Proc. Natl. Acad. Sci. USA
Vol. 94, pp. 5243–5248, May 1997
Identification of nitric oxide synthase as a protective locus
(Mycobacterium tuberculosis?infectious disease)
JOHN D. MACMICKING*?, ROBERT J. NORTH†, RON LACOURSE†, JOHN S. MUDGETT‡, SHRENIK K. SHAH§,
AND CARL F. NATHAN*¶
*Beatrice & Samuel A. Seaver Laboratory, Department of Medicine, Cornell University Medical College, New York, NY 10021;†Trudeau Institute, Saranac
Lake, NY 12983; and Departments of‡Immunology and Inflammation and§Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065
Communicated by Maclyn McCarty, The Rockefeller University, New York, NY, March 13, 1997 (received for review December 9, 1996)
helped identify response pathways necessary to combat tu-
berculosis. Several such pathways may function as activators
of a common protective gene: inducible nitric oxide synthase
(NOS2). Here we provide direct evidence for this gene con-
trolling primary Mycobacterium tuberculosis infection using
mice homozygous for a disrupted NOS2 allele. NOS2?/?mice
proved highly susceptible, resembling wild-type littermates
immunosuppressed by high-dose glucocorticoids, and allowed
Mycobacterium tuberculosis to replicate faster in the lungs than
reported for other gene-deficient hosts. Susceptibility ap-
peared to be independent of the only known naturally inher-
ited antimicrobial locus, NRAMP1. Progression of chronic
tuberculosis in wild-type mice was accelerated by specifically
inhibiting NOS2 via administration of N6-(1-iminoethyl)-L-
lysine. Together these findings identify NOS2 as a critical host
gene for tuberculostasis.
Mutagenesis of the host immune system has
Tuberculosis, the leading cause of death from infectious
disease (1), poses an even greater threat as immunodeficiency
spreads among the host population and drug resistance rises in
the pathogen, Mycobacterium tuberculosis (Mtb). A genetic
search for mammalian host resistance pathways has revealed
increased mycobacterial lethality in mice rendered deficient in
T cell subsets [via chromosomal disruption of ?2-microglobu-
lin, T cell receptor-?, T cell receptor-?, or recombination-
activating gene RAG2] (2–4) or in those cytokines and their
receptors responsible for macrophage activation [interferon-?
(IFN-?), IFN-? receptor, or tumor necrosis factor receptor-1
(TNFR1)] (5–8). Other mice harboring mutations in either the
interferon regulatory factor-1 or natural resistance-associated
macrophage protein 1 (NRAMP1) loci are more permissive for
growth of the attenuated bacillus Calmette–Gue ´rin (BCG)
vaccine strain of Mycobacterium bovis (9, 10).
At least five of the above transgenic lines have been exam-
ined for expression of the immunologically induced antimi-
crobial enzyme, nitric oxide synthase (NOS2) (11); all were
found to be deficient. Impaired NOS2 activity arising as a
secondary defect suggests that this gene may represent a point
of convergence for several mycobacteristatic pathways. In
some studies, the antimycobacterial responses of IFN-?-,
treated mouse and human macrophages have been blocked by
NOS inhibitors (12–15), and such inhibitors exacerbated the
course of disease in mice (16). Moreover, NOS2 was expressed
in pulmonary alveolar macrophages from patients with tubercu-
losis (17). These observations underscore that NOS2 may repre-
sent a pivotal protective locus against tuberculosis, a hypothesis
examined here using gene-targeted mice devoid of NOS2.
Mice. Adult (8–12 weeks) male and female NOS2?/?mice
(18) and their wild-type or heterozygous littermates were F2–3
(129?SvEv ? C57BL?6) intercross progeny derived from our
specific pathogen-free colony maintained at The Rockefeller
University, before shipment to the Trudeau Institute. Exper-
iments were performed according to each institutions’ guide-
lines for animal use and care.
Mycobacterial Infection and Enumeration. Mtb (Erdman
strain; Trudeau Mycobacterial Culture Collection No. 993)
was supplied as frozen (?70?C) log phase dispersed cultures in
Proskauer and Beck medium (Difco) containing 0.01% Tween
80 and live bacilli administered via the lateral tail vein as
previously described (3). Inoculum dose was corroborated by
plating aliquots of the original injectate and 24-hr-infected
organ homogenates (10-fold serial dilutions in PBS, 0.01%
Tween 80) onto enriched agar (Middlebrook 7H11, Difco) and
colony-forming units (CFU) enumerated in triplicate 21–28
days later. Mice were genotyped before experimentation via
Southern blot hybridization analysis (18) and verified imme-
diately postmortem by that or PCR.
