Mycobacterial infection in natural killer T cell knockout mice

Tuberculosis (2002) 82(2/3), 97--104 & 2002 Elsevier Science Ltd. All rights reserved doi:10.1054/tube.331, available online at http://www.idealibrary.com on

Mycobacterial infection in natural killer T cell knockout mice I. Sugawara,* H.Yamada,* S. Mizuno,* C.Y. Li,* T. Nakayama,w M.Taniguchiw *Department of Molecular Pathology,The Research Institute of Tuberculosis, 3-1-24 Matsuyama, Kiyose,Tokyo 204-0022, Japan w Department of Immunology, Chiba University, School of Medicine, Chiba 260, Japan

Summary To gain a better understanding of the pathological role of natural killer (NK) T cells in murine tuberculosis, NKT knockout (KO) mice (Ja281-/-mice) were utilized. Eight-week-old NKT KO mice of BALB/c origin and wild-type (WT) mice were infected with Mycobacterium tuberculosis Kurono strain by the airborne route using an airborne infection apparatus, and their capacity to control mycobacterial growth, granuloma formation, and cytokine production were examined.The NKT KO mice developed granulomatous lesions in the lungs; there was no statistically significant difference in pulmonary granuloma size between NKT KO and WT mice (po0.01).The average CFU valuesincreased 3 weekspost-infection, but decreased 9 and11weeks post-infection, in the lungs of NKT KO mice.When stimulated with Kurono strain in vitro, splenic cells from NKT KO mice produced less IFN-g than did those from WT mice. Expression of mRNA for IL-2, IL-4, IL-6, IL-10 and IL-12 p40 was slightly lower in NKT KO mice compared with WT mice. Our data indicate that NKT cells play a detrimental role in late-phase mycobacterial infection, althoughTh1cells are essential in early-phase mycobacterial infection. & 2002 Elsevier Science Ltd. All rights reserved

INTRODUCTION Murine natural killer (NK) T cells belong to a/b T cells and express NK-associated cell surface markers (NK1.1) and some members of the Ly-49 family.1,2 NKT cells are different in that they use an invariant T cell antigen receptor a chain consisting of rearranged Va14 and Ja281 segments and a conserved CDR3.3 Va14 NKT cells produce IFN-g and IL-4 on stimulation with their specific ligand, a-galactosylceramide.4 The ligand-activated Va14 NKT cells directly kill various target tumor cells by an NK-like mechanism after direct contact with the target cells.5 Moreover, it has been reported that decidual NKT cells are involved in abortion.6 On the other hand, the in vivo functions of NKT cells in infectious diseases are not fully described. NKT cells are involved in the IgG response to the GPI-anchored proteins of Plasmodium and Trypanosoma.7 It has recently been reported that murine NKT cells contribute to the granulomatous reaction caused by mycobacterial cell walls,8 but little is known about the function of NKT cells when mice are infected aerially with M. tuberculosis. Therefore, it is very important to clarify the physiologic role of NKT cells in the development of protective Correspondence to : Isamu Sugawara, Department of Molecular Pathology, The Research Institute of Tuberculosis, 3-1-24 Matsuyama, Kiyose, Tokyo 204-0022, Japan.Tel.:+81-424-93-5075; fax: +81-424-92-4600; e-mail: [email protected]

immunity against murine tuberculosis. We utilized NKT KO mice for this purpose. Using these NKT KO mice, we studied the role of NKT cells in M. tuberculosis-induced experimental tuberculosis.

MATERIALS AND METHODS

Animals BALB/c wild-type (WT) and BALB/c Ja281-/- mice were used.4 The former mice were purchased from Japan SLC Co. Ltd., Shizuoka, Japan, and the latter mice were obtained by backcrossing 129/Sv- nine times onto the BALB/c background. Genomic characteristics were confirmed by PCR using genomic DNA and BALB/c WT and Ja281-/- mouse genome-specific primers (Fig. 1). All the mice were housed in a biohazard level 3 facility and given mouse chow and water ad libitum after aerosol infection with M. tuberculosis.

