INFECTION AND IMMUNITY, Jan. 1987, p. 223-226
Copyright © 1987, American Society for Microbiology
Role of Natural Killer Cells in Infection with the Mouse Pneumonitis
Agent (Murine Chlamydia trachomatis)
DWIGHT M. WILLIAMS,'* JULIUS SCHACHTER,2 AND BARRY GRUBBS'
Infectious Diseases Section, Department ofMedicine, Audie L. Murphy Memorial Veterans Hospital, University of Texas
Health Science Center, San Antonio, Texas 78284,1 and the Department ofLaboratory Medicine, University of
California, San Francisco, California 941432
Received 4 August 1986/Accepted 10 October 1986
Natural killer (NK) activity is increased in both spleen and lung early in pulmonary infection by murine
Chlamydia trachomatis in both susceptible nude and resistant heterozygous (nul+) mice. Ablation of the rise in
NK activity by giving the mice antiasialo GM-1 antibody or stimulation of NK activity by immunomodulators
did not affect quantitative tissue counts of the mouse pneumonitis biovar ofC. trachomatis or significantly affect
survival. Studies are needed to further define the role of NK cells in host defense, immunoregulation, and
immunopathology during chlamydial infection.
Ample precedent exists for a possible role for natural killer
(NK) cells in host defense against nonviral intracellular
organisms such as Toxoplasma gondii (5, 6) and Salmonella
typhimurium (11), as well as other nonviral pathogens such
as Cryptococcus neoformans (10). Further, cytotoxic activ-
ity of relatively uncharacterized type has been shown during
infection with Chlamydia psittaci (2, 9) in mice by use of
spleen cells or spleen-derived cytokines as effectors. How-
ever, an attempt to find cytotoxic T-cell activity in mice
infected with Chlamydia trachomatis was unsuccessful (12),
as was an attempt with humans (13). We felt this provided
to further explore the issue of cellular
cytotoxicity in C. trachomatis infection by monitoring NK
activity during infection by use of the mouse pneumonitis
biovar of C. trachomatis (MoPn). Further, we decided to
examine NK cell activity in the actual organ infected (in this
case, lung) as well as in the spleen. We also decided to use
the standard NK assay with YAC-1 uninfected tumor target
cells instead of the macrophage and fibroblast targets used in
previous studies (2, 9). We were fortunate that a reliable
method for measuring NK function in the murine lung has
recently been described (14).
Our model of murine pneumonia caused by MoPn in both
nude (nulnu) (susceptible) and heterozygous (nul+) (resist-
ant) mice has been presented in detail in recent publications
MATERIALS AND METHODS
MoPn. MoPn was obtained and maintained as previously
described (16, 17). The undiluted titer of MoPn was 4 x 109
inclusion-forming units (IFU)/ml.
Inoculation of mice. Groups of 5 to 20 mice were inocu-
lated under pentobarbital anesthesia with 0.05 ml of MoPn
agent (16, 17). The infection dose was 5 x 103 IFU per
Mice. Specific-pathogen-free nulnu and nul+ mice (parent
strain BALB/c) were obtained and maintained as previously
described (16). They had been made superclean by using
germ-free foster mothers and repopulating them with limited
bacterial flora of organisms nonpathogenic to mice (16).
Mice of both sexes (6 to 8 weeks old) were used in these
Tumor cell lines. These studies made use of the Moloney
virus-induced lymphoma, YAC-1, which was grown as a
suspension culture in RPMI 1640 medium supplemented
with 10% fetal bovine serum, sodium pyruvate, nonessential
amino acids, and glutamine. This cell line was free of
mycoplasma and MoPn by culture.
NK assay. The spleen NK assay was the standard 4-h
assay (5, 6, 10, 11) with splenic effectors from infected and
uninfected nul+ or nulnu mice. The effector/target (E/T)
ratios were 100:1, 50:1, and 25:1 with determination of Na2
51CrO4 release. Incubation time was 4 h at 37°C and 5% C02,
at which time a 100-,ul sample of the supernatant was
counted in a gamma spectrometer (Beckman Instruments,
Inc., Fullerton, Calif.). Spontaneous release was determined
in media alone, and maximum release was determined with
5% Triton X. All tests were run in triplicate on single mice.
