Essential Role of Bystander Cytotoxic CD122?CD8?T Cells
for the Antitumor Immunity Induced in the Liver of Mice by
Ryusuke Nakagawa,* Takuo Inui,‡Ikuko Nagafune,§Yoshiko Tazunoki,§Kazuhiro Motoki,§
Akira Yamauchi,* Mitsuomi Hirashima,†Yoshiko Habu,‡Hiroyuki Nakashima,‡and
We recently reported that NK cells and CD8?T cells contribute to the antimetastatic effect in the liver induced by ?-galacto-
sylceramide (?-GalCer). In the present study, we further investigated how CD8?T cells contribute to the antimetastatic effect
induced by ?-GalCer. The injection of anti-CD8 Ab into mice 3 days before ?-GalCer injection (2 days before intrasplenic
injection of B16 tumors) did not inhibit IFN-? production nor did it reduce the NK activity of liver mononuclear cells after
?-GalCer stimulation. However, it did cause a reduction in the proliferation of liver mononuclear cells and mouse survival time.
Furthermore, although the depletion of NK and NKT cells (by anti-NK1.1 Ab) 2 days after ?-GalCer injection no longer decreased
the survival rate of B16 tumor-injected mice, the depletion of CD8?T cells did. CD122?CD8?T cells in the liver increased after
?-GalCer injection, and antitumor cytotoxicity of CD8?T cells in the liver gradually increased until day 6. These CD8?T cells
exhibited an antitumor cytotoxicity toward not only B16 cells, but also EL-4 cells, and their cytotoxicity significantly decreased by
the depletion of CD122?CD8?T cells. The critical, but bystander role of CD122?CD8?T cells was further confirmed by adoptive
transfer experiments into CD8?T cell-depleted mice. Furthermore, it took 14 days after the first intrasplenic B16/?-GalCer
injection for the mice to generate CD8?T cells that can reject s.c. rechallenged B16 cells. These findings suggest that ?-GalCer
activates bystander antitumor CD122?CD8?T cells following NK cells and further induces an adaptive antitumor immunity due
to tumor-specific memory CD8?CTLs. The Journal of Immunology, 2004, 172: 6550–6557.
their TCR (6, 7), and their development is dependent on the MHC
class I-like molecule, CD1d (8). A glycolipid Ag, ?-galactosylce-
ramide (?-GalCer),2is a synthetic ligand of NKT cells and induces
NKT cells to produce IFN-? and IL-4 (9). Liver NKT cells of mice
acquired a potent antitumor cytotoxicity after IL-12 stimulation in
vivo and inhibit hematogenous tumor metastases in the liver, lung,
and kidney (2, 4, 10–14). However, liver NKT cells in mice in-
jected with ?-GalCer have been reported to rapidly disappear (15–
17), while, in addition, it was recently shown that NKT cells do not
disappear, but transiently down-regulate their TCR and NK1.1 Ag
(18, 19). We recently demonstrated that NK cells that are activated
by IFN-? produced by ?-GalCer-activated NKT cells are the main
and direct antimetastatic effector cells in the liver and reject the
liver metastases of intrasplenically (i.s.) injected tumors, while
ouse NK1.1 Ag?T cells (NKT cells) are abundant in
the liver (1–5). These cells use V?14J?281 gene prod-
uct combined with V?8 and V?7 gene products for
CD8?T cells are also required for the antimetastatic effect in the
liver and mouse survival, whereas activated NKT cells caused the
hepatocyte injury (20, 21). The mice that rejected liver metastases
of i.s. injected tumors are already resistant to the same s.c. rechal-
lenged tumor cells, while they could not reject s.c. inoculated other
tumors (22, 23). Furthermore, the inhibitory effect of ?-GalCer on
s.c. rechallenged tumors was mainly mediated by CD8?T cells
(22, 23). These findings suggested that NK and NKT cells might
be important for innate immunity against tumor metastasis, while
tumor-specific memory CD8?T cells may play an important role
in the inhibition of s.c. inoculated tumor growth.
?-GalCer stimulation reportedly induces an Ag-independent
proliferation of memory phenotype T cells (bystander prolifera-
tion) (24). An extensive T cell proliferation occurs during viral
infections in mice (25), and in mice injected with LPS (26), CpG
DNA (27), or poly(I:C) (25), and for the most part, these prolif-
erative responses are restricted to memory phenotype CD8?T
cells (28). The proliferations of these CD8?T cells stimulated with
poly(I:C) occurred in MHC class I-deficient hosts (using bone
marrow chimeras). As a result, although the proliferative response
of the memory phenotype CD8?T cells represented a MHC-in-
dependent bystander response, it may be an important immune
response to protect the hosts against infections (25).
In the present study, we show that ?-GalCer induces the by-
stander proliferation of memory phenotype CD122?CD8?T cells
with an antitumor function in the liver. These CD122?CD8?T
cells are tumor-nonspecific CTLs; however, they become more
critical antitumor effectors than NK cells as early as 2–3 days after
?-GalCer injection. As a result, ?-GalCer-induced antitumor im-
munity of NK cells triggered by IFN-? produced by NKT cells is
Departments of *Cell Regulation and†Immunology and Immunopathology, Kagawa
Medical University, Kagawa, Japan;‡Department of Microbiology, National Defense
Medical College, Tokorozawa, Japan; and§Pharmaceutical Research Laboratory, Ki-
rin Brewery, Takasaki, Japan
Received for publication October 20, 2003. Accepted for publication March 18, 2004.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1Address correspondence and reprint requests to Dr. Shuhji Seki, Department of
Microbiology, National Defense Medical College, Tokorozawa 359-8513, Japan. E-
mail address: email@example.com
2Abbreviations used in this paper: ?-GalCer, ?-galactosylceramide; i.s., intrasplenic;
MNC, mononuclear cell.
The Journal of Immunology
Copyright © 2004 by The American Association of Immunologists, Inc.0022-1767/04/$02.00
rapidly taken over by CD122?CD8?T cells, and thereafter, adap-
tive antitumor immunity may further be induced by tumor-specific
memory CD8?T cells.