Screening of NKT KO (Ja281-/-) mice by polymerase chain reaction (PCR) For demonstration of Ja281-/- mice, PCR using Ja281specific primer sets was performed. The 1.8 kb PCR product using primer sets (CJ-2 and IM-neo2) indicated Ja281 mutants and the 1.5 kb PCR product using primer sets (CJ-2 and CJ-4) indicated WT mice (Fig. 1). The primer sequences were as follows: CJ-2

97

98

Sugawara et al.

used to evaluate in vivo growth of Mycobacteria. The lung and spleen tissues were homogenized with a mortar and pestle, and then 1ml of sterile saline was added. After homogenization, 100 ml of homogenate was taken and plated by 10-fold serial dilution on 1% Ogawa’s egg solid media. Colony numbers on the media were counted after incubation for 4 weeks at 371C. All CFU values are expressed as log10 CFU per whole organ.9,10

RT-PCR Fig. 1 Distinction betweenWTand NKT KO mice by PCR.Primer sets indicating disruption of the Ja281gene were utilized in PCR analysis. NKT KO mice showed a1.8 kb band, but did not show a1.5 kb band.The mice lacking the1.5 kb gene amplification were used in this study as NKT KO mice.

50 -AAAAGAGAGGATTCAGATAGCCCAGGAG-30 , CJ-4 50 TCTACTGCTTTTGATTCTGCCTTTGGTGTC-30 and IMneo2 50 -CGGTATCGCCGCTCCCGATTAGCAGCGCAT-30 . The samples were subjected to 35 cycles of denaturation (941C for 1 min), annealing (621C for 1 min), and extension (721C for 2 min).

Experimental infection A M. tuberculosis Kurono strain (ATCC 358121) passaged in a C57BL/6 mouse was grown in Middlebrook 7H9 broth for 2 weeks, then filtered with Sterile Acrodisc Syringe Filters (Pall Corporation, Ann Arbor, MI, USA) with pore size 5.0 mm. Then, aliquots of the filtered bacterial solution were stored in a freezer at 801C until use. Mice were infected via the airborne route by placing them into the exposure chamber of a Glas-Col aerosol generator (Glas-Col, Terre Haute, IN, USA).9,10 The nebulizer compartment was filled with 5 ml of a suspension of 106 CFU of Kurono strain, such that approximately 100 tubercle bacilli were deposited in the lung of each animal. The 10 pulmonary granuloma sizes of NKT KO and WT mice were measured with a micrometer (Nikon Optical Co., Tokyo, Japan).

CFU assay At 1, 3, 5, 7, 9, and 11 weeks after infection, three mice of both groups (except for 1 week, when two mice were used in each group) were anesthetized with pentobarbital sodium, the abdominal cavity was incised, and exsanguination was performed by splenectomy and transection of the left renal artery and vein. The lungs, spleen and liver were excised and weighed. The left lobe of the lung and a part of the spleen tissue were separately weighed and Tuberculosis (2002) 82(2/3), 97--104

The right lower lobe of the lung and the portion of spleen tissue left after CFU examination were used for RT-PCR analysis of expression of several cytokine mRNAs in these organs during M. tuberculosis infection. These tissue specimens were frozen in liquid nitrogen immediately after excision and weighing, and stored at 851C until use. RNA extraction was performed as described previously.11 Briefly, frozen tissues were thawed and homogenized in a 1.5 ml microcentrifuge tube with a tipblunted 1 ml pipet tip in liquid nitrogen. Then, the homogenates were treated with the total RNA isolation reagent TRIzol (GIBCO BRL), according to the manufacturer’s instructions. After isolation, each total RNA was reverse transcribed into cDNA with M-MLV reverse transcriptase (GIBCO BRL) following measurement of total RNA concentration and agarose gel electrophoresis. Polymerase chain reaction was performed with equivalent amounts of cDNA of each sample (5 ml), TaKaRa EX Taq (containing 10  buffer and dNTP mixture; Otsu, Shiga, Japan) and gene-specific primer sets for b-actin, IFN-g, TNF-a, interleukin (IL)-1b, IL-2, IL-6, IL-10, IL-12 p40, TGF-b, and iNOS. Primer sequences, annealing temperature and cycle number for amplification are listed in Table 1. Amplifications were performed with a DNA thermal cycler model 480 (Perkin-Elmer Cetus). Ten microliters of each PCR product was applied to electrophoresis in 4% agarose and NuSieve GTG (1:3) gel and visualized using ethidium bromide staining. The same amounts of b-actin RNA from the lung tissues were used as an internal control in the RT-PCR analysis. We made a densitometric analysis of electrophoretic RT-PCR using NIH image ver. 1.62. Relative densitometric ratios of various cytokines and iNOS mRNA to b-actin mRNA as an internal control were determined.