Percent cytotoxicity was calculated as follows: % cytotox-
ous)]/[(maximum releasable cpm)
Spontaneous release was always less than 10% of maximum
Isolation of pulmonary NK cells. Isolation of pulmonary
NK cells was done according to the method of Talmadge et
al. (14). Mice were exsanguinated, the pulmonary vascula-
ture was repeatedly perfused with normal saline to remove
residual blood cells, and the lungs were washed by
tracheobronchial lavage. The lungs were removed, thymic
and bronchial tissues were excised, and the remaining pul-
monary tissue was minced and dissociated by two sequential
30-min incubations with 0.14% collagenase and 0.03%
DNase. The resulting suspension was twice washed and
layered onto a Lympholyte-M Ficoll (Pharmacia Fine Chem-
icals, Piscataway, N.J.) gradient with centrifugation at 500 x
g for 30 min. The effector cells were harvested, washed, and
used at effector target ratios of 50:1 and 25:1 in the NK assay
Modulation of NK function. NK function was stimulated
by giving 100 ,ug of poly(I)poly(C) (Sigma Chemical Co.,
St. Louis, Mo.) intravenously (i.v.) 18 h before infection
with MoPn. NK function was ablated by giving 200 ,ug of
antiasialo GM-1 (Wako Chemicals) i.v. at days 0 and 3 of
= [(cpm effectors +
- (cpm spontane-
- (cpm spontaneous)].
Vol. 55, No. 1
WILLIAMS ET AL.
TABLE 1. Spleen NK function in nul+ and nulnu mice
% Cytotoxicitya for:
Infected mice, daysafter infection
11.2 ± 3.7
8.8 ± 3.4
5.7 + 2.1
16.1 + 5.0
10.8 ± 4.3
7.1 + 2.8
12.6 ± 4.0
9.7 ± 3.7
6.7 ± 2.5
17.4 ± 4.0
13.0 ± 6.7
9.3 ± 3.1
34.6 ± 7.4
27.7 ± 7.1
19.2 + 5.9
45.1 ± 5.7
37.0 ± 7.8
27.6 ± 4.2
15.5 ± 4.0
11.3 ± 3.6
6.7 ± 2.4
42.5 ± 10.7
31.11 ± 8.3
21.5 ± 4.8
12.8 + 7.8
8.5 + 5.9
6.0 ± 4.5
11.0 ± 0.2
aResults are the mean±standard deviation of data from three mice in each group.
bp < 0.5 for all values compared with those of uninfected mice.
c-, Insufficient cells obtained from very ill mice.
Quantitative culture of lungs for MoPn. Lungs were har-
vested, minced, and quantitatively cultured for MoPn in
McCoy cell culture as in our previous study (17).
Statistics. Statistics were by the Student t test with cor-
rection for unequal variances.
Table 1 shows the results of NK assays on spleens from
nul+ and nulnu mice at days 2, 5, 10, and 15 after intranasal
infection with MoPn, compared with results of assays on
spleens from uninfected nul+ and nulnu mice. A statistically
significant elevation of NK function occurred in both nul+
and nulnu spleen cells at day 5 postinfection with a return to
base line by day 10 in nul+ spleens. nulnu NK values,
however, remained elevated until day 15 (although complete
testing was not obtained on that day because of very small
spleens in terminally ill nulnu mice).
Of perhaps more biological relevance (Table 2), a similar
significant increase occurred in lung NK function (the organ
where the infection was occurring) at day 5 postinfection.
Thus, significant elevation of NK function occurred in a
significant anatomical compartment according to the stan-
dard NK assay during infection with a C. trachomatis biovar
To determine whether NK function changes had an effect
on infection with MoPn (a role in host defense as opposed to
immunomodulation alone), NK function was increased by
giving two doses of antiasialo GM-1 during infection with
subsequent quantitative culture of MoPn in lungs.
poly(C) significantly increased NK function in
lung 18 h later so that it would be elevated at the time of
infection in mice given poly(I)
(Table 3). The table also shows that antiasialo GM-1 pre-
vented the normal increase of pulmonary NK function seen
by day 5 of infection.