NOS2 PCR Genotyping. Tail biopsy DNA was prepared as
described by Laird et al. (19), and 400–600 ng was analyzed by
PCR in a 50-?l reaction volume containing 10 mM Tris?HCl
(pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 200 mM dNTPs, and 2.5
units of Taq polymerase (Perkin–Elmer). Primer pairs (200
nM each) were as follows: 5? primer (5?-ATCAGCCTT-
TCTCTGTCTCC-3?), 3? primer (5?-GGCTTTCTGCTGT-
TCTCTC-3?), wild-type allele (413-bp amplificand); 5? primer
3? primer (5?-GCCTGAAGAACGAGATCAGCAGCCTC-
TG-3?), targeted allele (1,288-bp amplificand). PCR amplifi-
at 94?C, 45 s at 62?C, and 3 min at 72?C before electrophoreti-
cally separating the products on a 1.2% agarose gel and
visualizing by ethidium bromide staining. A 1:1 correspon-
dence was observed with mice genotyped by Southern blot
enlisting a KpnI–EcoRI genomic probe (18).
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked ‘‘advertisement’’ in
accordance with 18 U.S.C. §1734 solely to indicate this fact.
Copyright ? 1997 by THE NATIONAL ACADEMY OF SCIENCES OF THE USA
PNAS is available online at http:??www.pnas.org.
Abbreviations: NOS2, inducible nitric oxide synthase; Mtb, Mycobac-
terium tuberculosis; BCG, bacillus Calmette–Gue ´rin; IFN-?, interfer-
on-?; TNF, tumor necrosis factor; TNFR1, TNF receptor-1;
NRAMP1, natural resistance associated macrophage protein 1; CFU,
colony-forming unit; RSNO, low molecular weight S-nitrosothiol; p.i.,
postinfection; AFB, acid-fast bacilli; HC, hydrocortisone; ?2M, ?2-
microglobulin; L-NIL, N6-(1-iminoethyl)-L-lysine; D-NIL, N6-(1-
¶To whom reprint requests should be addressed. e-mail:
?Present address: Laboratory of Immunology, The Rockefeller Uni-
versity, New York, NY 10021.
NRAMP1 Haplotyping. Identification of Bcg polymorphisms
at nucleotide 596 was undertaken via the allele-specific PCR
method of Medina et al. (20). NOS2-targeted 129?Sv AB2.1
ES cell (clone 6.16; ref. 18) and C57BL?6 genomic DNAs
served as Bcgrand Bcgsstandards, respectively.
Plasma NOx and Low Molecular Weight S-Nitrosothiol
(RSNO) Assays. Plasma samples subjected to ultrafiltration
(Centricon 10; Amicon) were measured in triplicate for NO2?
? NO3?by nitrate reductase-linked diazotization assay (18)
and RSNOs after Hg2?ion displacement (21). S-Nitroso-L-
as the RSNO standard, with linear detection limits of 1.5–200
?M. NO2?constituted ? 4% of total plasma NOx, and was
removed by treatment with 0.5% (wt?vol) H2NSO4NH4before
Immunohistochemistry. Ultrathin (2 ?m) sections of for-
malin-fixed, paraffin-embedded lung tissue were incubated
overnight at 4?C with rabbit anti-holo MuNOS2 IgG (22)
(1:1,000 dilution), and a biotinylated goat anti-rabbit IgG
applied before visualizing with avidin-biotin horseradish per-
oxidase (Vector Laboratories) using 3-amino-9-ethylcarbazole
(Sigma) as substrate (23). Sections were counterstained with
hematoxylin. Immunoreactive foci were absent in NOS2?/?
mice, and in NOS2?/?mice if either preimmune serum was
used or the primary IgG omitted. Substitution of the latter by
an anti-MuNOS2 peptide antibody (NO16; ref. 24) (1:500
dilution) gave staining concordant with the anti-holo Mu-
Histology. Formalin-fixed, paraffin-embedded tissues were
sectioned (4–6 ?m) and stained with hematoxylin and eosin or
by the Ziehl–Neelsen method for acid-fast bacilli (AFB) as
Cytokine ELISA Analysis. Undiluted plasma samples were
assayed in triplicate for MuIFN-? and MuTNF-? by ELISA
(Duo Set; Genzyme). Assay sensitivity ranged between 25 and
810 pg?ml for IFN-? and 70–2,250 pg?ml for TNF-?.
NOS2 Inhibitors and in Vivo Administration. N6-(1-
Iminoethyl)lysine (NIL; L- and D-enantiomers) were synthe-
sized as the hydrochloride salts (26) and supplied in acidified
(pH 2.7) drinking water (4 mM) ad libitum beginning on day
41, and the supply changed every 48 hr through day 70
postinfection (p.i.). Efficacy was assessed by plasma NOxand
RSNO assays before (day 30 p.i.) and during (day 69 p.i.) NIL
administration. Neither enantiomer affected food nor water
consumption, and necropsy revealed no sign of drug toxicity.
Hydrocortisone acetate (United Research Laboratories, Phil-
adelphia) was administered subcutaneously (2.5 mg; 100
mg?kg?1) on days 5 and 10 p.i. for primary Mtb infection (106
CFU) or 40 and 45 p.i. for Mtb reactivation (105CFU).