Light and electron microscopic examination For light microscopic examination, the right upper lung lobe of each animal was excised and fixed with 20% formalin neutral buffer–methanol solution (8% formaldehyde and 20% methanol; Mildform 20NM, Wako Pure Chemical Co., Osaka, Japan), dehydrated with graded

& 2002 Elsevier Science Ltd. All rights reserved

NKTand murine tuberculosis

99

Table 1 Primer sets used in this study Target

Sequence*

AnnealingT (C1)

Cycles Number

Product Lengthw

b-actin

S: 50 -TGT GAT GGT GGG AAT GGG TCA G-30 A: 50 -TTT GAT GTC ACG CAC GAT TTC C-30

65

23

514

IFN-g

S: 50 -TAC TGC CAC GGC ACA GTC ATT GAA-30 A: 50 -GCA GCG ACT CCT TTT CCG CTT CCT-30

65

30

405

TNF-a

S: 50 -ATG AGC ACA GAA AGC ATG ATC-30 A: 50 -TAC AGG CTT GTC ACT CGA ATT-30

65

30

276

IL-1b

S: 5’-CAG GAT GAG GAC ATG AGC ACC-30 A: 50 -CTC TGC AGA CTC AAA CTC CAC-30

65

30

447

IL-2

S: 50 -CTT CAA GCT CCA CTT CAA GCT-30 A: 50 -CCATCT CCT CAG AAA GTC CAC-30

65

40

400

IL-4

S: 50 -ACG GAG ATG GAT GTG CCA AAC GTC-30 A: 5’-CGA GTA ATC CAT TTG CAT GAT GC-30

65

40

279

IL-6

S: 50 -CAT CCA GTT GCC TTC TTG GGA-30 A: 30 -CAT TGG GAA ATT GGG GTA GGA AG-30

65

40

463

IL-10

S: 5-GTG AAG ACT TTC TTT CAA ACA AAG -30 A: 50 -CTG CTC CAC TGC CTT GCT CTTATT-30

56

40

274

IL-12 P40

S: 50 -ATC TCC TGG TTT GCC ATC GTT TTG-30 A: 50 -TCC CTT TGG TCC AGT GTG ACC TTC-30

65

30

527

TGF-b

S: 50 -CGG GGC GAC CTG GGC ACC ATC CAT GAC-30 A: 50 -CTG CTC CAC CTT GGG CTT GCG ACC CAC-30

65

24

371

iNOS

S: 50 -TGG GAATGG AGA CTG TCC CAG-30 A: 50 -GGG ATC TGA ATG TGATGT TTG-30

65

30

306

*S = sense, A = antisense. Product lengths were indicated by base pairs.

w

series of ethanol, treated with xylene, and embedded in paraffin. Five micrometer thick sections were obtained from each paraffin block and stained with either hematoxylin and eosin or Ziehl–Neelsen for acid-fast bacilli. For electron microscopy, the right middle lobe of each animal’s lung was fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4 (PB) at 41C overnight, washed three times with cold PB, post-fixed with 1% osmium tetroxide in PB at 41C for 1 h, dehydrated with graded series of acetone and finally embedded in Spurr’s low-viscosity resin. Spurr’s resin polymerization was done at 701C for 16 h in an oven. Ultra-thin sections were obtained with a Reichert Ultracut apparatus and stained with uranyl acetate and Sato’s lead solution. Examination was performed with a JEOL JEM-1230 apparatus (JEOL, Tokyo, Japan).12

