Regulation of IFN-γ production in granulocyte-macrophage colony-stimulating factor-deficient mice

0014-2980/98/1212-3980$17.50+.50/0 © WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998

Regulation of IFN- q production in granulocyte-macrophage colony-stimulating factor-deficientmice

Yuji Noguchi, Hisashi Wada, Michael W. Marino and Lloyd J. Old

Ludwig Institute for Cancer Research, New York Branch at Memorial Sloan-Kettering CancerCenter, New York, USA

Granulocyte-macrophage colony-stimulating factor-deficient (GM-CSF−/−) mice producefar lower serum levels of IFN- + in response to LPS than GM-CSF+/+ mice. CD4+ and CD8+

T cells from LPS-injected GM-CSF−/− mice showed a deficiency in IFN- + production andproliferative activity in response to IL-2 and IL-12, whereas IFN- + production by NK cells wasnot compromised. These defects of T cells were reversed by administration of GM-CSF invivo, but not by supplementation with GM-CSF in vitro. GM-CSF−/− mice do not have anintrinsic defect in IFN- + production, because IL-12 injection induces the same high levels ofIFN- + in GM-CSF−/− and GM-CSF+/+ mice. To investigate the inhibitory effect of LPS onGM-CSF−/− T cells and the indirect restorative activity of GM-CSF, we tested the action ofsupernatants from cultured dendritic cells (DC). A factor or factors in the DC supernatantnormalized serum IFN- + levels and T cell responses in LPS-injected GM-CSF−/− mice. IL-18reproduced some but not all of these in vivo and in vitro effects of DC supernatants. Ourresults indicate that GM-CSF is important in protecting T cells from inhibitory signals gener-ated during immunization or exposure to LPS, and that this effect of GM-CSF is indirect andmediated by factors produced by DC.

Key words: IFN- + / Granulocyte-macrophage colony-stimulating factor

Received 16/6/98Revised 26/8/98Accepted 27/8/98

[I 18524]

Abbreviations: DC: Dendritic cell DC-DF: DC-derivedfactor

1 Introduction

GM-CSF is well known for its role in regulating hemato-poiesis through its effects on the growth and differentia-tion of granulocytes and macrophages [1, 2]. Anotheractivity of GM-CSF that is now recognized is its criticalrole in the maturation and proliferation of dendritic cells(DC) [3]. GM-CSF-deficient (GM-CSF−/−) mice haverecently been generated [4, 5] and found to be relativelyresistant to the lethal effects of LPS [6]. In analyzing thisLPS resistance, Basu et al. [6] found that LPS-injectedGM-CSF-deficient mice produced lower levels of IFN- +than wild-type mice, whereas TNF production was unaf-fected. In an investigation of the immune response ofGM-CSF−/− mice, Wada et al. [7] found that the devel-opment of a humoral immune response was delayed inthese mice and that this was associated with an antigen-

specific deficiency in the proliferative response andIFN- + production by CD4+ T cells. This defect could bereversed by a factor or factors present in the supernatantof DC cultured with GM-CSF, and it was suggested thatthe factor(s) acted by protecting cells from inhibitory sig-nals generated during an immune response [7].

In the present study, we analyzed the basis for the poorIFN- + response in LPS-injected GM-CSF−/− mice, andreport that the LPS-induced defect in IFN- + productionwas also corrected by a GM-CSF-induced DC factor.

2 Results

2.1 Restoration of serum IFN- q in LPS-injectedGM-CSF−/− mice by GM-CSF

In initial studies, mice were injected i.p. with 100 ? g LPSand serum levels of cytokines were measured. Fig. 1shows the kinetics of cytokine production/release afterLPS administration. Confirming Basu et al. [6], the promi-nent finding was low serum levels of IFN- + in GM-

3980 Y. Noguchi et al. Eur. J. Immunol. 1998. 28: 3980–3988

Figure 1. Serum levels of cytokines following LPS adminis-tration. Mice were injected i.p. with 100 ? g LPS and bledfrom the retroorbital plexus. GM-CSF+/+ mice ( 1 ); GM-CSF−/− mice ( Æ ). Each data point represents the mean ± SDof four mice.

Figure 2. Restoration of IFN- + serum levels in LPS-injectedGM-CSF−/− mice by GM-CSF. GM-CSF−/− mice wereinjected i.p. with 100 ? g LPS along with 200 ng ( ‡ ), 100 ng( ß ), 50 ng ( , ), 25 ng GM-CSF (+) or no ( Æ ) GM-CSF andbled 7 h later. As controls, GM-CSF+/+ mice were injectedi.p. with 100 ? g LPS alone ( 1 ). Each symbol represents anindividual mouse and five mice were included in each group.