Next, quantitative culture of nul+ lungs for MoPn was
performed at various times after infection in immuno-
manipulated and unmanipulated
-poly(C) 18 h before infection or ablated by
poly(C) the day before
given 18 h before infection to increase NK activity at the
time of infection had no effect on MoPn IFU per lung,
compared with controls on either day 2 or 3 ofinfection (P >
0.30 compared with controls). Mean IFU per lung at day 2
were 3,390 ± 1,349 in mice given no poly(I)
3,486 ± 1,159 in mice given poly(I)
day 3 were 4,1351,821 and 4,775
the same two groups. Each value is the mean ± standard
deviation of results for five mice. Conversely, antiasialo
GM-1 antibody given i.v. to nul+ mice to block the normal
rise in NK activity by day 5 of infection (Tables 1 to 3) had
no effect on MoPn IFU per lung on day 5 postinfection.
Values were 13,848 ± 9,784 IFU per lung (unmanipulated
4,609 (nul+ given antiasialo GM-1), and
13,225 ± 7,509 (unmanipulated nulnu). Thus, no significant
effect on MoPn titers was achieved by immunomodulatory
techniques designed to increase NK function at the time of
poly(C) 18 h earlier] or prevent the nor-
mal increase ofNK activity with infection (antiasialo GM-1).
Finally, a series of experiments was performed in vivo,
giving mice immunomodulators known to increase NK func-
tion. We are aware that any immunomodulator is likely to
influence other immunologic parameters in addition to NK
function over the duration of an experiment, and we recog-
nize that these experiments are likely to be less specific than
the quantitative culture data just presented. The first exper-
iment used murine
San Diego, Calif.; 1.7 x 104 U) given i.v. every 2 days for a
total of four doses. Pulmonary NK activity measured 1 day
after the last dose in uninfected nul+ mice was significantly
increased compared with control mice given excipient con-
trol material (at E/T ratio of 50:1, control mice showed a
percent cytotoxicity of 0.8 ± 0.3, after interferon 5.1 + 0.7
[P < 0.003]). This admittedly modest increase in NK func-
tion had no significant effect on mortality in mice given the
same doses in in vivo studies. With nul+ mice, mortality at
day 20 in mice given 105 IFU ofMoPn intranasally was 100%
in 10 mice given excipient control and 80% in 10 mice given
four doses of interferon starting on day -1 (P > 0.30 by the
poly(C). The values at
2,070, respectively, in
a. plus ,B interferon (Lee Biomolecular,
TABLE 2. Lung NK function in nul+ and nulnu mice
% Cytotoxicity for:
Infected mice, days after infection
1.7 ± 2.5
1.0 ± 1.1
3.5 ± 5.0
3.5 ± 4.4
2.6 ± 1.5
1.9 ± 0.03
4.7 ± 1.2
3.5 ± 0.5
44.2 ± 0.8
29.1 ± 1.3
40.8 ± 8.0
40.1 ± 7.7
3.7 ± 1.5
3.5 ± 1.0
4.6 ± 1.4
4.4 ± 1.6
ap < 0.05 for all values compared with those of uninfected mice.
NK CELL FUNCTION IN CHLAMYDIA INFECTION
TABLE 3. Effect of immunomodulation on pulmonary NK function
Poly(I) * poly(C)a
GM-i, day 5
a100 pLg of poly(I) * poly(C) given i.v. 18 h earlier; mice not infected. P < 0.05 compared with control [no poly(I) * poly(C)] for all values.
b 200 ,ug of antiasialo given i.v. on days0and 3 of infection with MoPn. NK assay done on day 5 of MoPn infection.
c P < 0.05 compared with antiasialo for both values.