Cytokine assay To examine the cytokine secretion profiles of splenic cells of NKT KO mice stimulated with M. tuberculosis Kurono strain in vitro for 3 days, levels of IFN-g, TNF-a, IL-4, IL-6, IL-12, and IL-1 secretion were measured by enzymelinked immunosorbent assay (ELISA) (Biosource International, CA, USA). Single splenic cell suspensions were prepared from WT and NKT KO mice. The cell suspensions were plated (5  105 cells/well) in the presence of

& 2002 Elsevier Science Ltd. All rights reserved

Kurono strain (m.o.i.; 10:1) in 96-well flat-bottomed microculture plates and cultured for 3 days at 371C in CO2 in air. Various cytokines in the supernatants obtained 3 days later were measured by ELISA.

Statistical analysis Comparison of CFU values was performed with a randomization test for two independent samples.

RESULTS

Mycobacterial burden in the lung and spleen of infected mice BALB/c WT and NKT KO ( Ja281-/-) mice, 20 mice each, were aerially infected twice with the Kurono strain of M. tuberculosis at approximately 100 CFU per lung. At 1 week after infection, tubercle bacilli could be recovered from the lungs of both groups of mice; the average log10 CFUs were 2.6 in WT and 1.6 in NKT KO mice. At 3 weeks after infection, the average log10 CFU values of both lung and spleen in NKT KO mice were higher than in WT mice (6.1 vs. 5.2 in lung, 4.1 vs. 3.3 in spleen). However, these average CFU values of NKT KO mice were the highest observed throughout the experimental period, because at 5 weeks post-infection, average CFU values of NKT KO mice, both in lung and spleen, had declined slightly. In Tuberculosis (2002) 82(2/3), 97--104

100

Sugawara et al.

Fig. 2 CFU in lung and spleen tissues of WTand NKT KO mice infected aerially with Kurono strain. }- -, lung tissue from WT mice; ^--, lung tissues from NKT KO mice; &- - -, spleen tissues from WT mice and &--, spleen tissues from NKT KO mice.Tubercle bacilli are expressed as log10(CFU/whole organ).

Fig. 3 Histologic examination ofinfectedlung tissues of NKTand WT mice at 3 and11weeks afteraerialinfectionwith Kurono strain.Hematoxylin and eosin stain, 100. (A) NKT KO mouse 3 weeks post-infection, (B) WT mouse 3 weeks post-infection, (C) NKT KO mouse11weeks post-infection, and (D) WT mouse11weeks post-infection.

Tuberculosis (2002) 82(2/3), 97--104

& 2002 Elsevier Science Ltd. All rights reserved

NKTand murine tuberculosis 101

Fig. 4 Electron micrographs of lung tissue infected with Kurono strain, obtained at necropsy (7 weeks after infection). (A) NKT KO mouse. A tubercle bacillus is present in the vacuolated phagosome. 6500. (B) WT mouse.Tubercle bacilli are present in several phagosomes. 6500.

particular, the average lung CFU of NKT KO mice at week 11 was almost similar to their spleen CFU. On the contrary, lung and spleen CFU values of WT mice increased gradually up to 11 weeks (Fig. 2). The CFU values of NKT KO mice were significantly lower than those of WT mice at 7 weeks in the spleen (4.3 vs. 3.9, po0.025) and at 11 weeks in the lung (5.7 vs. 3.9, po0.05). Throughout the experimental period, no mice of either group succumbed to progressive mycobacterial infection.

