Thus, NOS2-derived NO is required for the activation of NK cells, but not T cells, by IL-12.

Figure 3

NK cells express NOS2 activity, which is required for tyrosine phosphorylation of Stat4. KY-1 cells (21, 22), IL-2–expanded splenic NK cells (20), or splenic T-cells (24) from C57BL/6 mice (NOS2+/+ or NOS2−/−) were stimulated with IL-12. L-NIL (1 mM) was added 24 hours before stimulation and DETA/NO or DETA was added 2 hours before stimulation. One of two to four experiments is shown. (A) Anti-NOS2 immunoblot (IB) after 24 hours of stimulation with IL-12 (38). (B and C) Detection of tyrosine-phosphorylated Stat4 and total Stat4 protein in NK or T cells after 60 min of stimulation by immunoprecipitation (IP) and subsequent immunoblotting (27). α, anti.

Because IL-12 signaling in NK cells is strictly dependent on the signal transducer and activator of transcription 4 (Stat4) (26), we investigated whether NOS2 activity was critical for the activation of Stat4 (27). Stimulation of KY-1 cells with IL-12 caused tyrosine phosphorylation of Stat4 at 15 to 90 min, which was abolished when L-NIL was added to KY-1 cells at least 2 hours before IL-12 [(Fig. 3B) (17)]. The same result was obtained with IL-2–expanded splenic NK cells from NOS2+/+versus NOS2−/− mice, whereas tyrosine phosphorylation of Stat4 in purified primary T cells was NOS2-independent (Fig. 3B). Tyrosine phosphorylation of Stat4 in NK cells was completely restored when L-NIL–treated cells were exposed to small amounts of DETA/NO (Fig. 3C). As in human NK cells after culture in IL-2 (28), KY-1 cells tyrosine phosphorylated Stat1α, but unlike Stat4, the tyrosine phosphorylation was neither induced nor modulated by IL-12 and was also not affected by inhibition of NOS2 activity (17). Thus, NOS2 activity is specifically required for IL-12–stimulated Stat4 activation in NK cells.

NO is known to activate soluble guanylyl cyclase, which results in elevated concentrations of cyclic guanosine monophosphate (cGMP) and activation of cGMP-dependent kinases and ion channels (29). We added a potent and highly specific activator of cGMP-dependent kinases and of cGMP-gated ion channels [8-(4-chlorophenylthio)-guanosine-3′,5′-cyclic monophosphate (8-pCPT-cGMP) (29)] to KY-1 cells that were stimulated with IL-12 in the presence of L-NIL. 8-pCPT-cGMP did not restore tyrosine phosphorylation of Stat4 [Web Fig. 2 (30)]. Thus, NOS2-derived NO did not activate Stat4 via cGMP-dependent pathways.

The Janus kinases Jak2 and Tyk2 are tyrosine phosphorylated and activated in response to IL-12 and are thought to phosphorylate the Stat4 transcription factor in hematopoietic cells, including NK cells (28, 31). We therefore analyzed the effect of NOS2/NO on the activation of Jak2 and Tyk2 (32). Stimulation of KY-1 cells with IL-12 caused tyrosine phosphorylation of Jak2 and Tyk2, which remained unaltered by L-NIL (Fig. 4A). Thus, IL-12–induced tyrosine phosphorylation of both kinases (and hence the expression of a functional IL-12 receptor) was not dependent on NOS2 activity. However, when both kinases were immunoprecipitated from KY-1 cells after stimulation with IL-12 and subsequently tested for their catalytic activity (33), inhibition of NOS2 by L-NIL blocked the autokinase activity of Tyk2 but not of Jak2 (Fig. 4A). These results were confirmed with IL-2–expanded splenic NK cells from NOS2+/+ and NOS2−/−mice [(Fig. 4B) (17)]. In contrast, the induction of Tyk2 kinase activity by IL-12 in purified splenic T cells or in the TH1 cell line B10BI was independent of NOS2 [(Fig. 4B) (17)]. Tyk2 kinase activity of KY-1 cells was restored when the intact cells (stimulated with IL-12 plus L-NIL) (Fig. 4C) or the respective anti-Tyk2 immunoprecipitates were treated with DETA/NO [Web Fig. 3 (34)]. DETA/NO did not alter the tyrosine phosphorylation of Tyk2 (17). These findings suggest that stimulation of NK cells with IL-12 first causes tyrosine (auto)phosphorylation and activation of Jak2, which then transphosphorylates Tyk2. For the activation of Tyk2 function, NOS2-derived NO is required as an independent second signal.