Materials and Methods
This study was conducted according to the guidelines of the Institutional
Review Board for the Care of Animal Subjects at the National Defense
Medical College (Tokorozawa, Japan).
Mice and preparation of hepatic mononuclear cells (MNCs)
Male C57BL/6 (B6) mice at 6 wk of age were obtained from Nippon SLC
(Hamamatsu, Japan). Mice were maintained and fed under standard laboratory
conditions. Hepatic MNCs were prepared essentially as described (14). In
brief, the liver was passed through a stainless steel mesh, and the resulting
dissociated cells were suspended in HBSS, washed, resuspended in an isotonic
33% Percoll solution (Amersham Biosciences, Arlington Heights, IL) contain-
ing heparin (100 U/ml (Sigma-Aldrich, St. Louis, MO)), and centrifuged at
2000 rpm (500 ? g) for 15 min at room temperature. The resulting pellet was
resuspended in RBC lysis solution and then was washed twice in RPMI 1640
medium supplemented with 5% FCS.
?-GalCer, or (2S,3S,4R)-1-O-(?-D-galactopyranosyl)-2-(N-hexacosanoyl-
amino)-1,2,4-octadecanetriol (KRN7000), was synthesized in our labora-
tory (29). The original solution of ?-GalCer (220 ?g/ml) was prepared with
0.5% polysorbate 20 (Nikko Chemicals, Tokyo, Japan) in saline, and then
was subsequently diluted with this solution (vehicle) or with saline before
i.v. injection at a dose of 100 ?g/kg body mass.
Flow cytometric analysis
The surface phenotypes of liver MNCs were characterized by a four-color
flow cytometric analysis. FITC-conjugated anti-TCR?? mAb (H57-597,
hamster IgG), FITC-conjugated anti-CD8 mAb (53-6.7, rat IgG2a), PE-
conjugated anti-NK1.1 mAb (PK136, mouse IgG2a), PE-conjugated anti-
CD122 (IL-2R?) mAb (TM-?1, rat IgG2b), allophycocyanin-conjugated
anti-CD44 mAb (IM7, rat IgG2b), and allophycocyanin-conjugated anti-
CD3 mAb (145-2C11, hamster IgG) were purchased from BD PharMingen
(San Diego, CA). Before staining with Abs, the MNCs were incubated for
10 min with Fc blocker (2.4 G2; BD PharMingen) to prevent any nonspe-
cific binding. Flow cytometry was performed with the FACSCalibur device
(BD Biosciences, San Jose, CA).
In vivo and in vitro cell depletion
Anti-CD4 mAb (L3T4), anti-CD8 mAb (Lyt-2.2), anti-NK1.1 mAb, and
anti-CD122 (IL-2R?) mAb were derived from GK1.5, 2.43, PK136, and
TM-?1 hybridoma cells, respectively, and each Ab was prepared from mice
ascites fluid (IBL, Gunma, Japan). We previously showed that a single i.v.
injection of an optimal dose of anti-NK1.1 Ab depleted both NK and NKT
anti-CD8 mAb, or anti-CD122 mAb, or 200 ?g of anti-NK1.1 mAb. For the
in vitro depletion of CD8?CD122?T cells, whole liver MNCs were stained
with FITC-conjugated anti-CD8 mAb and PE-conjugated anti-CD122 mAb,
Pharmaceutical, Miami, FL) and the purity of the sorted cells was checked by
Epics XL (Coulter Pharmaceutical).
NK cell-sensitive YAC-1 lymphoma cells, EL-4 lymphoma cells, and B16
melanoma cells (both of B6 origin) were used as target cells. Target cells
(3 ? 106) were labeled for 60 min at 37°C with 100 ?Ci of Na2
500 ?l of RPMI 1640 supplemented with 10% FCS. They were then
washed three times with medium alone and subjected to a cytotoxicity
assay. Labeled targets (2 ? 103cells/well) were incubated for 4 h at 37°C
in 96-well round-bottom microtiter plates containing RPMI 1640 (total
volume of 100 ?l) with liver MNCs obtained from mice injected 24 h
previously with ?-GalCer (100 ?g/kg). The plates were then centrifuged,
and the resulting supernatants were harvested and their radioactivity was
determined with a gamma counter. Cytotoxicity was calculated as the per-
centage of released radioactivity after correcting for spontaneous release,
which was ?15% of maximal release. In some experiments, liver MNCs or
sorted liver CD8?T cells were obtained from mice injected with ?-GalCer
6 days before, and cells were incubated with target cells for 16 h and
determine the role of CD8?T cells in the cytotoxicity of liver MNCs.
To determine the proliferation of the MNCs (2 ? 105cells/200 ?l) stim-
ulated with ?-GalCer, the cells were pulsed with 0.5 ?Ci/well [3H]thymi-
dine ([3H]TdR) 12 h before the cells were harvested. The radioactivities of
the harvested cells at the indicated culture time points were assessed by the
liquid scintillation counting method.
Measurement of serum IFN-?
The peripheral blood of individual mice was collected at the indicated time
points from the retro-orbital sinus. The serum concentration of IFN-? was
measured by using a cytokine-specific ELISA kit (Endogen, Boston, MA).
B16 model of hepatic metastasis
Hepatic metastases of B16 tumor cells were produced, as previously de-
scribed (22). In brief, the spleen of anesthetized mice was exposed to allow
the direct i.s. injection of 3 ? 106B16 cells in 0.1 ml of medium. The
spleen was then removed after clamping of the artery and vein, and the
abdomen and skin were surgically sutured. Using this method, B16 tumor
cells almost exclusively metastasized to the liver, but not to other organs.
Adoptive transfer of purified CD8?T cells
Liver CD8?T cells were positively purified by magnetic cell sorting (MACS
system; Miltenyi Biotec, Bergisch Gladbach, Germany) from ?-GalCer-in-
jected mice with or without anti-CD122 mAb pretreatment. Briefly, liver
MNCs were stained with magnetic bead-conjugated anti-CD8 mAb for 20 min
at 4°C, and then were washed twice in the medium. Next, they were resus-
pended in 500 ?l of the medium, and transferred onto a separation column for
Biotec). Any adherent CD8?MNCs in the column were washed with 500 ?l
of the medium and collected using a plunger when the column was removed
from the unit. The collected cells were resuspended in PBS, and 2 ? 106
CD8?cells (200 ?l)/mouse were injected i.v. into B16-inoculated mice that
had been injected with anti-CD8 Ab.