Light and electron microscopic observation of infected lung In agreement with the CFU observations, histopathological findings obtained from BALB/c WT and NKT KO mice did not show clear differences. At 7 weeks and 11 weeks after infection, when the CFU values of NKT KO mice were significantly lower than those of WT mice, the average size of the 10 pulmonary granulomatous lesions were not different (5450+380 mm vs. 5625+520 mm at 9 weeks after infection), and there were only a small number of acid-fast tubercle bacilli in the NKT KO mice (Fig. 3). Similarly, electron microscopic observation did not reveal prominent morphological differences between two groups of mice. Several tubercle bacilli were recognized in the phagosomes of epithelioid macrophages (Fig. 4).

& 2002 Elsevier Science Ltd. All rights reserved

Table 2 Cytokine secretion by spleen cells of NKT KO mice and WT mice stimulated by M. tuberculosis Kuruno strain Cytokine

IL-1 IL-4 IL-6 IL-12 IFN-g TNF-a

Amount of cytokine secreted (pg/ml) NKT KO

WT

20+2 30+2 50+5 205+15 255+20 710+65

25+3 25+2 60+5 195+10 470+30 815+75

Note: Values are means+SD. Spleen cells from WTand NKT KO mice (5 105/well) were cultured in 96-well flat-bottomed microculture plates in the presence of Kurono strain (m.o.i.; 10:1) at 371C for 3 days. Various cytokines in the culture supernatants obtained 3 days later were assayed by a sandwich ELISA.

RT-PCR analysis In addition to bacterial growth in organs and light and electron microscopic observation, we compared by RTPCR the expression of the mRNA for several cytokines in lung tissues of both groups of animals. As shown in Fig. 5, no statistically significant differences were observed in cytokine mRNA expression between the groups in terms of densitometric analysis (po0.01). IFN-g and IL-4 mRNA expression at 7 weeks post-infection in NKT KO mice was slightly higher than that in WT mice. IL-2, IL-10 and IL-12 p40 mRNA expression at 1–9 weeks post-infection, IL-6 Tuberculosis (2002) 82(2/3), 97--104

102

Sugawara et al.

Fig. 5 Densitometric analysis of in vivo expression of mRNA for various cytokines, iNOS and TGF-b in Kurono strain-infected mice. Relative densitometric ratios of cytokine, iNOS and TGF-b mRNA to b-actin mRNAwere determined by using NIH image ver.1.62.*-,WT mice; *- -, NKT KO mice.

Tuberculosis (2002) 82(2/3), 97--104

& 2002 Elsevier Science Ltd. All rights reserved

NKTand murine tuberculosis 103

mRNA expression at 1–7 weeks post-infection and IL-4 mRNA expression in NKT KO mice was slightly lower than those of WT mice.

Cytokine assay There were no statistically significant differences in secretion of IL-1, IL-4, IL-6, IL-12, and TNF-a between WT and NKT KO mice (po0.01) when the spleen cells from each group were stimulated in vitro with Kurono strain for 3 days. The spleen cells from NKT KO mice secreted less IFN-g than those from WT mice (po0.01) (Table 2). DISCUSSION We have found that pulmonary granulomas are induced by inoculation of an M. tuberculosis strain, but that expression levels of IL-2, IL-4, IL-6, IL-10 and IL-12 mRNA are slightly low in late-phase mycobacterial infection in NKT KO mice. However, it has been reported that a granuloma-like lesion does not occur in NKT KO mice when they are injected subcutaneously with deproteinized mycobacterial cell walls.8 This discrepancy between their data and ours may be explained as follows. Firstly, we used live M. tuberculosis to induce granulomas, whereas they used deproteinized cell walls of H37Rv. The deproteinized cell walls may not induce granulomas. Secondly, we observed chronic pulmonary lesions from 1 week until 11 weeks after infection, but they observed acute skin and muscle lesions at 1, 2, 3 and 7 days after subcutaneous injection. A duration of 7 days may be too short to evaluate granuloma-like lesions. Lastly, we used NKT KO mice with a BALB/c origin, whereas they used NKT KO mice with a C57BL/c origin. There may be strain differences in the ability to form granulomas. We also examined by PCR the existence of the Ja281 gene in the NKT KO mice used in this study; a 1.8 kb gene fragment was recognized in these mice, demonstrating their lack of the Ja281 gene (Fig. 1). Splenic cells from NKT KO mice stimulated with M. tuberculosis Kurono strain secreted less IFN-g compared with those from WT mice. As it has been reported that NKT cells secrete IFN-g upon stimulation with M. bovis BCG,13,14 the reduced IFN-g production in early-phase mycobacterial infection can be attributed to the lack of NKT cells; IFN-g is produced only by Th1 and NK cells in NKT KO mice. Th1 and NK cells may be important in early-phase mycobacterial infection. The average CFU values increased 3 weeks after infection, but decreased 9 and 11 weeks after infection, in the lungs of NKT KO mice. As the average CFU values increased 9 and 11 weeks after infection in the lungs of WT mice, NKT cells may play a detrimental role in latephase mycobacterial infection. Further study will be