CSF−/− mice. The production of other cytokines, withthe exception of GM-CSF, was comparable in GM-CSF−/− and +/+ mice. To investigate whether IFN- + pro-duction could be restored by GM-CSF, 100 ? g LPS plusvarious doses of GM-CSF was injected i.p. into GM-CSF−/− mice and the mice were bled 7 h later. As shownin Fig. 2, GM-CSF−/− mice injected with 100 ng GM-CSF and LPS produced levels of serum IFN- + compara-ble to those of LPS-injected GM-CSF+/+ mice. Lowerdoses of GM-CSF had less or no restorative activity, anda higher dose appeared to have a suppressive effect on

IFN- + levels. GM-CSF exerted no significant effect onserum levels of other cytokines, such as TNF- § , IL-10and IL-12 p40, in LPS-injected GM-CSF−/− mice.

2.2 Low IFN- q production by T cells from GM-CSF−/− mice

As T cells and NK cells are the two major sources ofIFN- + , CD4+ T cells, CD8+ T cells, and CD16/32+ NK cellswere isolated from LPS-injected mice and cultured withIL-2 and IL-12 at concentrations shown in Fig. 3. Follow-ing a 3-day incubation period, supernatants from thesecultures were tested for IFN- + by ELISA. Lower levels ofIFN- + were observed in cultures of both CD4+ and CD8+

T cells from LPS-injected GM-CSF−/− mice. In contrast,IFN- + production by CD16/32+ cells was not compro-mised (Fig. 3). These in vitro studies indicate that lowserum levels of IFN- + after LPS are due to reduced IFN- +production by GM-CSF−/− T cells.

2.3 Serum levels of IFN- q in T cell-depleted mice

GM-CSF−/− and +/+ mice were immunodepleted ofCD4+ and CD8+ T cells and injected with 100 ? g LPS. Incontrast to the low levels of serum IFN- + in non-depletedLPS-injected GM-CSF−/− mice, T cell-depleted GM-CSF−/− mice produced levels of IFN- + comparable tothose of GM-CSF+/+ mice (Table 1). This suggests thatT cells in LPS-injected mice have an inhibitory effect onIFN- + production by NK cells.

Eur. J. Immunol. 1998. 28: 3980–3988 Regulation of IFN- + production in GM-CSF−/− mice 3981

Figure 3. Low IFN- + production by T cells from LPS-injected GM-CSF−/− mice. Mice were injected i.p. with 100 ? g LPS andkilled 3 h later. CD4+ or CD8+ T cells (1 × 105) or 2 × 105 CD16/32+ cells per well were cultured with IL-2 and IL-12 at the indicatedconcentrations. Supernatants were tested for IFN- + by ELISA after a 3-day culture period. GM-CSF+/+ mice ( 1 ); GM-CSF−/−mice ( Æ ). Each data point represents the mean ± SD of four mice.

Table 1. Serum levels of IFN- + in T cell-depleted GM-CSF−/− or GM-CSF+/+ mice after LPS injectiona)

T cell immunodepletion Serum IFN- + (pg/ml ± SD)

GM-CSF−/− GM-CSF+/+

No 18.8 ± 10.7 313.1 ± 66.6

Yes 489.7 ± 144.4 526.5 ± 128.3

a) For immunodepletion of T cells, mice were injected i.v.with 50 ? l anti-CD4 mAb and 25 ? l anti-CD8 mAb. Oneweek after this treatment, mice were injected i.p. with100 ? g LPS and bled 7 h later. Each data point representsthe mean ± SD of four mice.

2.4 Normal IL-2 and IL-12 responses of T cellsfrom naive GM-CSF−/− mice

IL-12 has been reported to mediate the up-regulation ofIFN- + production induced by LPS [8]. We thereforetested the IFN- + response of GM-CSF−/− mice to IL-12.Mice were injected i.p. with different doses of IL-12 oncea day for 4 days and bled 2 h after the last injection.Table 2 shows that IL-12-induced serum IFN- + levels inGM-CSF−/− mice were comparable to levels in GM-CSF+/+ mice. T cell depletion resulted in an equal reduc-tion in serum IFN- + levels in IL-12-injected GM-CSF−/−and +/+ mice. To analyze the basis for low IFN- + induc-tion by LPS and high IFN- + induction by IL-12 in GM-