All data are expressed as the means ? SD, and differences among groups
were analyzed by the Mann-Whitney U test using Stat View software. Then
mouse survival rates were analyzed by the log-rank test. A p value of
?0.05 was considered to be statistically significant.
Depletion of CD8?T cells or CD4?T cells affects neither
cytotoxic activity of liver NK cells nor production of IFN-? in mice
injected with ?-GalCer, whereas depletion of CD8?T cells
markedly decreased the proliferative response of liver MNC
We and other researchers recently reported that the antitumor cy-
totoxicity of liver NK cells and serum IFN-? levels remarkably
increased after an ?-GalCer injection (21, 30, 31). The pretreat-
ment of mice with anti-CD4 or anti-CD8 Ab did not reduce either
the cytotoxicity of the liver MNCs or serum IFN-? levels in ?-Gal-
Cer-injected mice (Fig. 1A, a and b). Although an injection of
either anti-CD8 Ab or anti-CD4 Ab depleted CD8?T cells and
CD4?T cells, respectively (Fig. 1B, right panels), anti-CD4 Ab
injection unexpectedly increased the proportion of NK1.1?cells to
43.1% (Fig. 1B, left middle panel), and the proportion of NKT
cells was not substantially affected (or even slightly increased)
(Fig. 1B, middle central panel). CD4?NKT cells presumably
down-regulated surface expression of CD4 molecule after anti-
CD4 Ab injection (Fig. 1B, left middle panel, indicated by an ar-
row). Although ?60% of NKT cells are CD4?, they have been
shown to express CD4 only at a low level (10, 32). These findings
may explain why the cytotoxic activity, IFN-? production, and
proportion of NKT cells of liver MNC did not decrease by CD4 Ab
However, the proliferation of liver MNCs of anti-CD8 Ab-pre-
treated mice stimulated with ?-GalCer in vitro decreased signifi-
cantly (Fig. 1Ac). Sorting experiments revealed that depletion of
CD8?CD122?cells from whole liver MNCs also greatly decreased
6551The Journal of Immunology
the proliferation of liver MNCs stimulated with ?-GalCer, while de-
mainly CD8?CD122?cells proliferated, while CD8?CD122?cells
did not. Consistently, no remarkable change was seen in the prolif-
presumably because spleen MNCs contain a much smaller population
of CD8?CD122?cells than liver MNC (5).
Treatment of mice with anti-CD8 Ab decreases an antimetastatic
effect induced by ?-GalCer after B16 inoculation
To examine when NK cells or CD8?T cells are crucial to reject
tumors, we injected each Ab at various time points, as indicated in
Fig. 2A. The pretreatment of mice with either anti-CD8 Ab or
anti-NK1.1 Ab-reduced ?-GalCer induced a prolongation of sur-
vival time of B16-inoculated mice. The treatment of mice with
anti-CD8 Ab on day 3 also decreased the survival time (Fig. 2B),
whereas the treatment of anti-NK1.1 Ab on day 3 hardly inhibit the
?-GalCer-induced effect (Fig. 2C). Furthermore, the treatment of
anti-CD8 Ab on day 8 slightly reduced the survival rate of the
mice. These findings suggest that liver CD8?T cells are more
important antitumor effectors than NK cells in the liver beyond 3
days after ?-GalCer injection.
CD8?T cells acquired antitumor cytotoxicity against B16 cells
with or without B16 inoculation
Although anti-CD8 Ab pretreatment did not decrease the cytotox-
icity of liver MNCs 1 day after the ?-GalCer injection (Fig. 1A),
the liver MNCs prepared from anti-CD8 Ab-treated mice (3 days
after ?-GalCer injection) 6 days after ?-GalCer injection exhibited
a decreased antitumor cytotoxicity compared with that of liver
MNCs prepared from mice without Ab treatment (Fig. 3A). However,
because the treatment of mice with both anti-CD8 Ab and anti-NK1.1
Ab completely inhibited the cytotoxicity toward B16 (Fig. 3A) NK
cells may still demonstrate a cytotoxicity toward B16 at this time
point. Furthermore, isolated CD8?T cells in the liver MNCs of mice
with or without B16 inoculation at 6 days after ?-GalCer injection
exhibited similar up-regulated antitumor cytotoxicities toward B16
(Fig. 3B), thus suggesting that the killing activity of these CD8?T
cells may not be B16 specific.
Ab or anti-CD8 Ab 2 days previously, and 18 h after ?-GalCer (100 ?g/kg, i.v.) injection, liver MNCs were isolated and the cytotoxicities were determined. Ab,
Serum specimens were collected from mice from retro-orbital plexus at the indicated time points, and the amounts of IFN-? were determined by ELISA.
Ac, Mice were injected with anti-CD4 Ab or anti-CD8 Ab 2 days previously. MNCs were isolated from the liver and spleen, and cultured with or without
?-GalCer (100 ng/ml) in a 96-well round-bottom plate. The proliferations of MNCs were examined after 2-day culture. Data are from an experiment that
was repeated three times with similar results. d, MNCs were isolated from the liver and stained with anti-CD8 and anti-CD122. CD8?CD122?T cells were
depleted by sorting. Sorted MNCs and whole MNC were cultured with ?-GalCer (100 ng/ml) in a 96-well round-bottom plate. The proliferations were
examined after 2-day culture. B, Mice were injected with anti-CD4 Ab or anti-CD8 Ab 2 days before MNC preparation. MNCs were isolated from the liver
and stained with anti-CD4, anti-CD3, anti-CD8, and anti-NK1.1. An arrow shows the population of presumably CD4 down-regulated NKT cells.
