| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Microbiology and Tumor Biology Center, Karolinska Institute, Stockholm, Sweden; and
Howard Hughes Medical Research Institute and Department of Microbiology and Immunology, Vanderbilt University School of Medicine, Nashville, TN 37232
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
In a comparison of TAP1-/-, ß2m-/-, and TAP1/ß2m-/- mice, expression of MHC class I molecules is highest, relatively, in TAP1-/- mice (approximately 10% of wild-type mice for both H-2Kb and H-2Db) followed by ß2m-/- and TAP1/ß2m-/- mice (15). While low levels of MHC class I molecules have been detected on the cell surface of ß2m-/- cells (5, 6, 15, 16), no class I expression is detected on the cell surface of TAP1/ß2m-/- cells by either biochemistry or flow cytometry (15, 16). However, cells from these mice can be killed by MHC class I-specific CD8+ T cells, indicating that they may express small levels of H-2Db and/or Kb molecules at the cell surface (15). Although low in numbers, CD8+ T cells are detected in all three strains of mice, being highest, relatively, in TAP1-/- followed by ß2m-/- and TAP1/ß2m-/- mice (13, 15) (reviewed in 2 . Several in vitro studies have revealed that the residual CD8+ T cells in the TAP1-/-, ß2m-/-, and TAP1/ß2m-/- mice differ in reactivity compared with CD8+ T cells in wild-type mice. In particular, CD8+ T cells from these mutant mice exhibit a strong bias toward reactivity against cells from syngeneic wild-type mice, i.e., cells expressing normal levels of "self" MHC class I molecules (9, 12, 13, 15). This difference in reactivity has been interpreted as a direct consequence of CD8+ T cell selection on thymic cells expressing a low ligand density of MHC class I molecules and/or on MHC class I molecules bound by a limited number of peptides (9, 12) (reviewed in 2 .
In contrast to the defects seen for the CD8+ T cell compartment, MHC class I-deficient mice have normal numbers of NK cells. However, these NK cells have altered reactivities compared with similar cells in wild-type mice. In particular, these cells do not kill MHC class I-deficient lymphoblasts in vitro, and they do not reject MHC class I-deficient bone marrow grafts (17–21 and unpublished results) (reviewed in 22 .
In the present study, we have addressed the ability of TAP1-/-, ß2m-/-, and TAP1/ß2m-/- mice to mount rejection responses against allogeneic and syngeneic MHC class I-positive tumor grafts, and against MHC class I-deficient tumor grafts. Three questions prompted the present studies: 1) we were interested in comparing how the TAP1-/-, ß2m-/-, and TAP1/ß2m-/- mutations, affecting MHC class I expression and T cell selection at different degrees (13, 15), influenced the ability of these mice to mount cytotoxic responses to allogeneic tumor grafts. 2) We wanted to address two seemingly conflicting results with regard to immune responses of ß2m-/- mice against syngeneic tumors (7, 8, 9, 10). Despite strong in vitro CD8+ T cell responses against syngeneic tumor cells (9), ß2m-/- mice do not reject these tumors in vivo (7, 8, 10). 3) We wanted to measure rejection responses against ß2m-deficient tumor grafts in ß2m-/- mice, since such grafts are efficiently eliminated by NK cells in wild-type mice (23).
The results revealed a markedly potent ability of all mutant mouse strains to mount responses against large grafts of allogeneic tumors. In contrast to published data, we also observed rejection responses against syngeneic MHC class I-positive tumors in the mutant mice. Finally, in contrast to wild-type mice, MHC class I-deficient tumor grafts were accepted in MHC class I-deficient mice. The results are discussed in relation to the role of MHC class I molecules in selecting functional CD8+ T and NK cell repertoires, and the consequences of this process for the development of a cell-mediated immune response in vivo.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
All mice were bred and maintained at the Microbiology and Tumor Biology Center (MTC), Karolinska Institute, Stockholm, Sweden. The generation of ß2m-/- and TAP1-/- as well as MHC class-II-/- mice has been described (4, 11, 24). ß2m-/- mice were from the Jackson Laboratories (Bar Harbor, ME), ß2m-/- of the H-2k and H-2d backgrounds were a kind gift from Dr. D. Roopenian. MHC class II-/- mice were a kind gift from Drs. C. Benoist and D. Mathis. TAP1/ß2m-/- mice were obtained by intercrossing TAP1-/- and ß2m-/- mice as described (15). All mice, when not otherwise noted, were on an H-2b background. The ß2m-/- and TAP1-/- mice were back-crossed to C57BL/6 (B6) mice at least six times before use.