Nonimmunosuppressed control groups received vehicle (0.2
ml) PBS alone.
Statistical Analysis. Log rank product limit estimates were
applied to Kaplan–Meier survival data for determining sig-
t tests and one-way ANOVA at 95% confidence intervals were
performed using InStat PC software, Version 2.00 (GraphPad,
NOS2?/?Mice Rapidly Succumb to Mtb Infection. Whether
NOS2 is essential for the orchestrated cellular immune re-
sponse mounted by the host during primary tuberculosis was
evaluated by inoculating NOS2?/?mice i.v. with 105virulent
Mtb Erdman bacilli. Consolidating pneumonitis and death
were observed within 33–45 days (mean 37.7 ? 1.2 days) p.i.
(Fig. 1a). Wild-type or heterozygous 129?SvEv ? C57BL?6
F2–3littermates survived a mean 160.0 ? 17.1 days and 151.3 ?
16.5 days, respectively. AFB were abundant within large,
sometimes necrotizing, granulomatous lesions in NOS2?/?
lungs, livers, and spleens examined shortly before mice suc-
cumbed to infection. In contrast, few AFB were evident in the
same organs of control mice at day 30 p.i. (Figs. 1c and 2).
infection. (a) Survival curves of litter-matched NOS2?/?(n ? 15),
wild-type (n ? 11), and heterozygous mice (n ? 4) mice injected i.v.
with 105CFU of Mtb Erdman bacilli. Data are from two independent
experiments. Differences between NOS2?/?or NOS2???and
NOS2?/?were statistically significant (P ? 0.0001, log-rank test). (b)
Genotypic and haplotypic allele-specific PCR strategies. Primer po-
sitions (arrows) for murine NOS2 originate within the antisense
phosphoglycerate kinase-Neortargeted insertion (hatched box) or 5?
untranslated region sequences. Terminal 8-bp template sequences are
shown. Sizes of the expected amplificands are in brackets. NRAMP1
primer pairs flank intron 5, with the 3? polymorphic substitution
(nucleotide 596) underlined. (c) NRAMP1 haplotype distribution
versus Mtb growth (mean ? SEM) in the lungs (E, F), livers (?, ?),
mice (Bcgr, n ? 9; Bcg, n ? 5), respectively. (Inset) PCR amplificands
of inherited NRAMP1 variant alleles (r, Bcgr; s, Bcgs) and intact or
targeted NOS2 alleles (arrows), the latter shown with a NOS2???
Absence of NOS2 confers susceptibility to primary Mtb
5244Immunology: MacMicking et al. Proc. Natl. Acad. Sci. USA 94 (1997)
Immunohistochemistry revealed the presence of macro-
phage NOS2 antigen within lung granulomas of Mtb-infected
wild-type mice and its absence in NOS2?/?mice, although
granulomas appeared equally well formed in either genotype
(Fig. 2). Plasma NO2?? NO3?confirmed the marked immu-
nologic induction of high-output NO production in wild-type
hosts by Mtb infection (31.3 ? 4.2, 177.0 ? 5.1, and 136.3 ?
10.9 ?M at days 1, 15, and 30 p.i), while failing to increase NO
secretion in NOS2?/?mice (21.4 ? 2.7, 26.4 ? 2.0, and 24.8 ?
2.5 ?M at days 1, 15, and 30 p.i). S-nitrosothiols, recently
shown to mediate antibacterial effects during murine Salmo-
nella typhimurium infection (27), also were elevated in wild-
type (6.8 ? 2.1, 18.3 ? 3.2, and 14.4 ? 2.4 ?M at days 1, 15,
and 30 p.i., respectively) versus mutant mice (5.8 ? 1.6, 6.2 ?
1.2, and 4.8 ? 1.5 ?M at days 1, 15 and 30 p.i.). The lower
production of NO by heterozygotes (plasma NO2?? NO3?,
63.4–71.5%; RSNO, 54.6–82.4% that of the increase in
NOS2?/?mice) appeared sufficient to confer extended pro-
tection (?150 days).
NOS2 Controls Mtb Growth Independently of NRAMP1.
Parental inbred (129?Sv, C57BL?6) and hybrid descendants
respond similarly to Erdman infection (2, 6), suggesting the
introduced defect, and not strain polymorphisms, account for
the susceptibility of NOS2?/?offspring. Moreover, by using
intercross progeny at F2 or later generations, any additive
effects of the NRAMP1 (Bcg) locus with NOS2 can be tested,
because the 129?Sv and C57BL?6 strains possess resistant
(Gly169) and susceptible (Asp169) autosomal Bcg alleles, re-
spectively (10, 28). Nonconservative replacement of Gly169by
Asp169arises from a recessive point mutation at nucleotide
596; a PCR-based haplotype mapping strategy was enlisted to
discriminate between allelic variants and hence the Bcgrand
Bcgsphenotype (Fig. 1b). NRAMP1 segregated independently
of resistance to growth of Mtb Erdman, as noted for Mtb
H37Rv (20, 25), and of NOS2 status (Fig. 1c). Antimicrobial
synergism between these two loci has been speculated (29, 30),
but cooperative effects were not apparent in the present
experiments. At lower i.v. inocula (?103), minor resistance to
Mtb growth (?0.5 log10CFU) in Bcgrversus Bcgsmice has been
reported (31), although not at doses similar to those given here
(20, 25). In neither case was NOS2 status examined.