Figure 4

NOS2-derived NO is required for the activation of Tyk2 kinase but not of Jak1 or Jak2 kinase. (A throughC) KY-1 cells (21, 22) or splenic NK (20) or T cells (24) from C57BL/6 (NOS2+/+ or NOS2−/−) mice were stimulated with IL-12 or IFN-α/β (±1 mM L-NIL) for 15 min, lysed, and subjected to immunoprecipitation with antibody to Jak1 (α-Jak1), α-Jak2, or α-Tyk2 (in the absence or presence of blocking peptide), followed by anti-phosphotyrosine immunoblotting (32), or were immunoprecipitated for in vitro autokinase assays (33). To control for equal loading of the lanes, blots were reprobed with α-Jak1, α-Jak2, or α-Tyk2. One of three experiments is shown. (C) NO donor (DETA/NO) or DETA was added to L-NIL–treated KY-1 cells 1 hour before IL-12.

Finally, we tested whether Tyk2 kinase also requires NOS2/NO for activity when type I interferon (IFN-α/β) is used as an alternative stimulus (35). In KY-1 cells or primary splenic NK cells (from NOS2+/+, NOS2−/−, or Rag1−/− C57BL/6 mice) activated with IFN-α/β, the activity of Tyk2 kinase as well as the production of IFN-γ were dependent on endogenous NOS2 activity, whereas the activity of Jak1 remained unaffected in the absence of NOS2 [(Fig. 4, A and B) (WebFig. 4) (17, 36)]. In contrast, when NK cells were activated with IL-18, which does not signal via Tyk2 and Stat4 (37), the production of IFN-γ remained unaltered in the absence of NOS2 (36). In primary T cells or TH1 cell lines (B10BI and LNC-2) IFN-α/β–mediated activation of Tyk2 kinase and IFN-γ production were NOS2-independent [(Fig. 4B) (Web Fig. 4) (17,36)]. These results are in line with the strict requirement for endogenous NOS2 for a protective function of IFN-α/β at day 1 of L. major infection (9).

Several interactions between NOS2 and IL-12 have been reported in the past, including the inhibition of macrophage IL-12 production by NO, the possible induction of the IL-12 antagonist IL-12(p40)2by NO, and the NOS2-dependent suppression of T cell responses by IL-12 (6, 7). The latter might result from an inhibition of Jak2 and Jak3 kinase or a disruption of the Jak3/Stat5 signaling pathway (5). These negative regulatory functions of NO contrast with our study, which identifies NOS2-derived NO as an indispensable and positive regulatory element in the IL-12 signaling pathway of NK cells during the innate response to a protozoan parasite. The effect of NO, whose molecular nature remains to be determined, is selective because (i) it pertains to NK cells (expressing NOS2) but not to T cells (lacking NOS2), and (ii) it is a prerequisite for the activation of Tyk2 but not of two other Janus kinases (Jak1 and Jak2). At day 1 of infection with L. major, NOS2 was focally expressed by dermal macrophages, most likely as a consequence of the induction of IFN-α/β by the parasites (9). We hypothesize that the small quantities of NO generated very rapidly after infection by macrophages (or endogenously by NK cells after exposure to IL-2) capacitate NK cells to respond to IL-12 and IFN-α/β, which leads them to become cytotoxic and to release IFN-γ. We propose that this signaling function of NO is critical for the T cell–independent containment of bacterial and parasitic infections at a time when direct NO-mediated control of the microbes does not yet occur.

  • * Present address: Department of Molecular and Cell Biology and Cancer Research Laboratory, 485 LSA, University of California, Berkeley, CA 94720, USA.

  • To whom correspondence should be addressed. E-mail: christian.bogdan{at}


Stay Connected to Science

Navigate This Article