Tumors
P815 is a mastocytoma of DBA/2 (H-2d) origin. EL-4 is a T cell lymphoma of C57BL/6 (H-2b) origin. C4.4-25- is a ß2m-deficient mutant of EL-4 (23). All tumors were maintained as tissue culture cell lines or passaged as ascites lines in 400-rad irradiated H-2 syngeneic mice.
Immunizations
For immunizations, mice were given 5 x 106 10,000-rad irradiated tumor cells or 25 x 106 3,000-rad irradiated spleen cells i.p. 1 wk prior to inoculation with live tumor cells.
Tumor rejection studies
Mice used for in vivo experiments were 6 to 10 wk old at the start of the experiments, usually littermates, or otherwise age matched within 2 weeks. Graded doses of tumor cells were inoculated s.c., in the left flank when not otherwise noted, in a 0.2-ml volume of PBS. Tumors always appeared at the site of inoculation. Tumor growth was followed at least once weekly by palpations and measurements of the tumor size. Mice were killed when the tumors reached a size of >15 mm in diameter, and no signs of rejection were observed. Mice without any signs of tumor growth were kept under observation for at least 6 wk after inoculation. Small groups of mice, never exceeding five mice per group, were tested in several independent experiments throughout the study to minimize the risk of random fluctuations in the quantity or quality of cells inoculated. Results from several independent tests were pooled if not otherwise noted.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
B6 wild-type as well as TAP1-/-,
ß2m-/-, and TAP1/ß2m-/- mice were
grafted s.c. with titrated doses of P815 (H-2d) mastocytoma
cells. B6 control mice did not show any signs of tumor growth when
doses of 105 cells or less were grafted s.c. When
106 or 107 cells were grafted, tumors appeared
at the site of inoculation in a fraction of mice inoculated but were
rejected from all mice within 3 wk after inoculation (Fig. 1
) (data not
shown). A similar pattern was observed in TAP1-/- mice,
although a relatively higher number of mice developed tumors after
grafts of 106 or 107 cells. Nonetheless,
complete tumor rejection responses were observed within 4 wk after
tumor challenge in all mice studied (Fig. 1). In ß2m-/- mice,
tumors appeared at the site of inoculation after grafts of
105 tumor cells. Tumor rejection responses were similar or
only slightly delayed as compared with TAP1-/- mice, and
within 5 wk most mice had rejected grafted tumors.
TAP1/ß2m-/- showed a pattern similar to that of
ß2m-/- mice, with the exception of the lowest dose
inoculated (105 tumor cells), when a majority of the mice
developed tumors within 2 wk after tumor challenge. Yet, despite
initial formation of large s.c. tumors (sizes up to 13 mm in diameter)
nearly all TAP1/ß2m-/- mice rejected the tumors within
5 wk after challenge (Fig. 1). Measurable differences in tumor sizes
between the respective groups of mice were most clearly detectable
within 2 wk after inoculation of 105 and 106
cells, when the relatively largest tumor loads were observed in
TAP1/ß2m-/- mice followed by ß2m-/-,
TAP1-/-, and B6 mice (Fig. 2).
|
|
Previous studies with ß2m-/- mice have
demonstrated the generation of strong CD8+ cytotoxic
responses in vitro against syngeneic MHC class I-positive cells,
including tumor cells (9). However, despite these strong responses,
ß2m-/- mice do not reject syngeneic MHC class
I-positive tumors in vivo (8, 10) (reviewed in 2 . To address
these apparently conflicting results, we first determined the growth of
titrated (102–106) syngeneic MHC class
I-expressing EL-4 cells in B6, TAP1-/-,
ß2m-/-, and TAP1/ß2m-/- mice. These
studies revealed a strikingly similar pattern of tumor growth in all
MHC class I-deficient strains well as in B6 wild-type mice (Fig. 3) (data not shown), confirming and
extending published results in ß2m-/- mice (8, 10). To
exclude that these results were not a specific property of the
B6-derived lymphoma EL-4, we performed similar titrations of MHC class
I expressing RMA lymphoma cells and MC57X fibrosarcoma cells (both B6
derived). These studies yielded results similar to those obtained with
EL-4 cells (data not shown). To exclude that the present results were
due to a general inability of the EL-4 cells to be rejected in
ß2m-/- mice, these cells were grafted to
ß2m-/- mice of the H-2d and
H-2k haplotypes. In both of these strains, EL-4 cells were
readily rejected after initial growth (Fig. 4; H-2d
ß2m-/- mice not shown), indicating that rejection
responses could be mounted to the tumor even in mice of a
ß2m-/- background.
