Perforin-Deficient CD8?T Cells: In Vivo Priming and
Antigen-Specific Immunity Against Listeria monocytogenes1
Douglas W. White,* Adam MacNeil,†Dirk H. Busch,‡Ingrid M. Pilip,‡Eric G. Pamer,‡and
John T. Harty2*†
CD8?T cells require perforin to mediate immunity against some, but not all, intracellular pathogens. Previous studies with H-2b
MHC perforin gene knockout (PO) mice revealed both perforin-dependent and perforin-independent pathways of CD8?T cell-
mediated immunity to Listeria monocytogenes (LM). In this study, we address two previously unresolved issues regarding the
requirement for perforin in antilisterial immunity: 1) Is CD8?T cell-mediated, perforin-independent immunity specific for a
single Ag or generalizable to multiple Ags? 2) Is there a deficiency in the priming of the CD8?T cell compartment of PO mice
following an immunizing challenge with LM? We used H-2dMHC PO mice to generate CD8?T cell lines individually specific for
three known Ags expressed by a recombinant strain of virulent LM. Adoptive transfer experiments into BALB/c host mice
revealed that immunity can be mediated by PO CD8?T cells specific for all Ags examined, indicating that perforin-independent
immunity is not limited to CD8?T cells that recognize listeriolysin O. Analysis of epitope-specific CD8?T cell expansion by MHC
class I tetramer staining and ELISPOT revealed no deficiency in either the primary or secondary response to LM infection in PO
mice. These results demonstrate that the perforin-independent pathway of antilisterial resistance mediated by CD8?T cells is
generalizable to multiple epitopes. Furthermore, the results show that reduced antilisterial resistance observed with polyclonal PO
CD8?T cells is a consequence of a deficiency in effector function and not a result of suboptimal CD8?T cell priming. The Journal
of Immunology, 1999, 162: 980–988.
defense against certain viruses (1, 2), intracellular bacteria (3), and
protozoan pathogens (4).
One such bacterial pathogen, Listeria monocytogenes (LM),3
has been studied extensively as a model organism to dissect the
cellular immune response (5). Following injection into a mouse,
LM is phagocytosed by macrophages in the spleen and liver. Some
virulent LM, primarily through the actions of a secreted hemolysin
known as listeriolysin O (LLO), escape from the vacuole to the
cytoplasm of the infected eukaryotic host cell. Once in the cyto-
plasm, LM multiplies and polymerizes host-derived F-actin,
thereby initiating movement and direct spread to neighboring host
cells (6–9). In this fashion, LM can spread from host cell to host
cell without leaving the intracellular space.
The course of listeriosis in mice is acute: injection of a lethal
dose results in uncontrolled bacterial replication in the spleen and
D8?T cells play an important role in the immune re-
sponse against intracellular pathogens. A number of stud-
ies have demonstrated an obligate role for these cells in
liver and death in 4 to 6 days. However, mice challenged with a
sublethal primary dose clear the infection within 2 wk and are
subsequently resistant to challenges 100-fold greater than the LD50
in naive animals (10). Studies in mice that lack CD8?T cells have
demonstrated the importance of these cells in the adaptive re-
sponse to LM (3, 11–13).
The effectiveness of CD8?T cells in defense against LM and
other intracellular pathogens is based on their ability, via clonally
distributed TCRs, to recognize peptide Ags bound to MHC class I
molecules on the surface of an infected cell. Peptides displayed by
MHC class I molecules preferentially derive from proteins in the
cytoplasm of the host cell (14, 15). Thus, pathogens such as LM
with a component of their life cycle in the host cell cytoplasm are
often susceptible to protective immunity mediated by CD8?T
Once a CD8?T cell is activated, it is capable of elaborating a
number of effector functions that aid the immune system in the
clearance of the pathogen. Previously activated CD8?T cells
readily produce IFN-? and TNF in an Ag-specific fashion. Both of
these cytokines, which are also produced by cells other than CD8?
