| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2-Microglobulin- Deficient Mice1
* Howard Hughes Medical Institute, Rheumatology Division, Barnes-Jewish Hospital and Washington University School of Medicine, St. Louis, MO 63110; and
Division of Immunology, Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-microglobulin (2m) than murine
2m. However, it is still controversial that Ly-49D
recognizes H-2Dd, as we were unable to demonstrate
the specificity previously reported. Using this one ligand-one receptor
recognition system, function of an NK activation receptor was, for the
first time, investigated in NK cells that are tolerized in
2m-deficient mice. Surprisingly, Ly-49D-killing activity
against ligand-expressing targets was observed with
2m-deficient mouse NK cells, albeit reduced, even though
"tolerized" function of Ly-49D was expected. These results indicate
that Ly-49D specifically recognizes the Chinese hamster MHC class I
molecule associated with Chinese hamster 2m, and
indicate that the Ly-49D NK cell activation receptor is not tolerized
in 2m deficiency. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Detailed studies over the last decade have indicated that NK cell inhibitory receptors fall into two structural categories: Ig-like receptors with type I integral membrane protein orientation, as illustrated by the human killer Ig-like receptors, and mouse gp49B; and C-type lectin-like receptors with type II integral membrane orientation that include the prototypic mouse NK cell inhibitory receptor, Ly-49A. Despite the structural differences, both types of inhibitory receptors deliver negative signals via the same mechanism, involving the immunoreceptor tyrosine-based inhibitory motif (ITIM)3 in their cytoplasmic domain. Phosphorylation of the ITIM recruits and activates the tyrosine phosphatase, SHP-1, that then presumably dephosphorylates molecules in the activation cascade, leading to inhibition.
Other observations also suggested that NK cell target specificity is not solely due to the action of inhibitory receptors. Mouse NK cells have poor killing capacity against human tumor targets that are killed well by human NK cells, and vice versa, regardless of MHC class I expression on the targets. In several experimental systems, there was evidence supporting the existence of NK cell receptors that activate (5). Nevertheless, the activation receptors are not well defined. Candidate activation receptors are often highly related to the inhibitory receptors, and are encoded in the same genomic region. Putative activation receptors, however, do not contain cytoplasmic ITIMs and often contain charged residues in their transmembrane domain that facilitate association with other transmembrane molecules that can provide positive signaling action. For example, Ly-49 receptors lacking ITIMs (Ly-49D and H), but containing charged transmembrane residues, associate with the DNAX-activating protein of 12 kDa (DAP12)/killer cell activating receptor-associated protein (KARAP) molecule, which transduces activation signals through its cytoplasmic immunoreceptor tyrosine-based activation motifs (6). However, the ligands are poorly understood for most NK cell activation receptors. Hence, most candidate NK cell activation receptors are orphan receptors.
In previous studies, we exploited the profound differences between NK cells from C57BL/6 (B6) and BALB/c mice with respect to killing of Chinese hamster ovary (CHO) cells and ultimately defined the specificity of the Ly-49D activation receptor in B6 mice for CHO killing (7). In brief, most targets were killed equally well by B6 and BALB/c NK cells, but only B6 NK cells killed CHO cells efficiently. Subsequently, anti-Ly-49D mAb, 4E4, blocks the CHO cell killing by B6 NK cells, but does not recognize BALB/c NK cells. These data indicate that Ly-49D in B6 NK cells specifically recognizes a ligand on CHO cells and activates killing.