CSF−/− mice, in vitro T cell responses to IL-2 and IL-12were examined in naive and LPS-injected mice. CD8+

T cells from naive or LPS-injected mice were culturedwith IL-2 and IL-12 for 3 days and IFN- + production andproliferative response were measured. As shown inFig. 4A, T cells from LPS-injected GM-CSF−/− miceshowed suppressed proliferative responses as well asinhibited IFN- + production in comparison to LPS-injected GM-CSF+/+ mice. In contrast, the responses ofT cells from naive GM-CSF−/− mice to IL-2 and IL-12were comparable to the responses of naive GM-CSF+/+

Table 2. Serum levels of IFN- + in GM-CSF−/− or GM-CSF+/+ mice after injection of IL-12a)

Doses of IL-12(per injection)

Serum IFN- + (pg/ml ± SD)

GM-CSF−/− GM-CSF+/+

5 ? g 1655.3 ± 871.0 1735.0 ± 474.4

1 ? g 506.8 ± 246.7 504.5 ± 194.2

(125.6 ± 24.8) (110.9 ± 12.6)

0.1 ? g 45.4 ± 23.5 37.6 ± 8.0

a) Mice were injected i.p. with various doses of IL-12 once aday for 4 days and bled 2 h after the last injection. Serawere tested for IFN- + by ELISA. Data shown in parenthe-ses are serum IFN- + levels in T cell-depleted mice. Eachdata point represents the mean ± SD of four mice.


Figure 4. (A) IFN- + production and proliferative response of T cells from LPS-injected and naive GM-CSF−/− and +/+ mice. Iso-lated CD8+ T cells (1 × 105 cells for IFN- + assays and 2 × 105 cells for proliferation assays) were cultured for 3 days with 10 pg/mlIL-12 and 100 U/ml (in the case of T cells from LPS-injected mice) or 10 pg/ml IL-12 and 1000 U/ml IL-2 (in the case of T cellsfrom naive mice). (B) Effect of in vivo or in vitro supplementation with GM-CSF on IFN- + production and proliferative response ofCD8+ T cells. For in vivo supplementation with GM-CSF, mice were injected i.p. with 100 ? g LPS alone, or 100 ? g LPS with100 ng GM-CSF, and spleens were harvested 3 h later. For in vitro supplementation, 25 ng/ml GM-CSF was added to cultures ofCD8+ T cells from LPS-injected GM-CSF−/− mice. For IFN- + assays, 1 × 105 CD8+ T cells were cultured with 100 U/ml IL-2 and100 pg/ml IL-12 for 3 days. For proliferation assays, 2 × 105 T cells were cultured with 100 U/ml IL-2 for 3 days. IFN- + in superna-tants was measured by ELISA and proliferative responses were determined by incorporation of methyl-[3H]thymidine. Each datapoint represents the mean ± SD of four mice.

mice. Administration of GM-CSF to LPS-injected GM-CSF−/− mice restored the proliferative responses andIFN- + production by T cells, whereas supplementationwith GM-CSF in vitro had no effect (Fig. 4B). Theseresults indicate that T cells from GM-CSF−/− mice areendowed with the capacity to respond normally to IL-2and IL-12, but that LPS induces an inhibitory effect onthese responses. GM-CSF, in an indirect fashion, abro-gates this inhibitory activity.

2.5 Restoration of IFN- q production by T cellsfrom LPS-injected GM-CSF−/− mice by a DC-derived factor(s)

To analyze the indirect effect of GM-CSF on T cells, DCwere isolated from LPS-injected GM-CSF+/+ mice andcultured with 10 ng/ml GM-CSF and 2 ng/ml IL-4 for 4days. Supernatants from these cultures were transferredto cultures of CD8+ T cells from LPS-injected GM-CSF−/− mice. As shown in Fig. 5, supernatant from DC

cultures restored IFN- + production by CD8+ T cells, butdid not restore the defective proliferative response.Supernatant from DC cultured in a comparable mannerfrom LPS-injected GM-CSF−/− mice or naive GM-CSF+/+ mice had no effect on IFN- + production by CD8+