T cells, are known to be important in the normal immune response
against LM (10) and in other infectious disease models in which
CD8?T cells are important mediators of resistance (16). Our re-
sults have demonstrated that IFN-? is not required for the devel-
opment nor expression of CD8?T cell-mediated immunity to LM
(17). Recently, it has been shown that IFN-?-deficient CD8?, but
not CD4?, T cells are also capable of clearing a chronic LM in-
fection in SCID mice (18).
Efficient Ag-specific lysis of a target cell harboring an intracel-
lular pathogen is a function largely limited to CD8?T cells. At
least two molecular pathways have been identified by which CD8?
T cells can mediate cytolysis in vitro (19, 20). The granule exo-
cytosis pathway requires the coordinated activity of perforin and
granzymes, both of which are found in the granules of activated
*Interdisciplinary Graduate Program in Immunology and†Department of Microbiol-
ogy, University of Iowa, Iowa City, IA 52242; and‡Sections of Infectious Diseases
and Immunobiology, Yale University School of Medicine, New Haven, CT 06520
Received for publication August 11, 1998. Accepted for publication October 7, 1998.
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.
1This work was supported by National Institutes of Health Grants AI36864 and
AI42767 (J.T.H.). D.W.W. is a trainee in the Medical Scientist Training Program.
A.M. was supported as an undergraduate fellow by the Howard Hughes Medical
Institute. D.H.B. is a research fellow of the Deutsche Forschungsgemeinschaft (DFG).
E.G.P. is a Pew Scholar in Biomedical Sciences.
2Address correspondence and reprint requests to Dr. J. T. Harty, 3-512 Bowen Sci-
ence Building, Department of Microbiology, University of Iowa, Iowa City, IA
52242. E-mail address: firstname.lastname@example.org
3Abbreviations used in this paper: LM, Listeria monocytogenes; ELISPOT, enzyme-
linked immunospot assay; LCMV, lymphocytic choriomeningitis virus; LLO, list-
eriolysin O; NP, nucleoprotein; PE, phycoerythrin; PO, perforin deficient.
Copyright © 1999 by The American Association of Immunologists0022-1767/99/$02.00
CD8?T cells, to activate the caspase cascade of the target cell and
induce apoptosis (21–23). The absolute requirement for this path-
way in the clearance of lymphocytic choriomeningitis virus
(LCMV) has been demonstrated in mice with targeted disruption
of the gene for perforin (24, 25). Activated CD8?T cells also
express CD95 ligand, which can ligate CD95 (Fas, Apo-1) on a
target cell and induce apoptosis via the caspase cascade. This path-
way is probably most important in the elimination of self-reactive
T cells that are repeatedly exposed to Ag (26). Evidence for a
role of this pathway in CD8?T cell-mediated immunity against
viruses in vivo has been presented (27, 28), but its importance
as an effector function against most infectious agents remains
Two studies on the role of perforin in secondary resistance and
CD8?T cell-mediated immunity to LM have been reported pre-
viously (29, 30). The first report demonstrated that CD8?spleno-
cytes that lack perforin are deficient in their ability to transfer
antilisterial immunity (29). This finding indicates that perforin is
required for an optimal CD8?T cell response to LM. However, the
data presented in this report did not rule out the possibility that the
deficiency observed with PO splenocytes was due to suboptimal
priming of the CD8?compartment of PO donor mice.
The second report used H-2bPO CD8?T cells that had been
restimulated in vitro to identify a perforin-independent pathway by
which CD8?T cells are capable of mediating antilisterial immu-
nity (30). However, this report concentrated on CD8?T cells spe-
cific for a single LM Ag. It was therefore impossible to generalize
the findings regarding perforin-independent immunity to CD8?T
cells specific for other LM-derived Ags.
Both studies were performed using PO mice of the H-2bhap-
lotype in which precise LM-derived epitopes recognized by CD8?
T cells are unknown. The lack of known epitopes in the H-2b
system prevented the analysis of CD8?T cell priming in H-2bPO
mice as well as the analysis of CD8?T cells specific for more than
one LM Ag. To address these issues, we generated H-2dMHC PO
mice by backcross of H-2bPO mice with BALB/c. The CD8?T
cell response to LM is well characterized in the H-2dhaplotype
(31), allowing us to analyze multiple Ags as targets of PO CD8?