& 2002 Elsevier Science Ltd. All rights reserved

required to elucidate a role of NKT cells in mycobacterial infection. Expression of mRNA for IL-6 and IL-12 was low in latephase mycobacterial infection in NKT KO mice. Although IL-6 and IL-12 are produced by macrophages, alveolar macrophages in late murine tuberculosis may secrete less IL-6 and IL-12. We have previously reported that expression of IL-12 mRNA is low in nuclear factor (NF)-IL6 KO mice.10 In fact, expression of NF-IL6 mRNA was low in NKT KO mice (data not shown). It has been reported that lack of IFN-g and TNF-a leads to lethal disseminated necrotic lesions.15–19 However, lack of IL-1, IL-4 and IL-18 does not reduce protective immunity against mycobacterial infection.9,11,20 As IFN-g and TNF-a expression is maintained from 1 week until 11 weeks in NKT KO mice, this explains induction of granuloma formation as a defence mechanism against invasion of M. tuberculosis. mRNA expression of IL-2, a Th1 cytokine, and IL-4 and IL-10, both Th2 cytokines, is also slightly low in NKT KO mice. As various cytokines including Th1 and Th2 cytokines are expressed to a varying degree in NKT KO mice, it is difficult to clarify a role of NKT cells in mycobacterial infection. Flow cytometric analysis is under way to examine profiles of cells that secrete various cytokines at different time points.

ACKNOWLEDGEMENT This study was supported in part by an International Collaborative Study Grant to the chief investigator, Dr Isamu Sugawara, from the Ministry of Health, Labor and Welfare, Japan.

REFERENCES 1. Bendelac A, Lantz O, Quimby M E, Yewdell J W, Bennink J R, Brutkiewicz R R. CD1 recognition by mouse NK1+ T lymphocytes. Science 1995; 268: 863–865. 2. Ortaldo J R, Winckler-Pickett R, Mason A T, Mason L H. The Ly-49 family: regulation of cytotoxicity and cytokine production in murine CD3+cells. J Immunol1998; 160: 1158–1165. 3. Koseki H, Asano H, Inaba T, Miyashita N, Moriwaki K, Lindahl K F, Mizutani Y, Imai K, Taniguchi M. Dominant expression of a distinctive V14+ T-cell antigen receptor a chain in mice. Proc Natl Acad Sci USA 1991; 88: 7518–7522. 4. Kawano T, Cui J, Koezuka Y, Toura I, Kaneko Y, Motoki K, Ueno H, Nakagawa R, Sato H, Kondo E, Koseki H, Taniguchi M. CD1drestricted and TCR-mediated activation of Va14 NKT cells by glycosylceramides. Science 1997; 278: 1626–1629. 5. Kawano T, Cui J, Koezuka Y, Toura I, Kaneko Y, Sato H, Kondo E, Harada M, Koseki H, Nakayama T, Tanaka Y, Taniguchi M. Natural killer-like nonspecific tumor cell lysis mediated by specific ligand- activated Va14 NKT cells. Proc Natl Acad Sci USA 1998; 95: 5690–5693. 6. Ito K, Karasawa M, Kawano T, Akasaka T, Koseki H, Akutsu Y, Kondo E, Sekiya S, Sekikawa K, Harada M, Yamashita M, Nakayama T, Taniguchi M. Involvement of decidual Va14 NKT cells in abortion. Proc Natl Acad Sci USA 2000; 97: 740–744.