T cells from LPS-injected GM-CSF−/− mice, indicatingthat in vivo priming of DC with LPS is necessary toinduce the DC-derived factor and that DC from GM-CSF−/− mice may not be in the appropriate maturationor activation stage to produce the factor. To compare theeffects of DC supernatant with IL-18, a recently definedIFN- + -inducing cytokine, CD8+ T cells from LPS-injectedGM-CSF−/− mice were cultured with the following com-binations: (1) 20 ng/ml IL-18 plus 100 pg/ml IL-2 or1000 U/ml IL-12 and (2) 50 ? l of culture supernatant plus100 pg/ml IL-2 or 1000 U/ml IL-12. Supernatant from DCcultures up-regulated IFN- + production by CD8+ T cellscultured with either IL-12 or IL-2, whereas IL-18 waseffective in augmenting IFN- + production only in thepresence of IL-12 (Table 3). To extend this comparison,LPS plus 0.5 ml DC culture supernatant or LPS plus


Figure 5. Restoration of IFN- + production by CD8+ T cells from LPS-injected GM-CSF−/− mice by DC culture supernatants. DCsupernatant (50 ? l) with 100 U/ml IL-2 and 100 pg/ml IL-12 (total volume of 200 ? l) was added to 1 × 105 CD8+ T cells from LPS-injected GM-CSF−/− mice. For proliferation assays, 2 × 105 CD8+ T cells were cultured in a comparable manner with DC super-natant and 100 U/ml IL-2. For control purposes, the effect of 2.5 ng/ml GM-CSF and 0.5 ng/ml IL-4 added to cultures of 1 × 105

CD8+ T cells from LPS-injected GM-CSF−/− mice was tested. Following a 3-day culture period, supernatants were tested forIFN- + by ELISA. Proliferative responses were determined by incorporation of methyl-[3H]thymidine. Each data point representsthe mean ± SD of three mice.

Figure 6. (A) Restoration of serum IFN- + levels in LPS-injected GM-CSF−/− mice by DC culture supernatants. GM-CSF−/− micewere injected with 100 ? g LPS along with 0.5 ml DC culture supernatant or 100 ng IL-18 or 5 ng GM-CSF and 1 ng IL-4. For con-trol purposes, GM-CSF+/+ mice were injected with 100 ? g LPS alone. All mice were bled 7 h later. Each data point represents themean ± SD of four mice. (B) In vitro IFN- + production and proliferative response of CD8+ T cells from GM-CSF−/− mice injectedwith 100 ? g LPS along with 0.5 ml DC culture supernatant or 100 ng IL-18 or 5 ng GM-CSF and 1 ng IL-4. T cells harvested 3 hafter injection were cultured as described in the legend to Fig. 4. The data represent the mean ± SD of four mice.

100 ng IL-18 were injected into GM-CSF−/− mice. Asshown in Fig. 6A, the administration of DC culture super-natant or IL-18 up-regulated serum IFN- + levels in LPS-injected GM-CSF−/− mice. Injection of 5 ng GM-CSFand 1 ng IL-4 (the two cytokines used in the DC cultures)exerted no effect on serum IFN- + levels. Fig. 6B showsthe IFN- + production and proliferative response in vitro ofCD8+ T cells from GM-CSF−/− mice injected in a compa-rable manner with LPS along with DC supernatant orIL-18. CD8+ T cells from GM-CSF−/− mice injected withLPS and DC culture supernatant produced similar levelsof IFN- + and proliferated as well as those from LPS-injected GM-CSF+/+ mice, whereas administration ofIL-18 did not reverse the LPS-induced inhibition of T cell

proliferation and IFN- + production, indicating that factorsother than, or in addition to, IL-18 in DC culture superna-tants were responsible for up-regulating the IL-2 andIL-12 responsiveness of T cells from LPS-injected GM-CSF−/− mice.

3 Discussion

Since the initial recognition of IFN- + as an anti-viral agent[9], IFN- + has emerged as one of the central mediators ininflammation and immunity [10]. For this reason, the reg-ulation of IFN- + biosynthesis has been of great interest,and the identification of IL-12 as a potent IFN- + -inducing


Table 3. Effect of DC culture supernatant or IL-18 on IFN- +production by GM-CSF−/− CD8+ T cellsa)

IFN- + levels (U/ml) in supernatantfrom GM-CSF−/− CD8+ T cells

cultured with:

Addition of: 1000 U/ml IL-2 100 pg/ml IL-12

None 24.1 22.9

GM-CSF and IL-4 34.0 38.2

DC culture supernatant 90.2 97.8

IL-18 24.1 111.1

a) CD8+ T cells were isolated from five LPS-injected GM-CSF−/− mice and cultured with: (1) 20 ng IL-18 plus IL-2or IL-12, and (2) 50 ? l supernatant plus IL-2 or IL-12. Forcontrol purposes, the effect of 2.5 ng/ml GM-CSF and0.5 ng/ml IL-4 on IFN- + production was tested.