Although the enhanced NK cell killing of target cells lacking MHC class
I expression was first demonstrated in vitro with tumor cells, this
correlation has been extended to otherwise normal cells in vivo. Con A
blasts produced from 2-microglobulin
(2m)-deficient mice are more susceptible to in
vitro killing by NK cells from otherwise syngeneic wild-type mice
(8, 9). Moreover, bone marrow transplant grafts from
2m-deficient mice are rejected by NK cells in
syngeneic wild-type hosts (10). Interestingly, however, NK
cells from 2m-deficient mice do not have
reactivity against autologous 2m-deficient Con
A blasts, and could not reject an autologous
2m-deficient bone marrow graft, which was
strongly rejected in wild-type mice. In addition, these NK cells failed
to kill allogeneic Con A blasts (8, 9), and could not
reject an allogeneic bone marrow graft (11). NK cells from
2m-/- mice also have
depressed NK cell activity against various target cell lines
(12), indicating that
2m-/- mouse NK cells
are tolerant for various target cells, including MHC class I-deficient
missing self target. For some targets, in vitro culture of NK cells
with cytokines can abrogate this tolerance (13). These
data indicate a general tolerance for autologous and allogeneic cells
in 2m-deficient NK cells. The altered function
of NK cells in 2m-/-
mice has been explained by the higher expression levels of inhibitory
NK receptors (14). However, this hypothesis cannot explain
the impaired killing of 2m-deficient NK cells
against MHC class I-deficient target cells, which cannot induce
inhibitory signals. It is also possible that the tolerance could be due
to the dysfunction or the lower expression levels of activation
receptors. But, the specific functions of NK activation receptors in
2m-/- mice have not
yet been analyzed.
In this study, we report the ligand of Ly-49D receptor on CHO cells.
This ligand, a Chinese hamster molecule termed Hm1-C4, resembles
classical MHC class I molecules and is recognized more strongly by
Ly-49D when it is expressed with Chinese hamster
2m. Identification of a ligand for an NK cell
activation receptor also allowed us to examine the issue of NK cell
tolerance in 2m-/-
mice with respect to activation receptors.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B6, BALB/c, and
2m-/- mice backcrossed
to B6 11 times were obtained from The Jackson Laboratory (Bar Harbor,
ME). Mice were housed in a specific pathogen-free facility at
Washington University, under supervision of veterinarians in the
Division of Comparative Medicine. All experiments were approved by the
Insitutional Animal Studies Committee.
mAbs and cytokine
The following mAbs were used: mAb 4E4 (anti-Ly-49D mAb)
(15), PK136 (anti-NK1.1 mAb), B22/249
(anti-H-2Db mAb; gift from T. Hansen,
Washington University, St. Louis, MO) (16), and 2C11
(anti-CD3
mAb). Anti-human CD4 (hCD4) PE mAb was purchased from
BD Biosciences (Sunnyvale, CA). W6/32, anti-human HLA class I mAb,
was purchased from DAKO (Glostrup, Denmark). Human IL-2 was purchased
from Chiron (Emeryville, CA).
Cell lines
Plat-E was kindly provided by T. Kitamura (University of Tokyo,
Institute of Medical Science, Tokyo, Japan) (17).
X63Ag8-653.mIL-3 was kindly provided by H. Karasuyama (Tokyo Medical
and Dental University, Tokyo, Japan) (18). BWZ.36 was
previously described (19). BaF/3 was provided by the late
M. Thomas (Washington University School of Medicine). CHO
(Pro-5) was provided by P. Stanley (Albert
Einstein College of Medicine, Bronx, NY). RMA was provided by K.
Kärre (Karolinska Institute, Stockholm, Sweden). C1498, Daudi,
Phoenix-A, and 293T cells were obtained from American Type Culture
Collection (Manassas, VA). CHO cells were cultured in 
MEM
supplemented with penicillin G, streptomycin, L-glutamine,
and 10% heat-inactivated FCS (Harlan Bioproducts for Science,
Indianapolis, IN). Daudi and Phoenix-A cells were grown in DMEM
supplemented with penicillin G, streptomycin, L-glutamine,
and 10% FCS. Plat-E cells were cultured in DMEM supplemented with
penicillin G, streptomycin, L-glutamine, 10% FCS, 1
µg/ml puromycin, and 10 µg/ml blasticidin. All other cell lines
were cultured in RPMI 1640 supplemented with penicillin G,
streptomycin, L-glutamine, and 10% FCS. BaF/3 cells were
grown in the presence of 10% culture supernatant of
X63Ag8-653.mIL-3.