Monogenic NOS2 Defect Versus Broad Immunosuppres-
sion. Having identified NOS2 as a monogenic determinant of
resistance to Mtb, we next evaluated the impact of its absence
by comparison with broad immunosuppression. Parenteral
glucocorticoid administration has long represented the means
by which mice are rendered most vulnerable to tubercular
infection (32, 33), even in mice already immunocompromised,
such as those with severe combined immunodeficiency (3).
Accordingly, NOS2?/?and NOS2?/?mice were treated with
a hydrocortisone (HC) regimen previously shown to suppress
mycobacterial resistance both in severe combined immunode-
ficiency hosts and in uncompromised mice of identical major
histocompatibility complex haplotype (H-2b) to those used
here (3). Survival times and mortality rates were similar
between vehicle (PBS)-treated NOS2?/?mice (mean 28.0 ?
1.0 days, 100% lethality) or HC-injected wild-type (mean
26.7 ? 0.6 days, 100% lethality) and HC-recipient NOS2?/?
animals (mean 23.3 ? 1.1 days, 100% lethality) after i.v.
injection with 106CFU of Mtb Erdman (Fig. 3a). In all three
groups, bacillary burdens were significantly greater than for
vehicle-treated wild-type mice at days 20 and 25 p.i. (Fig. 3b).
In the lungs, Mtb multiplied 104.95-fold in unsuppressed
NOS2?/?mice, 105.10-fold in HC-suppressed NOS2?/?mice,
and 105.09-fold in steroid-treated NOS2?/?mice. In healthy
controls, bacilli multiplied by a factor of 102.91and plateaued
thereafter. A comparable pattern was evident within the liver
and spleen, albeit at lower titers (Fig. 3b).
Increases in wild-type plasma NO2?? NO3?and RSNO
levels examined 5 days after the last steroid injection were
73.9% and 86.5% lower than that for the PBS-treated group
(Fig. 4a). This suppression of inducible NO production per-
sisted through day 20 p.i, coincident with the onset of tuber-
cular signs and death (Fig. 4a). Reduced NOS2 expression
after HC treatment also was observed locally within the lung
by immunohistochemistry (data not shown).
NOS2 Is a Crucial Innate Resistance Gene: Comparison
with Other Tuberculostatic Loci. Clonal expansion of antimy-
cobacterial T cells peaks 3–4 weeks after i.v. Mtb Erdman
inoculation (34). By such time, the majority of NOS2?/?and
HC-treated mice either are already ill or have died (Fig. 3a),
suggesting that innate resistance is crucial in determining
survival outcomes in naive mice. Net in vivo doubling times of
Mtb during days 1–15 thus were used to assess host immunity
early after infection (3, 35). Comparison with all reported
studies involving i.v. challenge with Mtb Erdman in genetically
manipulated mice [IFN??/?, TNFR1?/?, or ?2-microglobulin
was fastest in NOS2?/?animals, with or without hydrocorti-
sone treatment (Table 1). However, Mtb doubling times were
comparable for IFN??/?, TNFR1?/?, and NOS2?/?mice as a
fold-increase over their respectively matched H-2b-compatible
Such unilocus similarities may arise because of known NOS2
deficiencies in IFN??/?and TNFR1?/?mice (5, 6) or, alter-
natively, because NOS2?/?mice might possess immunologic
defects involving IFN-? and TNF-?, which predispose them to
tuberculosis. To examine the latter possibility, plasma IFN-?
and TNF-? were measured; both were elevated by day 15 p.i.,
irrespective of NOS2 status (Fig. 4b). In fact, IFN-? levels were
2–3 times higher in NOS2?/?mice, perhaps in response to the
continued presence of replicating bacilli (Fig. 4b). Detectable
IFN-? and TNF-? declined through day 25 p.i., but still
remained above initial (24-hr p.i.) levels (data not shown). HC
in NOS2?/?and NOS2?/?hosts. Histological analysis of AFB in
diseased organs at day 30 p.i. stained by the Ziehl–Neelsen method.
Acid-fast bacterial rods appear red. Magnification 400? (lung) and
100? (spleen). Localized NOS2 expression in lung granulomas as-
sessed by immunohistochemistry using anti-MuNOS2 antibody. Mag-
nification 200?. (Inset) 1,000?. At least six mice of each group were
Immunology: MacMicking et al.Proc. Natl. Acad. Sci. USA 94 (1997) 5245
suppressed TNF-? by 55% to 63% in both groups, consistent
with its known pharmacologic actions (36).