factor has made it possible to define more precisely themechanism of IFN- + production in innate and acquiredimmunity [11]. LPS is a potent activator of IL-12 produc-tion by macrophages, and IL-12 in turn activates T cellsand NK cells, the two major cellular sources of IFN- + , toproduce this cytokine. For IFN- + production by T cells,IL-2 is a major cofactor with IL-12 [12], and TNF is amajor IL-12 cofactor for IFN- + production by NK cells[13]. IL-10, a cytokine with anti-inflammatory activitymediated in part by down-regulating IL-12 and IFN- +production, is also a product of LPS-activated macro-phages and is concomitantly produced with IL-12 [14]. Inthe present report, we confirm the observation of Basu etal. [6] that GM-CSF−/− mice produce low levels of serumIFN- + in response to LPS. Injection of GM-CSF restoredthe serum levels of IFN- + , providing direct evidence foran immediate action of GM-CSF rather than an alteredGM-CSF-dependent physiological state being responsi-ble for the low serum levels of IFN- + in LPS-injected GM-CSF−/− mice. In vitro studies demonstrated that a T celldeficiency rather than a defect in NK cells was the basisfor the low IFN- + production in GM-CSF−/− mice, andmAb depletion tests confirmed that the ability of NK cellsto produce IFN- + is intact in GM-CSF−/− mice. Sincereceptors for GM-CSF are not expressed on T cells [15],the biological effect of GM-CSF on T cells must be indi-rect. Evidence in support of this idea comes from thefinding that in vivo, but not in vitro, supplementation withGM-CSF restores IFN- + production by T cells from LPS-injected GM-CSF−/− mice.

T cells from LPS-injected GM-CSF−/− mice showedpoor IFN- + production and impaired proliferation inresponse to IL-2 and IL-12. In contrast, IFN- + productionand the proliferative response of T cells from naive GM-

CSF−/− and +/+ mice were comparable, and IL-12induced similar levels of serum IFN- + in GM-CSF−/− and+/+ mice. Based on these findings, it seems reasonableto conclude that T cells from GM-CSF−/− mice areendowed with normal basal responsiveness to IL-2 andIL-12, but that LPS induces an inhibitory state that ren-ders T cells less responsive to these cytokines, resultingin the low level of serum IFN- + in LPS-injected GM-CSF−/− mice. With regard to the sensitivity of T cells toIL-2 and IL-12, we have shown by flow cytometric analy-sis that the § - and g -chains of the IL-2 receptors are up-regulated in a comparable manner in T cells from LPS-injected GM-CSF−/− and GM-CSF+/+ mice after a 12-hculture period with IL-2. In addition, Jak2 associatedwith the IL-12 receptor [16] is normally phosphorylated inT cells from LPS-injected GM-CSF−/− mice after culturewith IL-12 (unpublished data). These findings give sup-port to the idea that T cells from LPS-injected GM-CSF−/− mice can respond to IL-2 and IL-12 signaling,but that full responsiveness is compromised because ofa lack protective or stimulatory signals indirectly deliv-ered by GM-CSF. This suggests the existence of a factoror factors induced by GM-CSF which up-regulatesresponsiveness of T cells to IL-2 and IL-12. The fact thatDC culture supernatant restores IFN- + production byT cells from LPS-injected GM-CSF−/− mice supportsthis possibility.

DC are well known as potent APC with high migratorycapacity and high expression of MHC class II and criticalco-stimulatory molecules [17]. DC have been reported toproduce several immunomodulating factors. Macatoniaet al. [18] found that mouse CD11c+ DC produced IL-12,thus fostering antigen-specific T cells along the Th1pathway. Human CD1a+ DC have been reported to pro-duce a soluble factor that stimulates growth of CD40-activated B cells and enhances Ig production [19].Recently, a chemokine named DC-DK-1 derived from DChas been cloned [20]. DC-DK-1 is chemotactic for naiveT cells, suggesting involvement of this chemokine in theinitiation of immune responses. These findings indicatethat DC exert their effects on T or B cells via soluble fac-tors as well as through membrane-bound molecules. Wehave recently described another factor of DC originwhich we designated DC-derived factor or DC-DF [7].DC-DF is the product of GM-CSF-activated DC and hasthe characteristics of normalizing IFN- + production andproliferative activity of antigen-specific CD4+ T cells fromGM-CSF−/− mice. In the case of this antigen-specificdefect in GM-CSF−/− CD4+ T cells, the T cells producehigh levels of IL-2 in spite of low IFN- + production andlow proliferative responses, indicating that diminishedresponsiveness to IL-2 may be the main cause of ineffi-cient clonal expansion of antigen-specific T cells in GM-CSF−/− mice. The fact that DC-DF corrects this defi-