RT-PCR
mRNAs were prepared from CHO cells, using Fast Track 2.0 kit
(Invitrogen, Carlsbad, CA). RACE- PCR was done using SMART RACE cDNA
amplification kit (Clontech Laboratories, Palo Alto, CA) with Syrian
hamster MHC class I consensus primer sets (AGTTCGTGCGCTTCGACAGCGAC,
CAGGTCCTCGTTCAGGGCCATGTAATC and GAGCCGCGGGCGCYGTGGATG,
TACCCGCGGAGGAGGCGCCCGTC) (20, 21). Based on the sequences
obtained from the 5' and 3' RACE-PCR products, sequence-specific primer
sets were designed and used for the amplification of
full-length Hm1-C1, C2, C5
(ACTCAGATCCGAGATGGGGGCGGTG, GTCTGTCACTCCGTCCATGCCAGGCAG),
Hm-C3 (GGCCCCCGGTAATGTGGACCATGGCTTCTGTTTCC,
GAACCTGTCCCCAGGAGGCAGCTGTCGGAA), and Hm-C4
(ACTCAGATCCGAGATGGGTGCAGTTGCC, GGCTGAACACATTGTCTGTCACTCCGTCCATG).
Murine 2ma and
2mb alleles, and human, rat, and
Chinese hamster 2m were amplified
by PCR with specific primers (GenBank accession no. X57112) (22, 23). Murine 2mc was
generated from murine 2mb allele
using PCR-based mutagenesis.
Retrovirus vectors and transduction
The retrovirus vectors pMX (24) and
pMX-IRES-GFP (pIG) (25) were kindly provided by T.
Kitamura. The pMX-IRES-hCD4 (pI4) was generated by replacing the
IRES-GFP with IRES-hCD4 from CD4-RV vector (gift from K. Murphy,
Washington University) (26). Hm1 and
2m cDNAs from various species were subcloned
into pI4 or pIG. The fusion constructs of Hm1 with red fluorescence
protein (Clontech Laboratories) were also generated to show the ability
of the molecules to be expressed on the cell surface. To introduce
these cDNAs into cells, the plasmids were transfected into Plat-E or
Phoenix-A packaging cell lines with Fugene6 (Roche Molecular
Biochemicals, Mannheim, Germany). Target cells were incubated with the
viral supernatant and 10 µg/ml polybrene (27) or in
culture plates coated with Retronectin (Takara, Osaka, Japan)
(28).
DAZ assay
The DAZ construct was made by fusing the mouse CD3
chain
cytoplasmic domain (aa 52–164), Ly-49A transmembrane domain (aa
40–66), and Ly-49D extracellular domain (aa 70–269). The Ly-49A
transmembrane domain was introduced to avoid any confounding
interaction with DAP12/KARAP. The DA construct consists of Ly-49A
transmembrane domain (aa 40–66) and Ly-49D extracellular domain (aa
70–269). These constructs were retrovirally transduced into BWZ.36
cells to generate DAZ and DA cells. One hundred thousand DAZ or DA
cells were coincubated with the same number of the target cells for
16 h in 96-well plates and lysed with 100 µl chlorophenol red
-galactoside (CPRG) working solution (0.15 mM CPRG, 100 mM 2-ME, 9
mM MgCl2, 0.125% Nonidet P-40 in PBS)
(19). Bacterial--galactosidase (lacZ) activity was
determined by colorimetric assay for CPRG substrate conversion by
reading absorption of each well at 595 and 630 nm for reference using
96-well microplate reader (Spectra Max Plus; Molecular Devices,
Sunnyvale, CA).
Flow cytometric analysis
Cells were stained with fluorescein-conjugated mAbs, as previously described and analyzed, using FACSCalibur flow cytometer (BD Biosciences) (29).