46.3 ± 14.3
44.0 t 17.1
59.9 ± 4.1
58.5 ± 5.7
0.9 t 0.4
0.1 ± 0.1
2.7 t 0.8
0.4 t 0.1
22.9 t 2.7
18.2 ± 3.7
1.7 t 0.4
1.3 t 0.5
Wilcoxon two-tail test). With 104 IFU of MoPn, mortality
was 35% in controls and 50% in interferon-treated mice (P >
0.30 by the Wilcoxon two-tail test; 20 mice per group).
doses significantly increased NK function in nul+ uninfected
mice compared with controls given phosphate-buffered sa-
line (at an E/T ratio of 50:1, control mice showed a percent
cytotoxicity of 3.5 ± 1.1, after poly(I)
The same regimen with MoPn-infected nul+ mice (starting
the day before infection) led to no significant mortality
difference [mortality: 100% in control; 100% in poly(I)
* poly(C)-treated mice at day 20 postinfection; P > 0.30 by
the Wilcoxon two-tail
Poly(I) poly(C) was not toxic to uninfected nul+ mice (0%
mortality at day 20).
Repeated doses of 200 ,ug of antiasialo given i.v. were
somewhat toxic to the mice in our study and not suitable for
Recognizing that these mortality experiments are not
specific studies of NK function and that positive benefits of
early NK stimulation could be counterbalanced or out-
weighed by other effects of these immunomodulators, no
significant beneficial effect was observed.
-poly(C) given at 100 ,ug i.v. every 3 days for four
poly(C) 70.8 ± 7.8 [P
10 mice per group].
These studies demonstrate that pulmonary infection with
murine C. trachomatis leads to an increase in both spleen
and lung NK cell tumoricidal activity, peaking early after
infection and declining thereafter. This increase is prevent-
able by treatment of the mice with antiasialo GM-1. The
significance of this NK activity is unclear. As discussed in
the introduction, NK cells may be directly cytotoxic for
some pathogens including nonviral pathogens such as T.
gondii (5) and Cryptococcus neoformans (10) and might play
a role in the control of these infections before specific
immunity develops. In addition, NK cells also may play an
immunoregulatory role. NK cells can suppress the genera-
tion of Lyt 2+ cytotoxic T cells by suppressing or eliminating
dendritic cells (4), can suppress B-cell function (1), and can
be involved in producing imunomodulators such as inter-
feron (3). Our previous data have shown that T cells are
critically important in host defense against pneumonia
caused by MoPn in our model (15-17). Nude mice, which
frequently have increased NK cell activity (7), were more
susceptible to MoPn than were nul+ animals, suggesting that
specific (T-cell-dependent) immunity is more important to
ultimate survival in our model than is early natural or
nonspecific immunity. However, our nulnu mice NK activity
was not as high endogenously as has been reported previ-
ously (7, 8), perhaps because our mice were superclean and
had low-level background stimulation. Thus, the fact that
our nulnu mice demonstrated no early increased resistance
to MoPn with similar lung titers to nul+ on day 5 cannot be
used to exclude a role for NK cells in host defense against
MoPn. However, the facts that ablation of the rise in NK
activity by antiasialo GM-1 antibody and stimulation ofNK
activity by NK inducers had no significant effect on MoPn
lung titers and that NK inducers did not delay mortality due
to MoPn suggest that the in vivo role ofNK cells in directly
controlling the infection is at best small. We are in the
process of further examining this point by studying the in
vitro effects of NK cells on MoPn by using Percoll
(Pharmacia) gradient-enriched populations. It appears likely,
however, that the in vivo role is one primarily of im-
As stated in the introduction, the generation of factors
cytotoxic to Chlamydia sp.-infected cells during C. psittaci
infection has been described (2, 9). Currently, the cell
type(s) responsible for generation of these factors is unclear,
as is its role in host defense or immunopathology. The in
vivo role, if any, could be either beneficial or detrimental to
the host. Concanavalin A-induced cytotoxic factors can be
generated during MoPn infection in our model as well (G.