ciency suggests that GM-CSF exerts its effect on T cellsvia DC-DF, and that DC-DF is involved in the cognateinteraction between T cells and antigen processing cells,capacitating T cells to utilize IL-2 and promoting theirclonal expansion. The absence of GM-CSF-induced DC-DF disorders the balance between the positive and neg-ative effects exerted on T cells by antigen stimulation.The DC-derived factor identified in the present study hascomparable normalizing activity on CD8+ T cells fromLPS-injected GM-CSF−/− mice. Whether these twoactivities in DC supernatant, which we designate DC-DF(CD4) and DC-DF (CD8), are attributable to the same fac-tor awaits cloning and molecular characterization.

Because of the IFN- + -inducing properties of IL-18, arecently cloned cytokine [21], we analyzed its possiblerole in the IFN- + abnormality seen in GM-CSF−/− miceand its relation to the DC factor or factors described byus [7]. IL-18 has been shown to be synergistic with IL-12but not with IL-2 in up-regulating IFN- + production [22].This selective synergy of IL-18 with IL-12 was confirmedin the present study and distinguishes IL-18 from DC-DF(CD8) which synergized with IL-2 as well as with IL-12.The finding that in vivo injection of DC-DF (CD8) but notIL-18 abrogates the inhibitory effects induced by LPS onproliferative responses and IFN- + production by T cellsalso indicates that factors other than IL-18 are responsi-ble for the up-regulation of IFN- + production in LPS-injected GM-CSF−/− mice. Since IL-1 § , - g , IL-10 andIL-12 p40 are detectable in supernatants from DC cul-tures, it cannot be excluded that a combination of fac-tors, including IL-18, rather than a single factor mediatesthe effects of DC-DF (CD8).

4 Materials and methods

4.1 Mice

GM-CSF−/− mice on a C57BL/6 × 129 background weregenerated at the Melbourne Branch of The Ludwig Institutefor Cancer Research [4] and breeding stocks were trans-ferred to the New York Branch. (C57BL/6 × 129)F2 micewere used as controls for GM-CSF−/− mice.

4.2 mAb

Biotinylated mAb were purchased from Pharmingen(San Diego, CA). Anti-L3T4 (CD4) and anti-Lyt-2.2 (CD8)mAb were provided by Drs. F. Fitch (University of Chicago)and U. Hammerling (Memorial Sloan-Kettering Cancer Cen-ter), respectively. Experiments with T cell-immunodepletedmice were performed 1 week after injection of 50 ? l anti-L3T4 and 25 ? l anti-Lyt-2.2 mAb i.v. in a total volume of200 ? l [23].

4.3 Cytokines

Recombinant mouse GM-CSF and recombinant mouse IL-4were purchased from Pharmingen. Recombinant humanIL-2 and recombinant mouse IL-18 were purchased fromGenzyme (Cambridge, MA) and Peprotech (Rocky Hill, NJ),respectively. Recombinant mouse IL-12 was a kind gift of Dr.S. Wolf at Genetics Institute (Cambridge, MA). GM-CSF,IL-18 and IL-12 were diluted in PBS (containing 1 % synge-neic mouse serum in the case of IL-12) and injected i.p. in atotal volume of 200 ? l.

4.4 LPS

Hundred micrograms LPS (from E. coli; serotype O127:B8,Sigma, St. Louis, MO) were dispersed in PBS and injectedi.p. in a total volume of 200 ? l. Three hours after LPS injec-tion, spleens were removed to prepare cell suspensions.Additional mice were bled from the retroorbital plexus at var-ious times after LPS injection for serum harvest and cytokineassays.

4.5 Cytokine assays

Serum levels of cytokines were measured by ELISA withreagents from the following sources: IFN- + , TNF- § , IL-12p70 (Genzyme); GM-CSF, IL-1 g , IL-6, IL-10, TGF- g 1 (R&D,Minneapolis, MN); IL-12 p40 (Biosource International,Camarillo, CA). Culture supernatants were tested for IFN- +by ELISA using reagents from Pharmingen.