|
|
In vitro studies have demonstrated that alloreactive
CD8+ T cells from MHC class I-deficient mice generated in
MLRs cross-react with syngeneic cells expressing normal levels of MHC
class I molecules (9). This observation led us to address whether it
was possible to induce rejection of syngeneic MHC class I expressing
EL-4 cells by MHC class I-deficient (ß2m-/-) mice upon
simultaneous inoculation with allogeneic P815 cells. However, neither
simultaneous inoculation of both tumors in different flanks in the same
mouse nor intermixing the tumors prior to s.c. inoculation led to any
significant rejection of the EL-4 tumor cells (Fig. 5). Mice in the latter group, inoculated
with 1:1 mixtures of syngeneic EL-4 and allogeneic P815 tumors, readily
developed s.c. tumors. When solid s.c. tumors were removed from these
mice (at a tumor size of approximately 10 mm in diameter), all tumor
cells were found to be H-2b positive and H-2d
negative as revealed by FACS analysis, indicating selective growth of
EL-4 tumor cells (data not shown).
|
In contrast to the observations above, we found that MHC class
I-deficient (ß2m-/-) mice that had previously rejected
allogeneic P815 tumor grafts showed weak, albeit clearly detectable,
rejection responses against EL-4 tumor grafts (Fig. 6) (data not
shown). This result was specific for the MHC class I-deficient mice
since EL-4 cells rapidly formed solid tumors with no signs of rejection
in B6 wild-type mice that previously had rejected P815 tumor grafts.
Similar results were obtained with ß2m-/- mice that
were preimmunized with irradiated P815 cells 1 wk prior to challenge
with live EL-4 tumor cells (Fig. 6).
These results mimicked previous in vitro results, in which allospecific
CD8+ T cells from ß2m-/- mice showed
cross-reactivity on target cells expressing syngeneic class I molecules
expressed at a normal ligand density (9) (reviewed in 2 .
|
The results above prompted us to preimmunize
ß2m-/- mice with irradiated syngeneic EL-4 tumor cells
or B6-derived splenocytes and subsequently graft these mice with live
EL-4 cells. While only relatively weak tumor rejection responses were
observed in ß2m-/- mice preimmunized with EL-4 cells,
strong rejection responses were observed when ß2m-/-
mice were preimmunized with B6-derived spleen cells, and subsequently
challenged with EL-4 cells (Fig. 7). Control experiments
confirmed that the EL-4 cells were not rejected when grafted to B6 mice
immunized with B6-derived spleen cells. Thus, the present observations
demonstrate that MHC class I-deficient mice can reject tumor grafts
expressing syngeneic MHC class I molecules, provided that they have
been preimmunized. The relative differences observed in EL-4 rejection
responses after immunization of ß2m-/- mice with
irradiated EL-4 cells and B6-derived splenocytes, respectively, could
be due to either the cell dose used for immunization, the more potent
ability of spleen cells than tumor cells to prime T cells, or possibly
other reasons. This matter was not investigated further.
|
The rejection responses observed against EL-4 cells in
preimmunized MHC class I deficient mice were dependent on the
expression of MHC class I molecules on the cells used for immunization.