T cells, as well as the expansion of LM Ag-specific CD8?T cells
after immunization. We present data demonstrating that the per-
forin-independent pathway is not limited to CD8?T cells specific
for a single epitope, but rather functions in T cells specific for all
Ags tested. We also present data that rule out suboptimal priming
of the CD8?T cell compartment in perforin-knockout mice as an
explanation for the apparent deficiency of PO CD8?splenocytes in
mediating antilisterial immunity.
Materials and Methods
BALB/c (H-2dMHC) mice were obtained from the National Cancer In-
stitute (Frederick, MD) and crossed with PO (H-2bMHC) mice kindly
provided by Dr. W. R. Clark (25). F1mice were backcrossed to BALB/c
and H-2d/dperforin?/?mice were identified by flow-cytometric analysis of
PBL using Abs specific for H-2d(SF1-1.1.1) and H-2b(Y-3) and Southern
blot analysis for perforin genotype, as described (25). H-2d/dperforin?/?
mice were backcrossed to BALB/c three additional times and then inter-
crossed to generate H-2d/dperforin?/?(H-2dPO) mice. H-2dPO mice were
maintained by brother-sister mating and housed under specific pathogen-
free conditions at the University of Iowa (Iowa City, IA) animal care unit.
All mice were used at 8–16 wk of age in an age- and sex-matched fashion.
LM strain 10403s (32) and recombinant strain XFL303 (derived from
10403s), which expresses the LCMV NP 118–126 epitope as a secreted
fusion protein (33), are both resistant to streptomycin and were used as
previously described (33–36). Briefly, bacteria were grown in tryptic soy
broth to an OD600of approximately 0.1, diluted in pyrogen-free 0.9% so-
dium chloride (Abbott Laboratories, North Chicago, IL), and injected i.v.
in 0.2 ml per animal. Aliquots were plated onto tryptic soy agar containing
50 ?g/ml streptomycin to verify the number of CFU injected.
Cell lines, Abs, flow cytometry, and T cell depletion
P815 is a DBA/2-derived mastocytoma (H-2dMHC) (American Type Cul-
ture Collection (ATCC), Manassas, VA; ATCC TIB-64); P815-LLO refers
to P815 cells stably transfected with a plasmid construct expressing the LM
Ag LLO and neo-resistance (37); P815-p60 refers to P815 cells stably
transfected with a plasmid construct expressing the LM Ag p60 and neo-
resistance (34). P815-Fas is a derivative of P815 that expresses 10-fold
more surface CD95 than P815 (38); L1210F?(25) and L1210F?(39) are
derivatives of the lymphoblastic cell line L1210 (ATCC CCL-219) that
have been transfected with Fas antisense and sense cDNA, respectively.
Cell lines were maintained in RPMI 1640 (Life Technologies, Grand Is-
land, NY) supplemented with 10% FCS, antibiotics, L-glutamine, HEPES
buffer, and 2-ME (RP10 (34)). Transfected cells were maintained in RP10
supplemented with G-418 at 400 ?g/ml.
mAbs, which were purified from culture supernatants and quantitated as
previously described (17), were: rat anti-mouse TNF IgG (XT22 and XT3
(40) used in combination at a mass ratio of 1:1), rat anti-mouse IFN-? IgG
(XMG1.2 (41)), rat anti-mouse CD8 (2.43 (42)), and rat anti-mouse CD4
(GK1.5 (43)). Control polyclonal rat IgG was purchased from Sigma (St.
Louis, MO). Flow-cytometric analysis was performed as previously de-
scribed (17) using FITC-conjugated anti-CD8 (53.6-7; Sigma), PE-conju-
gated anti-CD4 (H129.19; Sigma), mouse anti-mouse H-2KbIgG2b (Y-3;
ATCC), and mouse anti-mouse H-2KdIgG2a (SF1-1.1.1; ATCC). T cell
subset depletion by mAb and complement was performed in vitro using rat
anti-mouse CD4 (RL172) and rat anti-mouse CD8 (3.168), as previously
described (17). T cell subset depletion in vivo was conducted as previously
described using 2.43 and GK1.5 (17).