Effect of the depletion of either CD8?T cells or CD4?T cells in mice on the ?-GalCer-induced effect. Aa, Mice were injected with anti-CD4
6552 BYSTANDER ANTITUMOR CD8?T CELLS INDUCED BY ?-GalCer
Time course analysis of the activation of CD8?T cells
The cytotoxicity of the liver CD8?T cells prepared from ?-Gal-
Cer-injected mice (without B16 inoculation) gradually increased
and exhibited a maximum cytotoxicity toward B16 as well as
YAC-1 cells at 6 days after ?-GalCer injection (Fig. 4A). In ad-
dition, the population of liver CD8?T cells increased (Fig. 4B)
and continuously maintained an activating profile like IL-2R?
(CD122) and CD69 expression after ?-GalCer injection (Fig. 4C).
CD122?CD8?T cells at every time point showed lower TCR/
CD3 intensities than those of CD122?CD8?T cells (data not
Depletion of CD122?CD8?T cells reduced the antitumor
cytotoxicity of liver CD8?T cells
The administration of anti-CD122 Ab has been reported to deplete
CD122?cells, such as NK, NKT, and CD122?CD8?T cells (5,
33). To further confirm such depletion, we analyzed the expression
of CD44 because most of CD122?CD8?T cells are also known to
have a high expression of CD44 (34, 35). CD44highcells in liver
CD8?T cells of ?-GalCer- or vehicle-injected mice (without B16
inoculation) remarkably decreased by the subsequent injection of
anti-CD122 Ab (Fig. 5A). Liver CD8?T cells prepared from mice
at 6 days after ?-GalCer injection (without B16 inoculation) ex-
hibited the antitumor cytotoxicity against either B16 or EL-4 cells;
however, their cytotoxicity was suppressed by the depletion of the
CD122?CD8?T cells, but not CD122?CD8?T cells, prepared
from ?-GalCer-injected mice 6 days previously exhibited cytotox-
icity in an Ag-independent bystander-activated manner.
Ab (Fig.5B). Hence,
Anti-CD122 Ab treatment inhibited ?-GalCer-induced
antimetastatic effect, and adoptive transfer of ?-GalCer-
activated CD8?T cells without CD122?fraction did not exhibit
an antimetastatic effect
The depletion of NK and NKT cells with anti-NK1.1 Ab 3 days
after ?-GalCer injection (4 days after B16 injection) did not affect
the survival time of mice inoculated with B16 (Figs. 2B and 6A).
However, the survival time of B16-inoculated mice greatly re-
duced either by anti-CD122 Ab or anti-CD8 Ab injection at 3 days
after ?-GalCer injection (Fig. 6A). As a result, although NK cells
may have cytotoxicity against B16 cells, as suggested by the re-
sults in Fig. 3A, CD122?CD8?T cells rather than NK cells con-
tributed to the inhibition of B16 metastases beyond 3 days after
?-GalCer injection. Furthermore, to confirm the bystander anti-
metastatic activity of CD122?CD8?T cells, we conducted the
adoptive transfer of ?-GalCer-activated CD8?T cells with or
without CD122?CD8?T fraction (Fig. 6B). The adoptive transfer
of CD8?T cells of donor type 1 mice that were treated with
?-GalCer (6 days before) (without B16 inoculation) rescued B16-
inoculated recipient mice that were already pretreated with anti-
CD8 Ab and ?-GalCer. However, the adoptive transfer of CD8?T
cells prepared from donor type 2 mice that were depleted of the
CD122?fraction only slightly prolonged the survival time of B16-
inoculated mice (Fig. 6C). These findings revealed that the trans-
ferred CD122?CD8?T cells activated by ?-GalCer contributed to
the rejection of B16 cells without the prior sensitization by B16
cells (bystander antimetastatic activity), and that CD122?CD8?T
CD8 Ab (PBS as a control) on the ?-GalCer-induced prolongation of
mouse survival rate. A, The mice were i.s. inoculated with B16 cells and
injected with ?-GalCer (100 ?g/kg, i.v.) or vehicle and received the Ab
injection at the indicated time points before/after tumor inoculation. B,
Survival of anti-CD8 Ab-treated mice was evaluated at the indicated days
after injection of tumor cells. ?, p ? 0.05 vs day 3 and day 2 Ab-treated
groups. C, The survival of anti-NK1.1 Ab-treated mice was evaluated at the
indicated days after injection of tumor cells. All experimental mice groups
in B and C included five to eight mice. ?, p ? 0.05 vs day 2 Ab-treated
Effect of the treatment with either anti-NK1.1 Ab or anti-
mal mice were i.s. inoculated with B16 cells and then were injected with
?-GalCer (100 ?g/kg, i.v.) 24 h after tumor injection. Three days after
?-GalCer injection, some mice were injected with anti-CD8 Ab or both
CD8 Ab and NK1.1 Ab. Six days after ?-GalCer injection, liver MNCs
were isolated and incubated with51Cr-labeled B16 cells for 16 h, and
cytotoxic activities were determined. The data are the means ? SD of
values. B, Mice were i.s. inoculated with B16 cells and then were injected
with ?-GalCer (100 ?g/kg, i.v.) or vehicle 24 h after tumor injection. At
the same time, normal mice were injected with ?-GalCer (100 ?g/kg, i.v.)
or vehicle. Six days after ?-GalCer injection, liver MNCs were isolated
from the mice and CD8?T cells were purified by the magnet sorting. The
purities of sorted cells were ?95%, and the cytotoxic activity of sorted
liver CD8?T cells was determined. The data are the means ? SD of values
from five mice. All data are from experiments that were repeated three
times with similar results.