Byrne, D. Williams, and J. Schachter, unpublished data),
but optimal generation occurs later than day 5, and the role,
if any, of NK cells in the generation of such factors is
This study demonstrates a significant increase in NK
function during chlamydial infection. Further studies are
needed to define the role of NK cells in host defense,
immunoregulation, and immunopathology during chlamydial
This work was supported by Public Health Service grants
AI-22566 and AI-22380 from the National Institutes ofHealth and by
the General Medical Research Service of the Veterans Administra-
1. Arai, S., H. Yamamoto, K. Itoh, and K. Kumagai. 1983. Sup-
pressive effect of human natural killer cells on pokeweed
mitogen-induced B cell differentiation. J. Immunol. 131:651-
2. Byrne, G. I., and D. A. Krueger. 1985. In vitro expression of
factor-mediated cytotoxic activity generated during the immune
response to Chlamydia
3. Djeu, J. Y., N. Stocks, K. Zoon, G. J. Stanton, T. Timonen, and
R. B. Herberman. 1982. Positive self regulation of cytotoxicity
in human natural killer cells by production of interferon upon
exposure to influenza and herpes viruses. J. Exp. Med. 156:
in the mouse.J. Immunol.
VOL. 55, 1987
WILLIAMS ET AL.
4. Gilbertson, S. M., P. D. Shah, and D. A. Rowlsey. 1986. Natural
killer cells suppress the generation of Lyt 2+ cytolytic T cells by
suppressing or eliminating dendritic
5. Hauser, W. E., and V. Tsai. 1986. Acute Toxoplasma infection
of mice induces spleen natural killer cells that are cytotoxic for
Toxoplasma gondii in vitro. J. Immunol. 136:313-319.
6. Hauser, W. E., Jr., S. D. Sharma, and J. S. Remington. 1982.
Natural killer cells induced by acute and chronic Toxoplasma
infection. Cell. Immunol. 69:330-346.
7. Herberman, R. B., and H. T. Holden. 1979. Natural cell-
mediated immunity. Adv. Cancer Res. 27:305-377.
8. Herberman, R. B., M. E. Nunn, and D. H. Lavrin. 1975. Natural
cytotoxic reactivity of mouse lymphoid cells against syngeneic
and allogeneic tumors. Int. J. Cancer 16:216-229.
9. Lammert, J. K. 1982. Cytotoxic cells induced after Chlamydia
psittaci infection in mice. Infect. Immun. 35:1011-1117.
10. Murphy, J. W., and D. 0. McDaniel. 1982. In vitro reactivity of
natural killer (NK) cells against Crypbococcus neoformans. J.
11. Nencioni, L., L. Villa, D. Boraschi, B. Berti, and A. Tagliabue.
1983. Natural and antibody-dependent cell mediated activity
against Salmonella typhimurium by peripheral and intestinal
lymphoid cells in mice. J. Immunol. 130:903-907.
12. Pavia, C. S., and J. Schachter. 1983. Failure to detect cell-
mediated cytotoxicity against Chlamvdia trachomatis-infected
cells. Infect. Immun. 39:1271-1274.
13. Quigstad, E., and R. Hirschberg. 1984. Lack of cell-mediated
cytotoxicity towards Chlamydia trachomatis-infected target
cells in humans. Acta Pathol. Microbiol. Immunol. Scand. Sect.
14. Talmadge, J. E., R. B. Herberman, M. A. Chirigos, A. E.
Malvish, M. A. Schneider, J. S. Adams, H. Philips, G. B.
Thurman, L. Varesio, C. Long, R. K. Oldham, and R. H.
Wiltrout. 1985. Hyporesponsiveness to augmentation of murine
natural killer cell activity in different anatomical compartments
by multiple injections of various immunomodulators including
recombinant interferons and interleukin-2. J. Immunol. 135:
15. Williams, D. M., and J. Schachter. 1985. Role of cell-mediated
immunity in chlamydial infection: implications for ocular immu-
nity. Rev. Infect. Dis. 7:754-759.
16. Williams, D. M., J. Schachter, J. J. Coalson, and B. Grubbs.
1984. Cellular immunity to the mouse pneumonitis agent. J.
Infect. Dis. 149:630-639.
17. Williams, D. M., J. Schachter, D. J. Drutz, and C. V. Sumaya.
1981. Pneumonia due to Chlamydia trachomatis in the immu-
nocompromised (nude) mouse. J. Infect. Dis. 143:238-241.