Generation and maintenance of CD8?T cell lines
CD8?T cell lines specific for LLO 91–99, p60 217–225, or NP 118–126
were derived from BALB/c and H-2dPO mice using methods previously
described (37). Briefly, 2–4 ? 107splenocytes from mice injected 7–10
days previously with 103–4CFU of virulent LM strain 10403s or XFL303
were incubated in RP10 with 3 ? 106irradiated (150 Gy) stimulator cells
(P815-LLO cells or P815-p60 cells). Subsequent weekly restimulations
were conducted by combining 1–3 ? 106responder cells with 3 ? 106
irradiated stimulator cells and approximately 4 ? 107irradiated (30 Gy)
syngeneic splenocytes in RP10 supplemented with 5% supernatant from
Con A-stimulated rat spleen cells and 50 mM ?-methyl mannoside. In the
case of T cell lines specific for NP 118–126, P815-derived stimulator cells
were left out and 3 ? 107irradiated syngeneic splenocytes were incubated
with 100 nM synthetic NP 118–126 peptide for 1 h at 37°C and rinsed
three times before their addition to the T cell culture.
In vitro characterization of CD8?T cell lines
51Cr release assays were performed as previously described (44, 17, 30).
Briefly, labeled target cells were combined with effector cells at the indi-
cated ratios in RP10 in round-bottom 96-well plates. Following a 4–7.5-h
incubation (as indicated), supernatant was harvested and assayed for
51Cr release. Percent specific release of
formula: 100 ? (experimental cpm ? spontaneous cpm)/(total cpm ?
spontaneous cpm). Spontaneous release was less than 20% of total in all
TNF was quantitated using a WEHI 164 clone 13 bioassay (45), as
previously described (30). Briefly, supernatants from coincubations of
CD8?T cells and target cells were added to WEHI 164 cells in flat-bottom
96-well plates. Following overnight incubation, survival of incubator cells
was assayed by the addition of alamar blue (Acumed, West Lake, OH).
Death of the indicator cells, a relative measure of TNF production, was
determined 2–6 h after addition of alamar blue by measuring OD570-
OD600. Production of TNF by target cells in the absence of CD8?T cells
was not detected. Murine rTNF (Boehringer Mannheim, Indianapolis, IN)
was used as a control and to determine the detection limits of the WEHI
bioassay. Concentrations of 1–10 pg/ml of rTNF were routinely detected
using this assay.
IFN-? was quantitated by ELISA, as previously described (30). Briefly,
supernatants from overnight coincubations of effector cells and target cells
and rIFN-? controls were added to 96-well plates that had been previously
coated with rat anti-mouse IFN-? (XMG1.2) mAb. Rabbit anti-mouse
IFN-? (a gift from J. Cowdery at the University of Iowa), alkaline phos-
phatase-conjugated goat anti-rabbit Ig (Sigma), and alkaline phosphatase
51Cr was calculated by the
981The Journal of Immunology
substrate (Sigma) were added sequentially according to the manufac-
turer’s protocol. OD was measured at 405 nm. Limit of detection was
less than 50 U/ml.
Adoptive transfer experiments
The capacity of splenocytes derived from immunized animals and CD8?T
cell lines to mediate antilisterial immunity in vivo was determined using
adoptive transfer assays, as described previously (44, 17, 30). Briefly,
RBC-depleted splenocytes from donor mice immunized 7–10 days previ-
ously with 103–4virulent LM strain 10403s or XFL303 or CD8?T cells
restimulated in vitro 7 to 9 days previously were harvested, washed in
antibiotic-free buffer, and resuspended in pyrogen-free 0.9% sodium chlo-
ride. Cells were delivered i.v. in 0.5-ml into naive BALB/c host mice.
Within 2 h, host mice, including uninjected controls, were challenged i.v.
with the indicated dose of virulent LM 10403s or XFL303. CFU/spleen and
liver were determined 3 days postchallenge by homogenizing the spleens
and livers in 0.2% IGEPAL (Sigma), plating 10-fold serial dilutions onto
tryptic soy agar containing 50 ?g/ml streptomycin, and calculating colony
count averages after overnight incubation at 37°C.