The antitumor cytotoxicity of liver CD8?T cells. A, Nor-
6553The Journal of Immunology
cells could not transfer the antimetastatic effect. These findings
also suggest that CD122?CD8?T cells could not substitute
CD122?CD8?T cells, and they are therefore essentially different
Tumor-specific CD8?CTL exhibited their function 14 days after
tumor inoculation and ?-GalCer injection
We previously showed that ?-GalCer injection induced tumor-spe-
cific memory immunity (22, 23). B16 i.s. inoculated mice could
reject liver-metastasized B16 cells by ?-GalCer treatment, and
these survived mice could also reject s.c. rechallenged B16 cells,
but could not reject s.c. challenged EL-4 cells (36). Mice that
survived i.s. inoculated with B16 cells were selected and pretreated
with various Abs and were thereafter s.c. inoculated with B16. The
results showed that the tumor-specific memory immunity was
largely due to the CD8?T cells, but the NK1.1?cells (NK and
NKT cells) contributed very little (Fig. 7A). To confirm when
memory CTL generate and to reject specific tumor cells, we s.c.
inoculated B16 cells to mice at various time points that were al-
ready i.s. inoculated with B16 cells and injected with ?-GalCer
(Fig. 7B). The ?-GalCer injection did not cause a regression of the
s.c. inoculated B16 cells early after first B16 i.s. inoculation (days
1–7), while the most mice could reject s.c. inoculated B16 cells 14
days after first B16 i.s. inoculation (Fig. 7C). However, all mice
(n ? 5) that had received the i.s. B16 injection 14 days before
could not reject the s.c. inoculated EL-4 cells (data not shown).
These results suggest that it takes 14 days for tumor-specific
CD8?CTLs to be generated, and that the CD122?CD8?T cells
that expanded within 1 wk after ?-GalCer injection in the liver are
not tumor-specific CTLs.
The depletion of NK and NKT cells 2–3 days after ?-GalCer injec-
tion no longer inhibited antimetastatic effect of ?-GalCer against B16
tumor cells in the liver, whereas the depletion of CD8?T cells at this
time point significantly reduced the ?-GalCer-induced antimetastatic
effect and mouse survival time. CD8?CD122?T cells were found to
be critical for the proliferation of liver MNCs after ?-GalCer injec-
tion. Adoptive transfer experiments also showed CD8?CD122?T
cells, but not CD8?CD122?T cells, to be responsible for the
antitumor function in the liver, suggesting that CD8?CD122?T
cells and CD8?CD122?T cells are essentially different popula-
tions. These results suggest that although NK cells in the liver are
important antitumor effectors in the early phase after ?-GalCer
injection, CD8?CD122?T cells thereafter are more critical anti-
tumor effectors. However, these CD8?CD122?T cells expanded
within 1 wk appeared to be distinct from tumor-specific memory
CTLs and are bystander CTLs. It took 14 days for ?-GalCer to
generate/differentiate tumor-specific memory CD8?T cells.
Although we recently showed that liver NK cells proportionally
increased until 5 days after ?-GalCer injection (21), the present
study revealed that NK cells no longer play an essential role in the
antimetastatic effect in the liver as early as 2–3 days after ?-GalCer
injection, even though they still have antitumor cytotoxicity. The
cells after ?-GalCer injection. A, Cytotoxic activities of liver CD8?T cells
against B16 cells at the indicated time points after ?-GalCer injection.
Normal mice were injected with ?-GalCer (100 ?g/kg, i.v.) or vehicle. At
the indicated time points after ?-GalCer injection, liver MNCs were iso-
lated and the liver CD8?T cells were sorted. The purities of the sorted cells
were ?95%, and the cytotoxic activity of sorted liver CD8?T cells was
determined. B, The percentage of CD8?T cells in the liver MNCs. Normal
mice were injected with ?-GalCer (100 ?g/kg, i.v.) or vehicle. At the
indicated time points after ?-GalCer injection, liver MNCs were isolated
and stained with anti-CD3 and anti-CD8. C, CD122 and CD69 expressions
of liver CD8?T cells were examined at indicated time points after ?-Gal-
Cer injection. The data are from experiments that were repeated three times
with similar results.
The time course analysis of the activation of liver CD8?T
CD44 expression (A) and cytotoxic activity of liver MNCs (B). Normal
mice were injected with ?-GalCer or vehicle, and then anti-CD122 Ab was
injected 3 days after ?-GalCer injection. After an additional 3 days, liver
MNCs were isolated from the mice and the liver CD8?T cells were sorted.
The purities of sorted cells were ?95%; A, FACS analysis of sorted liver
CD8?T cells were conducted. B, The cytotoxicity of liver CD8?T cells
was determined in vitro with B16 cells (left panel) or EL-4 cells (right
panel). The data are from experiments that were repeated three times with
Effect of the injection of anti-CD122 Ab ?-GalCer on the
6554BYSTANDER ANTITUMOR CD8?T CELLS INDUCED BY ?-GalCer
antimetastatic role in the liver MNCs was taken from NK cells by
CD122?CD8?T cells at this time point. As a result, coordination
of antimetastatic function of the liver MNCs rapidly occurs among
NKT, NK, and CD122?CD8?T cells.
Antitumor cytotoxicity of CD122?CD8?T cells expanded
within 1 wk after B16 and ?-GalCer injection was not B16 specific
because their cytotoxicity was also exhibited toward EL-4 cells,
and their cytotoxicity against B16 cells could not be blocked by the
treatment of B16 cells with anti-H-2D and anti-H-2K Ab (data not
shown). In addition, the adoptive transfer of CD8?T cells from
mice injected with ?-GalCer 6 days before (but without B16 in-
jection) similarly inhibited liver metastasis of B16 cells and pro-
longed the mouse survival time. Moreover, the mice that received
the i.s. B16 tumor injection and ?-GalCer treatment could not
reject rechallenged s.c. inoculated B16 tumors until 10–14 days
after the first i.s. B16 injection. The CD122?CD8?T cells that
thus expanded within 1 wk by ?-GalCer stimulation are not likely
B16-specific CD8?T cells that can inhibit s.c. rechallenged B16
tumors. It is generally accepted that it takes 2 wk after an Ag
challenge for the development of either Ag-specific Ab or Ag-
specific CTL, and their manner of recognition by Ab or by TCR is
known to be highly restricted. From the view of kinetics and the
recognition of targets, CD122?CD8?T cells induced by ?-GalCer
injection within 1 wk are thus different from tumor-specific mem-
ory CTLs and are bystander CTLs.