4.6 Isolation and culture of T cells and CD16/32+ cells

To isolate defined cell populations, spleen cells were incu-bated with biotinylated mAb against mouse CD4, CD8 orCD16/32+ at 4 °C for 30 min. Cells were washed and incu-bated at 4 °C for 30 min with avidin-coated magnetic beads(Dynal, Lake Success, NY). Cells were pelleted magneticallyand washed. Purity of isolated cells ( G 90 %) was confirmedby flow cytometry. CD4+ or CD8+ T cells (1 × 105) or 2 × 105

CD16/32+ cells per well were cultured in complete medium(RPMI with 10 % FBS, 5 × 105 M 2-ME and glutamine) atvarious concentrations of IL-2 and IL-12 in 96-well plates ata total volume of 200 ? l for 3 days at 37 °C in a 5 % CO2

atmosphere. For proliferation assays, 2 × 105 CD8+ T cellswere cultured with IL-2 alone or IL-2 and IL-12 for 3 days at37 °C in a 5 % CO2 atmosphere, and the proliferativeresponse was determined by uptake of methyl-[3H]thymidine.

4.7 Isolation of DC and generation of DC-DF

DC were isolated from spleen cells of LPS-injected (C57BL/6 × 129)F2 mice using biotinylated anti-mouse CD11c mAband avidin-coated magnetic beads. CD11c+ cells (5 × 105


per well) were cultured in serum-free medium (Sigma) with10 ng/ml GM-CSF and 2 ng/ml IL-4 in 96-well plates for 4days at 37 °C in a 5 % CO2 atmosphere. Culture superna-tants (50 ? l) were transferred to cultures of CD8+ T cells fromLPS-injected GM-CSF−/− mice with IL-2 and/or IL-12 at atotal volume of 200 ? l for 3 days at 37 °C in a 5 % CO2 atmo-sphere. For proliferation assays, isolated CD8+ T cells werecultured with 100 U/ml IL-2 and 50 ? l culture supernatant ina total volume of 200 ? l for 3 days. For in vivo studies, 0.5 mlDC supernatants were injected i.p.

5 References

1 Metcalf, D., The molecular biology and functions of thegranulocyte-macrophage colony-stimulating factors.Blood 1986. 67: 257–267.

2 Moore, M. A. S., The clinical use of colony stimulatingfactors. Annu. Rev. Immunol. 1991. 9: 59–191.

3 Witmer-Pack, M. D., Oliver, W., Valinsky, J., Shuler, G.and Steinman, R. M., Granulocyte/macrophage colony-stimulating factor is essential for the viability and func-tion of cultured murine epidermal Langerhans cells. J.Exp. Med. 1987. 166: 1484–1498.

4 Stanley, E., Lieschke, G. J., Grail, D., Metcalf, D.,Hodgson, G., Gall, J. A., Maher, D. W., Cebon, J.,Sinickas, V. and Dunn, A. R., Granulocyte/macrophagecolony-stimulating factor-deficient mice show no majorperturbation of hematopoiesis but develop a character-istic pulmonary pathology. Proc. Natl. Acad. Sci. USA1994. 91: 5592–5596.

5 Dranoff, G., Crawford, A. D., Sadelain, M., Ream, B.,Rashid, A., Bronson, R. T., Dickersin, G. R.,Bachurski, C. J., Mark, E. L., Whitsett, J. A. and Mulli-gan, R. C., Involvement of granulocyte-macrophagecolony-stimulating factor in pulmonary homeostasis.Science 1994. 264: 713–716.

6 Basu, S., Dunn, A. R., Marino, M. W., Savoia, H.,Hodgson, G., Lieschke, G. J. and Cebon, J., Increasedtolerance to endotoxin by granulocyte-macrophagecolony-stimulating factor-deficient mice. J. Immunol.1997. 159: 1412–1417.

7 Wada, H., Noguchi, Y., Marino, M. W., Dunn, A. R. andOld, L. J., T cell functions in granulocyte/macrophagecolony-stimulating factor deficient mice. Proc. Natl.Acad. Sci. USA 1997. 94: 12557–12561.

8 Magram, J., Connaughton, S. E., Warrier, R. R., Car-vajal, D. M., Wu, C. Y., Ferrante, J., Stewart, C., Sar-miento, U., Faherty, D. A. and Gately, M. K., IL-12-deficient mice are defective in IFN gamma productionand Type 1 cytokine responses. Immunity 1996. 4:471–481.

9 Wheelock, E. F., Interferon-like virus-inhibitor induced inhuman leukocytes by phytohemagglutinin. Science1965. 149: 310–311.

10 Farrar, M. A. and Schreiber, R. D., The molecular andcell biology of interferon- + and its receptor. Annu. Rev.Immunol. 1993. 11: 571–611.