The prevalence of activated Ag-specific CD8?T cells in the spleens of
BALB/c and PO mice was determined by ELISPOT analysis, as previously
described (46–49). Briefly, splenocytes (5 ? 103-105/well) were cocul-
tured with P815-LLO target cells (105/well) for 24–48 h in flat-bottom
96-well plates that had been previously coated with rat anti-mouse IFN-?
mAb (R4-6A2; PharMingen). Following lysis of the cells with distilled
water and rinses with PBS containing 0.2% Tween-20, the plates were
incubated with rabbit anti-IFN-? polysera. After rinsing, the plates were
incubated with donkey anti-rabbit Ig conjugated to alkaline phosphatase
(Jackson). Further rinsing was followed by the addition of 5-bromo-4
chloro-3-indolyl phosphate (BCIP) substrate (Sigma) in AMP buffer im-
pregnated with 0.75% agarose. The reaction was developed at 37°C and
spots were counted using a dissection microscope. The average frequency
of responders from triplicate determinations was multiplied by the total
number of splenocytes to calculate responders per spleen.
Analysis of CD8?T cells using Kd-peptide tetramer complexes
Kd-peptide tetramer complexes were generated and used as previously de-
scribed (50), with minor modifications. For staining, approximately 5 ?
106lympholyte-M (Cedarlane) separated splenocytes were blocked with
anti-FcR mAb (2.4G2, gift of T. Waldschmidt at the University of Iowa)
before staining with FITC-conjugated anti-CD8 (53.6-7; Sigma) and PE-
conjugated Kd-peptide tetramer complexes for 1 h at 4°C. After washing,
the cells were resuspended in PBS containing 0.01% sodium azide, 1%
BSA, and 1 ?g/ml propidium iodide, and then analyzed on a FACScan
using CyCLOPS software (Cytomation, Fort Collins, CO). Lymphocytes
(determined by forward scatter and side scatter) that excluded propidium
iodide were analyzed for CD8 and Kd-peptide tetramer-specific staining.
The frequency of CD8?cells and CD8?, tetramer?cells was used with the
total splenocyte count to calculate the responders per spleen.
CD8?splenocytes derived from H-2dPO mice appear to be
deficient, compared with wild-type CD8?splenocytes, in
mediating antilisterial immunity
We bred the PO mutation onto the BALB/c (H-2d) background in
which multiple LM Ags have been described (31). Experiments in
which we compared naive H-2dPO mice with BALB/c mice did
not reveal a significant difference in the LD50of virulent LM (data
not shown). This result is consistent with studies that compared
H-2bPO mice with wild-type C57BL/6 mice (29, and data not
shown) and verifies that perforin is not required for resistance to
primary LM infection.
We next compared the antilisterial activity of immune spleno-
cytes derived from H-2dPO mice with those from BALB/c mice.
Since the resistance of naive PO and BALB/c mice to primary LM
infection was similar, we used standard sublethal doses of virulent
LM (?0.1 LD50) to immunize H-2dPO and BALB/c donor mice.
Seven days postimmunization, equivalent numbers of BALB/c or
PO donor splenocytes were transferred into naive BALB/c host
mice that were subsequently challenged with a high dose of viru-
lent LM. Bacterial counts in the spleens and livers of splenocyte
recipient and control mice 3 days postchallenge demonstrated that
BALB/c-derived splenocytes provided dramatic antilisterial pro-
tection in both the spleen (Fig. 1A) and the liver (Fig. 1B), reducing
bacterial recovery ?100,000-fold and ?10,000-fold, respectively,
compared with mice that did not receive splenocytes. This activity
was primarily mediated by CD8?cells since depletion of the
CD8?compartment with Ab and complement before transfer
eliminated the majority of the protection (Fig. 1, A and B). Immu-
nity mediated by splenocytes derived from H-2dPO mice was
comparable with that provided by wild-type splenocytes in the
liver (8,000-fold reduction), but somewhat less than that provided
by wild-type splenocytes in the spleen (1,000-fold reduction). De-
pletion of CD8?cells from H-2dPO-derived splenocytes had a
modest effect on the level of immunity transferred, but did not
eliminate immunity mediated by these cells (Fig. 1, A and B).