Tuberculosis (2002) 82(2/3), 97--104

104

Sugawara et al.

7. Schofield L, McConville M J, Hansen D, Campbell A S, Fraser-Reid B, Grusby M J, Tachado S D. CD1d-restricted immunoglobulin G formation to GPI-anchored antigens mediated by NKT cells. Science 1999; 283: 225–229. 8. Apostolou I, Takahama Y, Belmant C, Kawano T, Huerre M, Marchal G, Cui J, Taniguchi M, Nakauchi H, Fournie J –J, Kourilsky P, Gachelin G. Murine natural killer cells contribute to the granulomatous reaction caused by mycobacterial cell walls. Proc Natl Acad Sci USA 1999; 96: 5141–5146. 9. Yamada H, Mizuno S, Horai R, Iwakura Y, Sugawara I. Protective role of Interleukin-1 in mycobacterial infection in IL-1 a/b double-knockout mice. Lab Invest 2000; 80: 759–767. 10. Sugawara I, Mizuno S, Yamada H, Matsumoto M, Akira S. Disruption of nuclear factor-interleukin-6 (NF-IL6), a transcription factor, results in severe mycobacterial infection. Am J Pathol 2001; 158: 361–366. 11. Sugawara I, Yamada H, Kaneko H, Mizuno S, Takeda K, Akira S. Role of interleukin-18 (IL-18) in mycobacterial infection in IL-18gene- disrupted mice. Infect Immun 1999; 67: 2585–2589. 12. Yamada H, Mizuno S, Reza-Gholizadeh M, Sugawara I. Relative importance of nuclear factor (NF)-k B p50 in mycobacterial infection. Infect Immun 2001; 69: 7100–7105. 13. Emoto M, Emoto Y, Buchwalow I B, Kaufmann S H. Induction of IFN-b-producing CD4+ natural killer T cells by Mycobacterium bovis bacillus Calmette Guerin. Eur J Immunol 1999; 29: 650–659.

Tuberculosis (2002) 82(2/3), 97--104

14. Emoto M, Zerrahn J, Miyamoto M, Perarnau B, Kaufmann S H. Phenotypic characterization of CD8+NKT cells. Eur J Immunol 30: 2300–2311. 15. Cooper A M, Dalton D K, Stewart T A, Griffin J P, Russel D, Orme I M. Disseminated tuberculosis in interferon- ggene-disrupted mice. J Exp Med 1993; 178: 2243–2247. 16. Flynn J L, Chan J, Triebold K J, Dalton D K, Stewart T A, Bloom B R. An essential role for interferon gin resistance to Mycobacterium tuberculosis infection. J Exp Med 1993; 178: 2249–2254. 17. Sugawara I, Yamada H, Kazumi Y, Doi N, Otomo K, Aoki T, Mizuno S, Udagawa T, Tagawa Y, Iwakura Y. Granulomas in interferon g gene-disrupted mice are inducible by avirulent Mycobacterium, but not by virulent Mycobacterium. J Med Microbiol 1998; 47: 871–877. 18. Flynn J L, Goldstein M M, Chan J, Triebold KJ, Pfeffer K, Lowenstein C J, Schreiber R, Mak T W, Bloom B R. Tumor necrosis factor-a is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 1995; 2: 561–572. 19. Kaneko H, Yamada H, Kazumi Y, Mizuno S, Sekikawa K, Sugawara I. The role of tumor necrosis factor-a in Mycobacterium-induced granuloma formation in TNF-a deficient mice. Lab Invest 1999; 79: 379–386. 20. Sugawara I, Yamada H, Mizuno S, Iwakura Y. IL-4 is required for defense against mycobacterial infection. Microbiol Immunol 2000; 44: 971–979.

& 2002 Elsevier Science Ltd. All rights reserved