Preimmunization of ß2m-/- mice with either MHC class
II-/- or control B6 wild-type splenocytes was sufficient
to induce rejection of grafted EL-4 tumor cells, while preimmunization
of ß2m-/- mice with ß2m-/- splenocytes
did not induce rejection of grafted EL-4 tumor cells (Fig. 8). These responses were also dependent
on the expression of MHC class I molecules at the level of the target
cell. MHC class I-deficient EL-4 cells (C4.4-25-, a
ß2m-deficient variant of EL-4) were not rejected under conditions in
which rejection of EL-4 wild-type cells was observed.
|
After studying the fate of allogeneic and MHC class I-positive
syngeneic tumor grafts in MHC class I-deficient mice, we finally
addressed the ability of these mice to reject MHC class I-deficient
tumor grafts. MHC class I-deficient tumors are eliminated by NK cells
in wild-type B6 mice (23, 25, 26). Several studies have demonstrated
that MHC class I-deficient mice fail to efficiently kill MHC class I
deficient cells in vitro, and to resist MHC class I-deficient bone
marrow grafts in vivo (17, 18, 19, 20, 21) (reviewed in 22 . In agreement with
these findings, we here demonstrate that ß2m-/- mice
readily accept ß2m-deficient EL-4 (C4.4-25-) tumor
grafts (Fig. 9).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
). The results
reveal a potent ability of MHC class I-deficient mouse strains,
including TAP1/ß2m-/- mice, to reject large grafts of
allogeneic tumors. The ability to reject allogeneic tumor grafts
correlated with numbers of CD8+ T cells in the mice. In
contrast to the conclusions of previous reports, rejection responses
against syngeneic MHC class I-expressing tumors were observed in MHC
class I-deficient mice while similar responses were not detected in
wild-type mice. Finally, MHC class I-deficient tumor grafts were
accepted in MHC class I-deficient mice while similar grafts were
rejected in wild-type mice.
|
Earlier reports have discussed the origin of the CD8+ T cells detected in ß2m-/- mice (reviewed in 2 . It was speculated that these cells could develop either through positive selection on low levels of MHC class I molecules in the thymus or, alternatively, represent a small population of CD8+ T cells that differentiated independently of class I recognition. In an initial analysis of this phenomenon, it was observed that the small pool of CD8+ T cells selected in ß2m-/- mice not only reacted with the allogeneic cells but also cross-reacted and killed cells expressing self MHC class I expressed at normal levels (9). Similar findings were also observed in studies of TAP1-/- and TAP1/ß2m-/- mice (12, 13, 14, 15). This bias toward reactivity against syngeneic MHC class I expressed at a normal ligand density strongly indicated that at least a part of the residual CD8+ T cell population had been selected on the low levels of MHC class I molecules expressed in these mice. The studies in ß2m-/- mice also demonstrated that CD8+ T cells could be induced readily in a primary MLC after stimulation with H-2b class I-positive cells, whereas allogeneic cells generally failed to elicit a response under similar conditions (9). In the latter case, in vivo priming was necessary for the generation of efficient responses. In retrospect, this observation may explain why Zijlstra et al. as well as Koller et al. in their initial studies failed to detect CD8+ T cell-mediated reactivity in the ß2m-/- mice (3, 4). Taken together, these results suggested that events occurring during T cell selection influence the reactivity of CD8+ T cells selected in MHC class I-deficient mice (9) (reviewed in 2 . These findings have been taken as evidence for a novel specificity of the CD8+ T cells in MHC class I-deficient mice, i.e., an ability to specifically recognize and kill target cells expressing syngeneic MHC class I expressed at a normal ligand density (9, 12). Similar conclusions have more recently been made from studies in mice expressing MHC class II molecules loaded with a single or a limited set of peptides (27, 28, 29, 30).