These results recapitulate those from H-2bmice (29, 30) and fur-
ther support the model that perforin plays a role in the normal
immune response to LM, especially in the spleen, and that spleno-
cytes other than CD8?T cells may play a significant role in adap-
tive immunity to LM in the PO mouse.
Splenocytes from immunized PO mice also transferred immu-
nity into naive PO host mice at levels that were indistinguishable
from that observed in BALB/c hosts (data not shown). This result
argues against a model in which perforin, derived from host cells,
plays a role in antilisterial immunity observed in adoptive transfer
from H-2dBALB/c or PO mice. Three BALB/c recipients per group were
injected i.v. with 50 ? 106splenocytes from the indicated donor mice that
had been immunized i.v. with 3.3 ? 103virulent LM 7 days previously.
Transferred splenocytes were mock depleted or depleted of CD8?cells in
vitro, as indicated, with Ab and complement. Depletion efficiency was
?95%, as determined by flow cytometry (data not shown). Following the
transfer of immune splenocytes, recipient mice were challenged i.v. with
1.1 ? 105virulent LM strain 10403s (approximately 10 LD50). CFU anal-
ysis was performed on day 3 postchallenge. One control mouse that did not
receive any immune splenocytes died before the CFU assay. Data are pre-
sented as mean log10CFU ? SD. Student’s t test was used in statistical
analysis; p-values are shown for each group compared with the control
group that did not receive splenocytes.
Antilisterial immunity mediated by splenocytes derived
982PERFORIN-DEFICIENT CD8?T CELLS AND IMMUNITY TO L. monocytogenes
Generation and characterization in vitro of CD8?T cell lines
specific for three known antigenic peptides expressed by
recombinant LM XFL303
Experiments with LLO-specific CD8?T cell lines from H-2bPO
mice previously identified a perforin-independent pathway by
which CD8?T cells mediate antilisterial immunity (30). To de-
termine whether this result was restricted to LLO-specific CD8?T
cells, we generated and characterized CD8?T cell lines specific
for multiple Ags from H-2dPO and syngeneic control (BALB/c as
well as H-2dperforin?/?) mice that had been immunized with
virulent LM XFL303. XFL303 is a recombinant strain of virulent
LM that secretes the LCMV-derived NP 118–126 epitope in the
context of a dihydrofolate reductase (DHFR) fusion protein and
strongly activates CD8?T cells that recognize NP 118–126 bound
to H-2Ldin the H-2dmouse (33). Splenocytes from H-2dPO mice
immunized 7 days previously with a sublethal dose of virulent LM
XFL303 were restimulated in vitro with H-2dtarget cells express-
ing LLO 91–99, p60 217–225, or NP 118–126 peptide. CD8?T
cell lines specific for NP 118–126 were also generated from
control (BALB/c and perforin?/?) mice. BALB/c-derived
CD8?T cell lines specific for LLO 91–99 and p60 217–225
have been described previously (44, 34).
After several in vitro restimulations, all lines were ?95%
CD8?CD4?, as measured by flow cytometry (data not shown). To
verify Ag specificity and cytolytic activity, we performed51Cr
release assays. All CD8?T cell lines mediated cytolysis of H-2d
MHC target cells in the presence, but not in the absence, of the
appropriate Ag (Fig. 2, A–D). As expected, CD8?T cells derived
from control mice (Fig. 2D) mediated higher levels of specific lysis
in shorter time period than did CD8?T cells derived from PO mice
(Fig. 2, A–C). Delayed cytolysis in the absence of perforin is con-
sistent with previous studies that have documented the importance
of perforin in cytolytic assays in vitro (24, 25, 51, 52).