Vigilance of NOS2 in Preventing Tubercular Progression.
The prolonged survival of wild-type mice (Fig. 1a) prompted
us to consider whether NOS2 also might be responsible for
restricting microbial growth and chronic progression in the
late, clinically quiescent phase of the disease. Organ myco-
bacterial burdens have begun to decline by day 25 p.i. (Fig. 1c),
with few AFB detectable by day 30 p.i. (Fig. 2). Beginning on
day 41 p.i., clinically stable NOS2?/?mice received either
N6-(1-iminoethyl)-L-lysine (L-NIL), the most isoform-specific
NOS2 inhibitor known to be potent in other models of
infectious disease (23, 37), or its inactive enantiomer, N6-(1-
iminoethyl)-D-lysine (D-NIL). NOS2 can be ?90% inhibited
when 9 mM L-NIL is imbibed (23); here, limited supply
confined the dosage to 4 mM over 30 days. By day 69 p.i. (day
28 of drug treatment), L-NIL afforded 72% suppression of
plasma NO2?? NO3?and 64% of RSNO compared with
D-NIL, and triggered Mtb growth (0.80–2.06 log10 CFU ?
end of treatment (Fig. 5a). As a positive control (38), another
group was injected subcutaneously with HC on days 45 and
50 p.i. This also led to disease exacerbation, with 75% mor-
tality by day 70 p.i. (Fig. 5a) associated with large increases in
bacillary burden (Fig. 5b). In contrast, D-NIL-treated mice
continued to restrict Mtb replication throughout the period of
treatment (days 41–70 p.i.; Fig. 5b) and survived for the
duration of the experiment (day 90 p.i.; Fig. 5a).
These results formally demonstrate that the NOS2 locus is
necessary to control primary tuberculosis in mice. Absence of
NOS2 led to rapid bacterial growth, necrotic granulomatous
pneumonitis, and death. Despite a 30% to 40% reduction in
NO release accompanying the loss of one NOS2 allele, het-
erozygotes still retained sufficient NOS2 activity to extend
their survival as long as mice possessing the full chromosomal
complement. Additional loci may partially compensate for this
loss, although based on the inability of Bcgrinheritance to
affect Mtb growth in either NOS2?/?or NOS2?/?mice at the
inocula studied here, NRAMP1 appears an unlikely candidate.
Murine macrophage-derived oxidants such as H2O2, O2?
and OH? could represent another avenue of compensation.
Singly they appear ineffectual against Mtb Erdman (15);
however, evidence is accumulating that both H2O2and O2?
can synergize with NO-derived species to enhance microbial
killing (39, 40). Respiratory burst responses in IFN-?-activated
peritoneal macrophages triggered by a mycobacterial agonist
were intact in NOS2?/?mice (18). The possibility exists,
therefore, that in heterozygous mice these oxidative species
may react with the smaller amounts of NO produced to
embellish its antitubercular action.
Cooperative antimicrobial effects of NO and another redox
partner also may extend to the RSNOs, which increased
coid-immunosuppressed mice infected with Mtb. (a) Survival times for
Mtb Erdman-infected (106CFU i.v.) mice given 2.5 mg of HC s.c. on
days 5 and 10 p.i. NOS2?/?(n ? 16); NOS2?/?(n ? 18). Nonimmu-
nosuppressed control groups received vehicle (0.2 ml PBS) alone.
NOS2?/?(n ? 17); NOS2?/?(n ? 19). Both HC-treated groups and
NOS2?/?mice were significantly different from PBS-treated wild-type
controls (P ? 0.0002, log rank). Data represent two independent
experiments. (b) Organ mycobacterial burdens (means ? SEM) of
infected mice (n ? 5–10 per time point). HC-treated groups were
examined earlier (day 20) due to overt illness. ?, P ? 0.0002 versus the
PBS-treated wild-type group, ANOVA.
Phenotypic similarity of NOS2-deficient and glucocorti-
tuberculostatic cytokines. (a) Release of NO2?? NO3?and RSNO
during the course of infection and the effects of HC on their secretion.
Symbols are as in Fig. 3 and represent individual sera assayed in
triplicate; horizontal bars denote group means. ?, P ? 0.01, ??, P ?
0.0001, unpaired t test. (b) Plasma TNF-? or IFN-? responses (mean ?
SEM) of PBS- and HC-treated mice determined in triplicate by
ELISA. ?, P ? 0.048, ??, P ? 0.038.
Suppressive effects of HC on inducible NO production and
5246Immunology: MacMicking et al. Proc. Natl. Acad. Sci. USA 94 (1997)
several-fold in plasma during Mtb Erdman infection. However,
only 5% to 8% of detectable NO was bound to low molecular
weight thiols, perhaps calling into question their significance in
the present studies. Countering this argument is an established
history of RSNOs as potent virustatic, parasiticidal, and
bactericidal agents (reviewed in ref. 11). Moreover, NO en-
tering the bacterial cell via a thiol adduct (39, 41) may be
protected from facile interactions with Mtb outer envelope
glycolipids, which can scavenge oxidant species (42). Hence,
despite their small quantities, RSNOs nonetheless could have
contributed to the antimycobacterial effect.