11 Kobayashi, M., Fitz, L., Ryan, M., Herwick, R. M.,Clark, S. C., Chan, S., Loudon, R., Sherman, F., Perus-sia, B. and Trinchieri, G., Identification and purificationof natural killer cell stimulatory factor (NKSF), a cytokinewith multiple biologic effects on human lymphocytes. J.Exp. Med. 1989. 170: 827–845.

12 Trinchieri, G., Interleukin-12: a proinflammatory cyto-kine with immunoregulatory functions that bridge innateresistance and antigen-specific adaptive immunity.Annu. Rev. Immunol. 1995. 13: 251–276.

13 Tripp, C. S., Wolf, S. F. and Unanue, E. R., Interleukin12 and tumor necrosis factor- § are costimulators ofinterferon- + production by natural killer cells in severecombined immunodeficiency mice with listeriosis, andinterleukin-10 is a physiologic antagonist. Proc. Natl.Acad. Sci. USA 1993. 90: 3725–3729.

14 Berg, D. J., Kuhn, R., Rajewsky, K., Muller, W.,Menon, S., Davidson, N., Grunig, G. and Rennick, D.,Interleukin-10 is a central regulator of the response toLPS in murine models of endotoxin shock and theSchwartzman reaction but not endotoxin tolerance. J.Clin. Invest. 1995. 96: 2339–2347.

15 Park, L. S., Friend, D., Gillis, S. and Urdal, D. L., Char-acterization of the cell surface receptor for human gran-ulocyte/macrophage colony-stimulating factor. J. Exp.Med. 1986. 164: 251–262.

16 Bacon, C. M., McVicar, D. W., Ortaldo, J. R., Rees, R.C., O’Shea, J. J. and Johnston, J. A., Interleukin 12 (IL-12) induces tyrosine phosphorylation of Jak2 and Tyk2:Differential use of janus family tyrosine kinases by IL-2and IL-12. J. Exp. Med. 1995. 181: 399–404.

17 Steinman, R. M., The dendritic cell system and its role inimmunogenecity. Annu. Rev. Immunol. 1991. 9: 271–296.

18 Macatonia, S. E., Hosken, N. A., Litton, M., Vieira, P.,Hsieh, C. S., Culpepper, J. A., Wysocka, M., Trinchieri,G., Murphy, K. M. and O’Garra, A., Dendritic cells pro-duce IL-12 and direct the development of Th1 cells fromnaive CD4+ T cells. J. Immunol. 1995. 154: 5071–5079.

19 Dubois, B., Vanbervliet, B., Fayette, J., Massacrier,C., Van Kooten, C., Briere, F., Banchereau, J. andCaux, C., Dendritic cells enhance growth and differenti-ation of CD40-activated B lymphocytes. J. Exp. Med.1997. 185: 941–951.

20 Adema, G. J., Hartgers, F., Verstraten, R., de Vries, E.,Marland, G., Menon, S., Foster, J., Xu, Y., Nooyen, P.,McClanahan, T., Bacon, K. B. and Figdor, C. G., A


dendritic-cell-derived C-C chemokine that preferentiallyattracts naive T cells. Nature 1997. 387: 713–717.

21 Okamura, H., Tsutui, H., Komatsu, T., Yutsudo, M.,Hakura, A., Tanimoto, T., Torigoe, K., Okura, T.,Nukada, Y., Hattori, K., Akita, K., Namba, M., Tanabe,F., Konishi, K., Fukuda, S. and Kurimoto, M., Cloningof a new cytokine that induces IFN- + production byT cells. Nature 1995. 378: 88–91.

22 Ahn, H.-G., Maruo, S., Tomura, M., Mu, J., Hamaoka,T., Nakanishi, K., Clark, S., Kurimoto, M., Okamura,H. and Fujiwara, H., A mechanism underlying synergybetween IL-12 and IFN- + -inducing factor in enhancedproduction of IFN- + . J. Immunol. 1997. 159: 2125–2131.

23 Nakayama, E. and Uenaka, A., Effect of in vivo admin-istration of Lyt antibodies. Lyt phenotype of T cells inlymphoid tissue and blocking of tumor rejection. J. Exp.Med. 1985. 161: 345–355.

Correspondence: Lloyd J. Old, Ludwig Institute for CancerResearch, New York Branch at Memorial Sloan-KetteringCancer Center, 1275 York Avenue, New York, NY 10021,USAFax: +1-2 12-7 17 31 00e-mail: LOld — LICR.org


Documents

Regulation of IFN-γ production in granulocyte-macrophage colony-stimulating factor-deficient mice