To verify that PO CD8?T cells mediate cytolysis via CD95,
cells of the H-2dhaplotype that vary in their expression of CD95
were utilized as target cells in
Whereas peptide-coated, CD95-expressing (P815-Fas (38)) target
cells were lysed at high levels in a51Cr release assay by PO-
derived CD8?T cells, peptide-coated P815 cells that express very
low levels of CD95 (P815) were not lysed (Fig. 3A). PO CD8?T
cell lines specific for p60 and NP also exhibited increased lysis of
target cells expressing CD95 compared with target cells that ex-
press minimal CD95 (data not shown). Similar results were ob-
tained with another pair of target cells that differ in expression of
CD95, namely L1210F?and L1210F?(39, 25). We observed spe-
cific lysis of CD95-expressing targets (Fig. 3B) and background
levels of lysis of targets that do not express CD95 (data not
shown). Finally, Ag-specific lysis of L1210F?cells by PO CD8?
T cells was inhibited by a mAb specific for CD95, but not by
51Cr release assays (Fig. 3).
CD8?T cell lines, specific for LLO 91–99 (circles), p60 216–225
(squares), or NP 118–126 (triangles), derived from PO (circles, squares,
and upward triangles) or BALB/c (downward triangles) mice, were incu-
bated with P815 target cells (open symbols) or P815 target cells coated
with the appropriate peptide (closed symbols) in a 7.5-h (A–C) or 4-h (D)
51Cr release assay. These data are representative of at least three indepen-
dent experiments with similar results. E–H, Supernatants from overnight
coincubations of CD8?T cell lines and target cells (as in A–D) were
assayed for IFN-? by ELISA. These data represent the mean of triplicate
determinations made in four independent experiments. SD did not exceed
21% for any data point. I–L, Supernatants from 10-h coincubations of
CD8?T cell lines and target cells (as in A–D) were assayed for TNF. TNF
was quantitated in a bioassay using WEHI 164 clone 13 cells (45), which
die in the presence of TNF. These data represent at least three independent
experiments with similar results for each CD8?T cell line.
In vitro characterization of PO CD8?T cell lines. A–D,
line. A, LLO-specific CD8?T cells derived from PO mice were incubated
for 5.5 h with CD95-low P815 target cells (squares) or CD95-high P815-
Fas target cells (circles) that were uncoated (open symbols) or coated with
LLO 91–99 peptide (closed symbols) in a51Cr release assay. These results
represent two independent experiments with similar results, which mea-
sured the CD95-dependent cytolytic activity of three individual PO CD8?
T cell lines with specificity for LLO 91–99, p60 216–225, and NP 118–
126. B, LLO-specific CD8?T cells derived from PO mice were incubated
for 4.5 h with CD95-expressing L1210F?target cells uncoated (open bars)
or coated with LLO 91–99 peptide (closed bars) in a51Cr release assay.
E:T ? 1:1. Anti-CD95 mAb (Jo2) or control IgG were added at 5 ?g/ml.
CD95-dependent cytolysis mediated by a PO CD8?T cell
983The Journal of Immunology
control rat IgG (Fig. 3B). Complete inhibition was observed at
most E:T ratios using 5 ?g/ml of anti-CD95 mAb, a concentration
that did not kill target cells nor the CD8?T cells in the time frame
of the51Cr release assay (data not shown). While these results do
not rigorously rule out perforin-independent, CD95-independent
mechanisms of lysis by PO-derived CD8?T cells, they do estab-
lish the presence of a CD95-dependent pathway.
All CD8?T cell lines that were generated also produced IFN-?
(Fig. 2, E–H) and TNF (Fig. 2, I–L) in an Ag-specific fashion.
IFN-? production by H-2dPO CD8?T cells specific for LLO and
p60 was also verified by ELISPOT analysis (data not shown).
In general, we observed higher levels of IFN-? production by
PO-derived CD8?T cells compared with wild-type cells (Fig. 2),
a finding that is consistent with data from Sad and colleagues (53).
In vivo immunity can be mediated by PO CD8?T cells specific
for a range of LM-derived peptides
Our previous studies demonstrated that LLO-specific CD8?T
cells from H-2bPO mice can provide significant immunity to LM.
To test whether perforin-independent CD8?T cell-mediated im-
munity is generalizable to other Ags, we tested the ability of H-2d
PO CD8?T cells specific for LLO 91–99, p60 217–225, and NP
118–126 to mediate immunity against rLM XFL303 in vivo.