?-GalCer injection was recently reported to induce the expan-
sion of CD122?CD8?T cells, and this phenomenon is called by-
stander proliferation because of their Ag independency (24). Viral
infections and IFN-?? (25) or IL-15 (34) also reportedly increase
these bystander CD122?CD8?T cells, and these cells have been
called memory phenotype CD8?T cells as distinguished from
memory CD8?T cells (34, 37, 38). LPS (26) and poly(I:C), both
of which are activators of NK cells, also reportedly induce the
proliferation of CD122?CD8?T cells in vivo in mice (34). We
previously reported that CD122?CD8?T cells secreted a much
larger amount of IFN-? and acquired a more potent antitumor cy-
totoxicity by anti-CD3 stimulation than did CD122?CD8?T cells
(39). CD122?CD8?T cells were abundant in the liver, and ?-Gal-
Cer injection further expanded these cells. As a result, bystander
CD122?CD8?T cells may be important as a Th1 effector popu-
lation in both antitumor and anti-infection immunity and may cre-
ate a bridge from innate immunity to adaptive immunity.
rate of mice inoculated with B16 cells. Three days after ?-GalCer injection
(4 days after B16 inoculation), B6 mice were injected (i.p.) with anti-
NK1.1 Ab, anti-CD8 Ab, or anti-CD122 Ab (200 ?g) (PBS as a control).
B, Schedule of the adoptive transfer of CD8?T cells with/without CD122?
fraction. Recipient mice were treated with anti-CD8 Ab before B16 and
?-GalCer injection at the indicated time point. Donors were normal mice
that were not injected with B16 cells. CD8?T cells prepared from donor
type 1 mice contained CD122?fraction, while CD8?T cells prepared from
donor type 2 mice lacked CD122?fraction (indicated by the flow cyto-
metric profiles). These CD8?T cells were adoptively transferred to recip-
ient mice twice at indicated time points (day 1 (3 h after ?-GalCer injec-
tion) and day 5). C, Effect of adoptive transfer of ?-GalCer-activated CD8?
T cells with (donor type 1) or without CD122?fraction (donor type 2).
Survival was evaluated at the indicated time points after B16 inoculation.
All experimental mice groups in A and C included five to eight mice. ?$,
p ? 0.05 vs group transferred from donor type 2 mice.
A, Effect of the treatment with various Abs on the survival
s.c. rechallenged with B16 tumor cells. Mice that had already rejected i.s.
inoculated B16 cells were s.c. rechallenged with B16 cells. Various Ab
were treated weekly from 2 days before tumor rechallenge, and survival
was evaluated at the indicated times after B16 inoculation. All experimen-
tal mice groups included five to eight mice. ?, p ? 0.05 vs anti-CD8
Ab-treated group. B, The schedule of s.c. inoculation of B16 cells to mice
that had received an i.s. inoculation of B16 and ?-GalCer treatment. The
B16 (i.s.)- and ?-GalCer-injected mice were s.c. rechallenged with B16
cells at the indicated days after first performing the i.s. inoculation of B16
cells (day 0). C, The growth of s.c. rechallenged B16 cells in mice. The
mice were s.c. rechallenged with B16 cells, as shown in B. The control
group (without previous i.s. B16 inoculation) included six mice, while the
other groups included five mice each.
A, Effects of various Ab treatment on the survival of mice
6555The Journal of Immunology
It is unlikely that conventional CD122?CD8?T cells are acti-
vated by ?-GalCer and became CD122?CD8?T cells because
depletion of CD122?CD8?T cells from liver MNCs greatly cause
a reduction in the proliferation of liver MNCs. In addition, the
depletion of CD8?T cells did not significantly reduce ?-GalCer-
induced proliferation of the spleen MNCs in which CD122?CD8?
T cells are predominant and there were fewer CD122?CD8?T
cells than in the liver MNCs (3, 5). Furthermore, the finding that
CD122?CD8?T cells could not inhibit tumor metastasis, as evi-
denced by adoptive transfer experiments, supports that activating
cells are CD122?CD8?T cells, but not CD122?CD8?T cells,
and these cells are essentially different populations.
It has been proposed that CD122?CD8?T cells may develop
independently of the thymus in the liver and also other sites be-
cause these cells increase age dependently in parallel with the thy-
mus involution in normal mice as well as in athymic nude mice (5,
40–43). In addition, CD122?CD8?T cells emerged early in the
livers of mice before thymus reconstitution in radiation bone mar-
row chimera mice (44). The report that CD122?CD8?T cells
were positively selected by H-Y Ag outside of the thymus in H-Y
Ag transgenic mice suggests that some CD122?CD8?T cells are
dependent on autologous Ags for their development and expansion
(40, 41) and may be autoreactive. These findings together with
present results suggest that thymus-independent T cells are in-
cluded in bystander CD122?CD8?T cells and expand against
infections and tumors.
Finally, it should be noted that although ligand-activated NKT
cells induce severe hepatocyte damage in aged mice through Fas/
Fas-ligand system (21), we have recently found that anti-TNF-?
Ab pretreatment of mice completely inhibited liver injury without
attenuating antitumor and antimetastatic effect of ?-GalCer.3
In conclusion, ?-GalCer induces the activation of bystander
CD122?CD8?T cells following the activation of NKT and NK
cells, and finally, tumor-specific memory CD8?T cells develop
?14 days after tumor/?-GalCer injection.
We thank Hitomi Tomura for her help in experiments.
1. Ohteki, T., and H. R. MacDonald. 1994. Major histocompatibility complex class
I related molecules control the development of CD4?8?and CD4?8?subsets of
natural killer 1.1?T cell receptor-?/??cells in the liver of mice. J. Exp. Med.
2. Hashimoto, W., K. Takeda, R. Anzai, K. Ogasawara, H. Sakihara, K. Sugiura,
S. Seki, and K. Kumagai. 1995. Cytotoxic NK1.1 Ag??? T cells with interme-
diate TCR induced in the liver of mice by IL-12. J. Immunol. 154:4333.