Cytotoxicity assay
Lymphokine-activated killer (LAK) cells were generated, as described previously, with minor modifications (7). On day 5 or 6, adherent LAK cells were separated by mAb 4E4 and CELLection Pan Mouse IgG kit, according to the manufacturer’s instruction (Dynal Biotech, Oslo, Norway). The purity of the Ly-49D+ or Ly-49D- LAK cells was always >80%. The percentages of Ly-49A+ cells were similar on both sorted populations. Separated LAK cells were expanded in culture until day 9–12 and used in cytotoxicity assays that were performed in a standard 4-h 51Cr release assay. Alternatively, fresh NK cells were enriched by depletion of surface Ig+ B cells from total splenocytes using CELLection Pan Mouse IgG kit. Subsequently, NK cells were purified using DX5-coated microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) and used in a standard 4-h 51Cr release assay. Effector cells were incubated with 51Cr-labeled targets at various E:T ratios. After 4 h, supernatants were harvested, and their radioactivity was determined. The percent specific cytotoxicity (% cytotoxicity) was calculated using the following formula: % cytotoxicity = (experimental lysis - spontaneous lysis)/(maximum lysis - spontaneous lysis) x 100. All cytotoxicity assays were carried out in triplicate wells. Data are presented as mean ± SD.
Retrovirus-mediated gene transfer
Retrovirus-mediated gene transfer was done, as described, with minor modifications (30). Bone marrow hemopoietic precursors were isolated following i.p. injection of 5-fluorouracil (150 mg/kg). Cells flushed from femora and tibias of the BALB/c mice were cultured for 2 days in DMEM supplemented with penicillin G, streptomycin, L-glutamine, 15% FCS, stem cell factor (100 ng/ml), IL-6 (100 ng/ml), and IL-3 (10 ng/ml). Bone marrow cells were washed and resuspended in the same medium with an equal volume of viral supernatant in the presence of polybrene and the same cytokines. This infection protocol was repeated on the following day. Recovered bone marrow cells were injected into the tail vein of 8 Gy irradiated recipient BALB/c mice. The splenocytes were harvested for LAK preparation 4 wk after the transplant.
Lung clearance assay
Lung clearance assay was performed as described (7). In brief, target cells were preincubated with 2.5 µg/ml 5-fluoro-2'-deoxyuridine (Sigma-Aldrich, St. Louis, MO) and radiolabeled with 125I-labeled 5-iodo-2'-deoxyuridine (Amersham Pharmacia Biotech, Piscataway, NJ). Mice were injected i.v. with 1 x 106 target cells. Mice were killed 5 h later, and the lungs were counted with a gamma counter. The percentage of the residual radioactivity was calculated as being equal to (residual radioactivity in the lungs/total injected radioactivity) x 100. Data are presented as mean ± SDs from five mice in each group.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Inasmuch as inhibitory receptors of the Ly-49 family recognize MHC
class I ligands and other studies suggested that Ly-49D may recognize a
murine MHC class I molecule (31), the Ly-49D ligand on CHO
cells may also be an MHC class I molecule. We therefore cloned five
Chinese hamster MHC class I cDNAs from CHO cells by the RACE method and
named Hm1-C1-C5 (Fig. 1
A,
GenBank accession nos. AY064386–AY064390), reflecting Hm1 as the
official designation for the hamster MHC locus. We have indicated our
cDNAs as Chinese hamster origin by the "C," and each different cDNA
was numbered. Hm1-C1, C2, C4, and C5 are similar to each other and have
a high degree of homology to classical mouse MHC class I genes (Fig. 1B). Hm1-C1 appears to be an alternatively spliced form of
Hm1-C5, because it is identical except for the deletion of 48 bp in its
cytoplasmic region. Hm1-C3 is less related to the other four molecules,
but is homologous to the nonclassical mouse MHC class I molecule, Qa-1
(data not shown).