CD8?T cells were transferred into naive BALB/c host mice that
were subsequently challenged with approximately 10 LD50of vir-
ulent LM XFL303. PO-derived CD8?T cells specific for LLO
91–99, p60 217–225, and NP 118–126 all provided antilisterial
immunity in the liver (Fig. 4, E–F). PO-derived CD8?T cells also
reduced bacterial counts in the spleen (Fig. 4, A–C), albeit to a
lesser degree. BALB/c-derived CD8?T cells specific for NP 118–
126 provided high levels of antilisterial immunity in both the
spleen (Fig. 4D) and the liver (Fig. 4H). Immunity mediated by
BALB/c-derived CD8?T cells specific for LLO 91–99 and p60
217–225 has been described previously (44, 34). Consistent with
our previous studies, the degree of immunity in the spleen medi-
ated by PO-derived CD8?T cells was typically less compared
with that which is usually observed with BALB/c-derived CD8?T
cells (Fig. 4, A–C versus D).
Studies were performed to confirm that the reduction in CFUs
provided by PO CD8?T cells correlated with survival of the an-
imals to an otherwise lethal challenge. One hundred percent of
mice that received PO-derived CD8?T cells (specific for LLO
91–99 or NP 118–126) survived at least 12 days without any overt
signs of illness following a challenge of XFL303 that killed all
mice that did not receive any CD8?T cells (Table I). Similarly, all
mice that received PO CD8?T cells specific for LLO survived a
lethal challenge with virulent LM 10403s (Table I).
The Ag specificity of immunity mediated by PO CD8?T cell
lines was confirmed in experiments in which PO CD8?T cells
specific for NP 118–126 mediated immunity, as measured by sur-
vival, against virulent LM XFL303, which expresses the NP
epitope, but failed to protect against the parental LM strain 10403s
that lacks the NP epitope (Table I).
These results demonstrate that perforin-independent mecha-
nisms of antilisterial immunity mediated by CD8?T cells are not
restricted to CD8?T cells that recognize LLO.
Analysis of priming and expansion of the CD8?T cell
compartment in PO versus syngeneic BALB/c mice
We previously hypothesized that splenocytes from PO mice might
be deficient in mediating antilisterial immunity due to inefficient
priming of LM-specific CD8?T cells in these mice (30). To ad-
dress this issue, we quantitated LLO-specific CD8?T cells in PO
CD8?T cells specific for multiple LM-derived peptides. Naive BALB/c
mice were injected i.v. with CD8?T cells (hatched bars) derived from PO
(A–C and E–G) or BALB/c (D and H) mice, and within hours challenged
with virulent LM. Control mice did not receive any T cells (open bars).
CFU from the spleen (A–D) and liver (E–H) were quantitated 3 days post-
challenge. Data are presented as mean log10CFU ? SD for five to nine
animals per group. These data are pooled from two to three independent
experiments for each CD8?T cell line. Student’s t test was used in sta-
tistical analysis; p-values are shown for each group compared with the
control group in the same experiment that did not receive T cells. T cells
transferred: 1–1.6 ? 107per mouse. Challenge with XFL303 (in the case
of NP-specific CD8?T cells) or 10403s (in the case of LLO- or p60-
specific CD8?T cells): 0.5–1.7 ? 105CFU per mouse.
In vivo immunity against virulent LM mediated by PO
Table I. CD8?T cell lines derived from H-2dPO mice provide Ag-specific immunity as measured by survivala
Host MiceT CellsLM StrainSurvival
10403s (wild type)
10403s (wild type)
10403s (wild type)
10403s (wild type)
aBALB/c- or PO-derived CD8?T cells (5–10 ? 106) specific for LLO 91–99 (LLO) or NP 118–126 (NP) were injected i.v. into BALB/c host mice
that were subsequently challenged with 1–2 ? 105NP 118–126-expressing virulent rLM (XFL303) or the parental strain 10403s that does not express
NP 118–126. Survival was monitored for 12 days. All animals that died succumbed within 5 days of challenge.
984 PERFORIN-DEFICIENT CD8?T CELLS AND IMMUNITY TO L. monocytogenes