3. Watanabe, H., C. Miyaji, Y. Kawachi, T. Iiai, K. Ohtsuka, T. Iwanage,
H. Takahashi-Iwanaga, and T. Abo. 1995. Relationships between intermediate
TCR cells and NK1.1?T cells in various immune organs: NK1.1?T cells are
present within a population of intermediate TCR cells. J. Immunol. 155:2972.
4. Seki, S., Y. Habu, T. Kawamura, K. Takeda, H. Dobashi, T. Ohkawa, and
H. Hiraide. 2000. The liver as a crucial organ in the first line of host defense: the
roles of Kupffer cells, natural killer (NK) cells and NK1.1 Ag?T cells in T helper
1 immune responses. Immunol. Rev. 174:35.
5. Tsukahara, A., S. Seki, T. Iiai, T. Moroda, H. Watanabe, S. Suzuki, T. Tada,
H. Hiraide, K. Hatakeyama, and T. Abo. 1997. Mouse liver T cells: their change
with aging and in comparison with peripheral T cells. Hepatology 26:301.
6. Makino, Y., H. Koseki, Y. Adachi, T. Akasaka, K. Tsuchida, and M. Taniguchi.
1994. Extrathymic differentiation of a T cell bearing invariant V?14J?281 TCR.
Int. Rev. Immunol. 11:31.
7. Lantz, O., and A. Bendelac. 1994. An invariant T cell receptor ? chain is used by
a unique subset of major histocompatibility complex class I-specific CD4?and
CD4?8?T cells in mice and humans. J. Exp. Med. 180:1097.
8. Bendelac, A. 1995. Positive selection of mouse NK1?T cells by CD1-expressing
cortical thymocytes. J. Exp. Med. 182:2091.
9. Kawano, T., J. Cui, Y. Koezuka, I. Toura, Y. Kaneko, K. Motoki, H. Ueno,
R. Nakagawa, H. Sato, E. Kondo, et al. 1997. CD1d-restricted and TCR-mediated
activation of V?14 NKT cells by glycosylceramides. Science 278:1626.
10. Takeda, K., S. Seki, K. Ogasawara, R. Anzai, W. Hashimoto, K. Sugiura,
M. Takahashi, M. Satoh, and K. Kumagai. 1996. Liver NK1.1?CD4??? T cells
activated by IL-12 as a major effector in inhibition of experimental tumor me-
tastasis. J. Immunol. 156:3366.
11. Seki, S., W. Hashimoto, K. Ogasawara, M. Satoh, H. Watanabe, Y. Habu,
H. Hiraide, and K. Takeda. 1997. Antimetastatic effect of NK1?T cells on
experimental haematogenous tumour metastases in the liver and lungs of mice.
12. Cui, J., T. Shin, T. Kawano, H. Sato, E. Kondo, I. Toura, Y. Kaneko, H. Koseki,
M. Kanno, and M. Taniguchi. 1997. Requirement for V?14 NKT cells in IL-12-
mediated rejection of tumors. Science 278:1623.
13. Kawamura, T., K. Takeda, S. K. Mendiratta, H. Kawamura, L. Van Kaer,
H. Yagita, T. Abo, and K. Okumura. 1998. Critical role of NK1?T cells in
IL-12-induced immune responses in vivo. J. Immunol. 160:16.
14. Ogasawara, K., K. Takeda, W. Hashimoto, M. Satoh, R. Okuyama, N. Yanai,
M. Obinata, K. Kumagai, H. Takada, H. Hiraide, and S. Seki. 1998. Involvement
of NK1?T cells and their IFN-? production in the generalized Shwartzman
reaction. J. Immunol. 160:3522.
15. Osman, Y., T. Kawamura, T. Naito, K. Takeda, L. Van Kaer, K. Okumura, and
T. Abo. 2000. Activation of hepatic NKT cells and subsequent liver injury fol-
lowing administration of ?-galactosylceramide. Eur. J. Immunol. 30:1919.
16. Leite-de-Moraes, M. C., A. Herbelin, C. Gouarin, Y. Koezuka, E. Schneider, and
M. Dy. 2000. Fas/Fas ligand interactions promote activation-induced cell death of
NK T lymphocytes. J. Immunol. 165:4367.
17. Matsuda, J. L., O. V. Naidenko, L. Gapin, T. Nakayama, M. Taniguchi,
C. R. Wang, Y. Koezuka, and M. Kronenberg. 2000. Tracking the response of
natural killer T cells to a glycolipid antigen using CD1d tetramers. J. Exp. Med.
18. Crowe, N. Y., A. P. Uldrich, K. Kyparissoudis, K. J. Hammond, Y. Hayakawa,
S. Sidobre, R. Keating, M. Kronenberg, M. J. Smyth, and D. I. Godfrey. 2003.
Glycolipid antigen drives rapid expansion and sustained cytokine production by
NK T cells. J. Immunol. 171:4020.
19. Wilson, M. T., C. Johansson, D. Olivares-Villagomez, A. K. Singh, A. K. Stanic,
C. R. Wang, S. Joyce, M. J. Wick, and L. Van Kaer. 2003. The response of
natural killer T cells to glycolipid antigens is characterized by surface receptor
down-modulation and expansion. Proc. Natl. Acad. Sci. USA 100:10913.
20. Inui, T., R. Nakagawa, S. Ohkura, Y. Habu, Y. Koike, K. Motoki, N. Kuranaga,
M. Fukasawa, N. Shinomiya, and S. Seki. 2002. Age-associated augmentation of
the synthetic ligand-mediated function of mouse NK1.1 Ag?T cells: their cy-
tokine production and hepatotoxicity in vivo and in vitro. J. Immunol. 169:6127.
21. Nakagawa, R., I. Nagafune, Y. Tazunoki, H. Ehara, H. Tomura, R. Iijima,
K. Motoki, M. Kamishohara, and S. Seki. 2001. Mechanisms of the antimetastatic
effect in the liver and of the hepatocyte injury induced by ?-galactosylceramide
in mice. J. Immunol. 166:6578.