Regardless of whether NO or one of its derivatives predom-
inated, NOS2 appeared so central for tuberculostatic defense
that its disruption had an impact on murine tuberculosis
quantitatively equivalent to that of glucocorticoids. Nonsup-
pressed NOS2?/?mice closely resembled glucocorticoid-
suppressed wild types, whereas glucocorticoids had little ad-
ditive effect in mutant animals. This is consonant with the view
that in HC-immunosuppressed wild-type mice, susceptibility
may have largely reflected the diminution in macrophage
NOS2 expression (43–45). While the lymphoablative effects of
glucocorticoids are well documented (reviewed in ref. 36),
studies in severe combined immunodeficiency mice demon-
strated the existence of residual, glucocorticoid-sensitive re-
sistance to Mtb (3). Whether this non-T cell, non-B cell-
dependent resistance pathway involved NK cells cannot be
discounted; however, it is more likely to be manifested in
macrophages, the cells in which mycobacteria chiefly reside
and replicate (3). The experiments described herein provide
support for macrophage NOS2 being a major steroid-sensitive
TNF-? (55% to 63%) and to a lesser extent IFN-? (?22%)
also were inhibited by glucocorticoids, though neither as much
elevated in the serum of nonsuppressed NOS2?/?mice to the
same extent as in wild-type mice, presumably reflecting pro-
duction at the sites of infection, because the latter represents
the stimulus for their release. Indeed, experiments conducted
to date have failed to identify immunodeficiencies in NOS2?/?
mice other than the introduced defect that could render them
susceptible to Mtb infection. Many of the cellular components
considered important for the immune response to tuberculosis
are unaffected in these mice: CD4 and CD8 single positive T
cell populations; TH1 proliferative responses (e.g., to other
intracellular pathogens); IFN-?-up-regulated major histocom-
patibility complex Class II expression; BCG-triggered respi-
ratory burst responses; and elaboration of IFN-? and TNF-?,
the only cytokines known to activate murine macrophages to
inhibit Mtb (18, 46). Of course, other factors not yet examined
could be dysregulated.
When germ-line mutations were introduced into several of
these protective loci (e.g., IFN-?, TNFR1, T cell receptor-?, T
cell receptor-?, and ?2M), they, too, allowed virulent Mtb to
clinically quiescent phase of tuberculosis in wild-type mice. (a) Mor-
tality in clinically stable wild-type mice (105CFU i.v.) receiving 4 mM
L-NIL supplied within the period bracketed (n ? 8) or parenteral HC
treatment (2.5 mg) administered s.c. on days 45 (HC-1) and 50 p.i.
(HC-2) (n ? 8). Controls were untreated NOS2?/?(n ? 10) and
NOS2?/?mice (n ? 10) and NOS2?/?mice given 4 mM D-NIL (n ?
8). Both L-NIL and HC-treated groups were significantly different
from D-NIL-treated wild-type controls (P ? 0.0005, log rank). (b)
Mycobacterial titers (means ? SEM) at the start (day 41 p.i.), within
4 per time point, with a common pretreatment group). The day 70 p.i.
time point was omitted for the HC group due to earlier mortality.
Isolated symbols represent recovered inocula on day 1 p.i. ?, P ? 0.01,
??, P ? 0.0001 versus the D-NIL group, ANOVA.
NOS2 inhibition accelerates disease progression during the
Table 1. Comparison of net in vivo doubling times for Mtb Erdman in genetically deficient hosts
Doubling time, hr
(B6 ? 129 F2-3)
(B6 ? 129 F2-3)
(B6 ? 129 F2-3)
(B6 ? 129 F2-3)
(B6 ? 129 Fn)
(B6 ? 129 Fn)
Doubling times calculated from mycobacillary counts according to ref. 3 during days 1–15 p.i. with the exception of (iv) (days 10–14 p.i.). Genetic
background given in parentheses, and n equals the total number of mice for both time points. All inocula administered i.v. R denotes resolving
(net replication at day 15 p.i. below day 1 p.i.). *Present study,†from ref. 5,‡from ref. 2, and§from ref. 6.
Immunology: MacMicking et al. Proc. Natl. Acad. Sci. USA 94 (1997) 5247
establish overwhelming infections (2, 4–6). Of these, IFN-
??/?, TNFR1?/?, and ?2M?/?mice were examined under
comparable conditions (105–106CFU of Mtb Erdman i.v.) and
were of matched or similar genetic background to those
NOS2?/?animals used here. Interstudy comparison of in vivo
Mtb doubling times thus could be undertaken. Within the lung,
these times were shorter in NOS2?/?mice (27–31 hr) than in
the other knock-outs (38–78 hr), approaching the intrinsic
replication rate (?20 hr) for Erdman in culture (47). This
represents a remarkably permissive environment for Mtb
growth and underscores the importance of NOS2 in providing
a hostile one. The latter’s deficiency in IFN-??/?, IFN-?R?/?,
and TNFR1?/?mice implies that NOS2 acts distally to these
loci, a suggestion lent further credence by the findings that
IFN-? and TNF-? expression were intact in NOS2?/?mice, as
was granuloma formation.