The findings that MHC class I-deficient mice can generate
CD8+ T cell responses against cells expressing syngeneic
MHC class I also provides an explanation for earlier experiments with
ß2m-/- mice using skin grafts (31, 32). These mice were
shown to reject syngeneic MHC class I-positive skin grafts. The
discovery of a residual CD8+ T cell pool in the
ß2m-/- mice, and the strong bias of these
CD8+ T cells toward reactivity with syngeneic class I
molecules expressed at a high ligand density, suggests that
CD8+ T cells could contribute to the rejection response
against syngeneic MHC class I-positive skin grafts in the
ß2m-/- mice. This interpretation was also used to
explain the rejection of MHC class I-positive skin grafts in
TAP1/ß2m-/- mice (15). However, the idea of a
self-biased CD8+ T cell repertoire in
ß2m-/- mice was not fully compatible with the notion of
the inability of ß2m-/- mice to reject
H-2b-expressing tumor grafts (8, 10). In a more detailed
analysis of this phenomenon, Jhaver et al. suggested that these
findings could be explained by defects of the CD8+ T cells
from ß2m-/- mice to proliferate and secrete cytokines
(e.g., IFN-
or IL-3/granulocyte-macrophage-CSF) upon stimulation
with targets expressing normal levels of self MHC class I molecules
(33). This led them to suggest that the CD8+ T cells from
ß2m-/- mice are in a state of partial but not complete
tolerance (i.e., split tolerance), leading to an inability to reject
H-2b tumors while generating strong cytotoxic activity
against the same tumors in vivo. To explain the discrepancy in the
ability to reject skin grafts but not tumor grafts, it was speculated
that generation of cytotoxic responses might be sufficient to reject
skin grafts despite no (or only moderate) proliferative responses,
whereas eradication of rapidly growing tumors may require additional
mechanisms such as the ability to rapidly proliferate and secrete
cytokines (33). The results in the present study indeed confirm that
naive ß2m-/- mice are unable to reject syngeneic MHC
class I-positive grafts, yet they demonstrate that preimmunization
(with either allogeneic or syngeneic cells) readily primes the mice to
reject syngeneic MHC class I-expressing tumor grafts. Thus, the present
observations suggest that the state of "split tolerance" toward MHC
class I-positive tumor grafts in MHC class I-deficient mice may be
relative rather than absolute.
Our observation that MHC class I-deficient mice accept MHC class I-deficient tumor grafts agrees well with the previous findings that these mice accept MHC class I-deficient bone marrow grafts (17, 20, 21). However, the present results do not exclude the possibility that MHC class I-deficient mice, such as the ß2m-/- strain, can kill ß2m-/- or other MHC class I-deficient cells under certain conditions. Indeed, Höglund and collaborators have recently demonstrated that purified NK cells from ß2m-/- mice can selectively discriminate between MHC class I-positive and class I-deficient tumor cells in vitro, and selectively kill the latter (P. Höglund, personal communication).
In conclusion, the present study illustrates how an MHC class I-deficient environment affects the ability of mice to respond to allogeneic, syngeneic, and MHC class I-deficient tumor grafts. The mice have retained an ability to reject allogeneic tumors; they have acquired an ability to reject syngeneic MHC class I-positive tumor grafts and, in contrast to wild-type mice, have lost the ability to reject MHC class I-deficient tumors.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Hans-Gustaf Ljunggren, Microbiology and Tumor Biology Center, Box 280, Karolinska Institute, S-171 77 Stockholm, Sweden.
3 Abbreviations used in this paper: ß2m, ß2-microglobulin; B6, C57BL/6 mice.
Received for publication June 26, 1997. Accepted for publication September 19, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsDiscussionReferences |
|---|
ß T-cell receptor cytotoxic T cells with major histocompatibility complex (MHC) class I antigen-directed cytotoxicity in ß2-microglobulin, MHC class I-deficient mice. Proc. Natl. Acad. Sci. USA 90:2837.This article has been cited by other articles:
|
|
 |
|
  E. Schott, R. Bonasio, and H. L. Ploegh Elimination In Vivo of Developing T Cells by Natural Killer Cells J. Exp. Med., October 20, 2003; 198(8): 1213 - 1224. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Hayashi and D. L. Faustman Role of Defective Apoptosis in Type 1 Diabetes and Other Autoimmune Diseases Recent Prog. Horm. Res., January 1, 2003; 58(1): 131 - 153. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. D. Schell, L. M. Mylin, S. S. Tevethia, and S. Joyce The assembly of functional {beta}2-microglobulin-free MHC class I molecules that interact with peptides and CD8+ T lymphocytes Int. Immunol., July 1, 2002; 14(7): 775 - 782. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Hayashi and D. Faustman NOD Mice Are Defective in Proteasome Production and Activation of NF-kappa B Mol. Cell. Biol., December 1, 1999; 19(12): 8646 - 8659. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Korten, E. Wilk, J. E. Gessner, D. Meyer, and R. E. Schmidt Altered Donor and Recipient Ly49+ NK Cell Subsets in Allogeneic H-2d -> H-2b and H-2b -> H-2d Bone Marrow Chimeras J. Immunol., December 1, 1999; 163(11): 5896 - 5905. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