22. Nakagawa, R., K. Motoki, H. Nakamura, H. Ueno, R. Iijima, A. Yamauchi,
S. Tsuyuki, T. Inamoto, and Y. Koezuka. 1998. Antitumor activity of ?-galac-
tosylceramide, KRN7000, in mice with EL-4 hepatic metastasis and its cytokine
production. Oncol. Res. 10:561.
23. Nakagawa, R., K. Motoki, H. Ueno, R. Iijima, H. Nakamura, E. Kobayashi,
A. Shimosaka, and Y. Koezuka. 1998. Treatment of hepatic metastasis of the
colon26 adenocarcinoma with an ?-galactosylceramide, KRN7000. Cancer Res.
24. Eberl, G., P. Brawand, and H. R. MacDonald. 2000. Selective bystander prolif-
eration of memory CD4?and CD8?T cells upon NK T or T cell activation.
J. Immunol. 165:4305.
25. Tough, D. F., P. Borrow, and J. Sprent. 1996. Induction of bystander T cell
proliferation by viruses and type I interferon in vivo. Science 272:1947.
26. Tough, D. F., S. Sun, and J. Sprent. 1997. T cell stimulation in vivo by lipo-
polysaccharide (LPS). J. Exp. Med. 185:2089.
27. Sprent, J., X. Zhang, S. Sun, and D. Tough. 2000. T-cell proliferation in vivo and
the role of cytokines. Philos. Trans. R. Soc. London B Biol. Sci. 355:317.
28. Sprent, J. 2003. Turnover of memory-phenotype CD8?T cells. Microbes Infect.
29. Kobayashi, E., K. Motoki, T. Uchida, H. Fukushima, and Y. Koezuka. 1995.
KRN7000, a novel immunomodulator, and its antitumor activities. Oncol. Res.
30. Carnaud, C., D. Lee, O. Donnars, S. H. Park, A. Beavis, Y. Koezuka, and
A. Bendelac. 1999. Cutting edge: cross-talk between cells of the innate immune
system: NKT cells rapidly activate NK cells. J. Immunol. 163:4647.
31. Eberl, G., and H. R. MacDonald. 2000. Selective induction of NK cell prolifer-
ation and cytotoxicity by activated NKT cells. Eur. J. Immunol. 30:985.
32. Seki, S., D. H. Kono, R. S. Balderas, and A. N. Theofilopoulos. 1994. V? rep-
ertoire of murine hepatic T cells: implication for selection of double negative
???T cells. J. Immunol. 153:637.
33. Toyabe, S., S. Seki, T. Iiai, K. Takeda, K. Shirai, H. Watanabe, H. Hiraide,
M. Uchiyama, and T. Abo. 1997. Requirement of IL-4 and liver NK1?T cells for
concanavalin A-induced hepatic injury in mice. J. Immunol. 159:1537.
34. Zhang, X., S. Sun, I. Hwang, D. F. Tough, and J. Sprent. 1998. Potent and
selective stimulation of memory-phenotype CD8?T cells in vivo by IL-15. Im-
35. Ohteki, T., R. Okuyama, S. Seki, T. Abo, K. Sugiura, A. Kusumi, T. Ohmori,
H. Watanabe, and K. Kumagai. 1992. Age-dependent increase of extrathymic T
3T. Inui, H. Nakashima, Y. Habu, R. Nakagawa, M. Fukasawa, M. Kinoshita, and S.
Seki. Neutralization of TNF? abrogates hepatic failure induced by ?-galactosylcer-
amide without attenuating its antitumor effect in aged mice. Submitted for publication.
6556 BYSTANDER ANTITUMOR CD8?T CELLS INDUCED BY ?-GalCer
cells in the liver and their appearance in the periphery of older mice. J. Immunol.
36. Nakagawa, R., I. Serizawa, K. Motoki, M. Sato, H. Ueno, R. Iijima,
H. Nakamura, A. Shimosaka, and Y. Koezuka. 2000. Antitumor activity of
?-galactosylceramide, KRN7000, in mice with the melanoma B16 hepatic me-
tastasis and immunohistological study of tumor infiltrating cells. Oncol. Res.
37. Nishimura, H., T. Yajima, Y. Naiki, H. Tsunobuchi, M. Umemura, K. Itano,
T. Matsuguchi, M. Suzuki, P. S. Ohashi, and Y. Yoshikai. 2000. Differential roles
of interleukin 15 mRNA isoforms generated by alternative splicing in immune
responses in vivo. J. Exp. Med. 191:157.
38. Judge, A. D., X. Zhang, H. Fujii, C. D. Surh, and J. Sprent. 2002. Interleukin 15
controls both proliferation and survival of a subset of memory-phenotype CD8?
T cells. J. Exp. Med. 196:935.
39. Takayama, E., S. Seki, T. Ohkawa, K. Ami, Y. Habu, T. Yamaguchi,
T. Tadakuma, and H. Hiraide. 2000. Mouse CD8?CD122?T cells with inter-
mediate TCR increasing with age provide a source of early IFN-? production.
J. Immunol. 164:5652.
40. Yamada, H., T. Ninomiya, A. Hashimoto, K. Tamada, H. Takimoto, and
K. Nomoto. 1998. Positive selection of extrathymically developed T cells by
self-antigens. J. Exp. Med. 188:779.
41. Yamada, H., G. Matsuzaki, Q. Chen, Y. Iwamoto, and K. Nomoto. 2001. Re-
evaluation of the origin of CD44high“memory phenotype” CD8 T cells: com-
parison between memory CD8 T cells and thymus-independent CD8 T cells. Eur.
J. Immunol. 31:1917.
42. Terra, R., N. Labrecque, and C. Perreault. 2002. Thymic and extrathymic T cell
development pathways follow different rules. J. Immunol. 169:684.
43. Clegg, C. H., J. T. Rulffes, P. M. Wallace, and H. S. Haugen. 1996. Regulation
of an extrathymic T-cell development pathway by oncostatin M. Nature
44. Watanabe, H., C. Miyaji, S. Seki, and T. Abo. 1996. c-kit?stem cells and thy-
mocyte precursors in the livers of adult mice. J. Exp. Med. 184:687.
6557The Journal of Immunology