In mice, as in people, the sterile eradication of Mtb is rarely
achieved (33), suggesting that long-term CD4?memory T cells
must continually enlist the aid of macrophages to maintain
bacterial dormancy. A requirement for NOS2 later during
infection therefore could be expected if the host is to avoid
disease recrudescence. Specifically inhibiting NOS2 with L-
NIL during the late phase of clinical stability supported this
hypothesis, because infection progressed more quickly and led
to earlier mortality. This was observed despite submaximal
(?70%) NOS2-specific inhibition. It also may explain why
disease acceleration was greater with HC, which, in addition to
its suppressive action on mouse macrophage NOS2 expression
(44), has direct inhibitory effects on T cell memory (38).
To the extent that findings in mice are germane to man, the
fact that NOS2 appears necessary to control mycobacterial
growth may have implications for the global incidence of
human tuberculosis, because Mtb currently infects over one-
third of the world’s population (1). A new set of factors thus
could influence whether infected individuals develop disease:
polymorphisms within the NOS2 locus, the balance of NOS2-
inducing or NOS2-inhibiting cytokines (11), clinical use of
NOS inhibitors, or the expression of microbial genes that
confer resistance to nitroxergic products (41, 48).
We thank L. Riley and B. Rogerson for helpful discussions; L. Ryan
for histology; and P. Davies and M. McCoss for generous support. This
work was supported by National Institutes of Health Grants HL51967
and AI34543 (C.F.N.) and HL51960 (R.J.N.).
1.Raviglione, M. C., Snider, D. E., Jr., & Kochi, A. (1995) J. Am.
Med. Assoc. 273, 220–226.
Flynn,J.L,Goldstein,M. M.,Triebold,K. J.,Koller,B.&Bloom,
B. R. (1992) Proc. Natl. Acad. Sci. USA 89, 12013–12017.
North, R. J. & Izzo, A. A. (1993) J. Exp. Med. 177, 1723–1733.
Ladel, C. H., Blum, C., Dreher, A., Reifenberg, K. & Kaufmann,
S. H. E. (1995) Eur. J. Immunol. 25, 2877–2881.
Flynn, J. L., Chan, J., Triebold, K. J., Dalton, D. K., Stewart,
T. A. & Bloom, B. R. (1993) J. Exp. Med. 178, 2249–2254.
Flynn, J. L., Goldstein, M. M., Chan, J., Triebold, K. J., Pfeffer,
K., Lowenstein, C. J., Schreiber, R., Mak, T. W. & Bloom, B. R.
(1995) Immunity 2, 561–572.
Dalton, D. K., Pitts-Meek, S., Keshev, S., Figari, I. S., Bradley, A.
& Stewart, T. A. (1993) Science 259, 1739–1742.
Kamijo, R., Le, J., Shapiro, D., Havell, E. A., Huang, S., Aguet,
M., Bosland, M. & Vilcek, J. (1993) J. Exp. Med. 178, 1435–1440.
Kamijo, R., Harada, H., Matsuyama, T., Bosland, M., Gereci-
tano, J., Shapiro, D., Le, J., Koh, S. I., Kimura, T., Green, S. J.,
Mak, T. W., Taniguchi, T. & Vilcek, J. (1994) Science 263,
Vidal, S., Tremblay, M. L., Govoni, G., Gauthier, S., Sebastiani,
G., Malo, D., Skamene, E., Olivier, M., Jothy, S. & Gros, P.
(1995) J. Exp. Med. 182, 655–666.
MacMicking, J., Xie, Q.-W. & Nathan, C. (1997) Annu. Rev.
Immunol. 15, 323–350.
Denis, M. (1991) Cell. Immunol. 132, 150–157.
Denis, M. (1991) J. Leukocyte Biol. 49, 380–387.
14. Flesch I. E. A. & Kaufmann, S. H. E. (1991) Infect. Immun. 59,
Chan, J., Ying, Y., Magliozzo, R. S. & Bloom, B. R. (1992) J. Exp.
Med. 175, 1111–1122.
Chan, J., Tanaka, K., Carroll, D., Flynn, J. & Bloom, B. R. (1995)
Infect. Immun. 63, 736–740.
Nicholson, S., Bonecini-Almeida, M. da G., Lapa e Silva, J. R.,
Nathan, C., Xie, Q.-W., Mumford, R., Weidner, J. R., Calaycay,
J., Geng, J., Boechat, N., Linhares, C., Rom, W. & Ho, J. L.
(1996) J. Exp. Med. 183, 2293–2302.
MacMicking, J. D., Nathan, C., Hom, G., Chartrain, N., Fletcher,
D. S., Trumbauer, M., Stevens, K., Xie, Q.-W., Sokol, K.,
Hutchinson, N., Chen, H. & Mudgett, J. S. (1995) Cell 81,
Laird, P. W., Zijerveld, A., Linders, K., Rudnicki, M. A., Jae-
nisch, R. & Berns, A. (1991) Nucleic Acids Res. 19, 4293.
Medina, E., Rogerson, B. J. & North, R. J. (1996) Immunology
Saville, B. (1958) Analyst 83, 670–672.
Xie, Q.-W., Cho, H. J., Calaycay, J., Mumford, R. A., Swiderek,
K. M.,Lee,T. D.,Ding,A.,Troso,T.&Nathan,C.(1992)Science
Stenger, S., Donhauser, N., Thurling, H., Rollinghoff, M. &
Bogdan, C. (1996) J. Exp. Med. 183, 1501–1514.
Xie, Q.-W., Cho, H. J., Kashiwabara, Y., Baum, M., Weidner,
J. R., Elliston, K., Mumford, R. & Nathan, C. (1994) J. Biol.
Chem. 269, 28500–28505.
Medina, E. & North, R. J. (1996) J. Exp. Med. 183, 1045–1051.
Moore, W. M., Webber, R. K., Jerome, G. M., Tjoeng, F. S.,
Misko, T. P. & Currie, MG. (1994) J. Med. Chem. 37, 3886–3888.
De Groote, M. A., Testerman, T., Xu, Y., Stauffer, G. & Fang,
F. C. (1996) Science 272, 414–417.
Malo, D., Vogen, K., Vidal, S., Hu, J., Cellier, M., Schurr, E.,
Fuks, A., Bumstead, N., Morgan, K. & Gros, P. (1994) Genomics
Vidal, S. M., Malo, D., Vogen, K., Skamene, E. & Gros, P. (1993)
Cell 73, 469–485.
Formica, S., Roach, T. I. A. & Blackwell, J. M. (1994) Immunol-
ogy 82, 42–50.
Brown, D. H., Miles, B. H. & Zwilling, B. S. (1995) Infect.
Immun. 63, 2243–2247.
Batten, J. C. & McClune, R. M., Jr. (1957) Br. J. Exp. Path. 38,
McClune, R. M., Jr., Feldmann, F. M., Lambert, H. P. & Mc-
Dermott, W. (1966) J. Exp. Med. 123, 445–468.
Orme, I. M. (1987) J. Immunol. 138, 293–298.
Molloy, A., Laochumroonvorapong, P. & Kaplan, G. (1994) J.
Exp. Med. 180, 1499–1509.
Barnes, P. J. & Adcock, I. (1993) Trends Pharmacol. Sci. 14,
Stenger, S., Thurling, H., Rollinghoff, M., Manning, P. & Bog-
dan, C. (1995) Eur. J. Pharmacol. 294, 703–712.
Cox, J. H., Knight, B. C. & Ivanyi, J. (1989) Infect. Immun. 57,
De Groote, M. A., Granger, D., Xu, Y., Campbell, G., Prince, R.
& Fang, F. C. (1995) Proc. Natl. Acad. Sci. USA 92, 6399–6403.
Pacelli, R., Wink, D. A., Cook, J. A., Krishna, M. C., DeGraff,
W., Friedman, N., Tsokos, M., Samuni, A. & Mitchell, J. B.
(1995) J. Exp. Med. 182, 1469–1479.
Hausladen, A., Privalle, C. T., Kneg, T., DeAngelo, J. & Stamler,
J. S. (1996) Cell 86, 719–729.
Chan, J., Fan, X.-D., Hunter, S. W., Brennan, P. J. & Bloom,
B. R. (1991) Infect. Immun. 59, 1755–1761.
Rook, G. A. W., Steele, J., Ainsworth, M. & Leveton, C. (1987)
Eur. J. Respir. Dis. 71, 286–291.
DiRosa, M., Radomski, M., Carnuccio, R. & Moncada, S. (1990)
Biochem. Biophys. Res. Commun. 172, 1246–1252.
Kunz, D., Walker, G., Eberhart, W. & Pfeilschifter, J. (1995)
Proc. Natl. Acad. Sci. USA 93, 255–259.
Wei, X-Q., Charles, I. G., Smith, A., Ure, J., Huang, F.-P, Xu, D.,
Muller, W., Moncada, S. & Liew, F. Y. (1995) Nature (London)
Respir. Dis. 115, 1066–1069.
Nunoshiba, T., De Rojas-Walker, T., Tannenbaum, S. R. &
Demple, B. (1995) Infect. Immun. 63, 794–798.
5248Immunology: MacMicking et al.Proc. Natl. Acad. Sci. USA 94 (1997)