|
Specificity of Ly-49D for Hm1-C4
To determine whether Hm1 molecules could be recognized by Ly-49D,
we used a heterologous indicator cell system using an expression
construct for a chimeric molecule consisting of the extracellular
domain of Ly-49D with the Ly-49A transmembrane domain and cytoplasmic
tail of CD3. The resulting DAZ construct was stably expressed in
BWZ.36 cells, containing a NF-AT repeat construct driving the
expression of lacZ. Immobilized anti-Ly-49D mAb, but not other
mAbs, stimulates DAZ cells to produce -galactosidase, which can be
detected by conversion of the colorimetric substrate CPRG (data not
shown), indicating that the DAZ cell line is responsive to
cross-linking of Ly-49D. Upon exposure of DAZ cells to the BaF/3 cells
expressing the different Hm1 cDNAs, only the Hm1-C4 transfectant
induced lacZ activity, suggesting that it was recognized by Ly-49D
(Fig. 2A). Whereas lacZ
activity induced by Hm1-C4 was somewhat lower than that stimulated by
CHO cells, the H-2Dd transfectant did not induce
any lacZ activity and no lacZ activity was induced by any other
transfectant. Although the addition of mAb 4E4 (anti-Ly-49D) alone
can induce weak lacZ activity in DAZ cells, the lacZ activity induced
by the Hm1-C4 transfectant was inhibited by mAb 4E4 (Fig. 2B). By contrast, F(ab')2 of B22/249
did not block CHO cell killing by B6 LAK cells (data not shown).
Comparable results were obtained with Hm1-C4 and
H-2Dd transfectants of another target cell
(C1498) (data not shown). Finally, the BWZ.36 transfectants expressing
the DA construct (DAZ without CD3) were not stimulated by Hm1-C4.
Therefore, the DAZ assay indicated that Hm1-C4 is a potential ligand of
Ly-49D.
|
2m (CHO2m),
as it was recently demonstrated that Ly-49A recognized
H-2Dd + murine 2m, but
not H-2Dd + human 2m
(34, 35). To investigate this possibility, double BaF/3
transfectants of Hm1 genes and CHO2m were
generated by transfection of the various Hm1 cDNAs with pI4 vector and
CHO2m with pIG vector, respectively, and
analyzed in the DAZ assay (Fig. 2
, C and D). Of
the Hm1 + CHO2m double transfectants, lacZ
activity was induced only by the Hm1-C4 cell line. This activity was
higher than that induced by the Hm1-C4 + pIG double transfectant, was
similar to the level induced by CHO cells, and was inhibited by mAb 4E4
(Fig. 2E). These data indicate that none of the other Hm1
molecules are ligands for Ly-49D and that the complex of Hm1-C4 and
CHO2m appears to be recognized by Ly-49D to a
greater extent than Hm1-C4 and murine 2m (see
below). Recognition of Hm1-C4 in cytotoxicity assays
We next investigated whether ligand specificity in the functional
DAZ assay correlated with Ly-49D-mediated cytotoxicity by primary NK
cells. For these studies, we exploited our previous findings that
Ly-49D+ B6 LAK cells killed CHO cells (Fig. 3A), whereas
Ly-49D- B6 LAK cells did not (Fig. 3B). As predicted from the DAZ assay,
Ly-49D+ B6 LAK cells killed Hm1-C4 transfectants,
but not Ba/F3 (Fig. 3C). They were more efficient in killing
of the Hm1-C4 + CHO2m double transfectant
(Fig. 3F). However, none of the other Hm1 transfectants were
killed. Interestingly, the H-2Dd transfectants
also were not killed. Killing of the Hm1-C4 and the Hm1-C4 +
CHO2m double transfectants was due to
recognition by Ly-49D because Ly-49D- B6 LAK
cells did not kill any Hm1 transfectant (Fig. 3, D and
G). Furthermore, the killing of CHO cells, Hm1-C4
transfectant, and Hm1-C4 + CHO2m double
transfectants was blocked by mAb 4E4 (Fig. 3, E and
H). Therefore, we conclude that the NK cell killing assays
recapitulate Ly-49D recognition of Hm1-C4, as indicated by the DAZ
assay.
|
2m on Ly-49D recognition
Having noticed that the recognition of Hm1-C4 may be influenced by
CHO2m, we further analyzed the effects of
2m from other species on recognition by
Ly-49D. To analyze this issue more carefully, we used the human Daudi
target cell, which is deficient in 2m
expression, and double transfectants were produced expressing Hm1-C4
with 2m from murine a, b, or c alleles, and
human, rat, or Chinese hamster. Comparable transfection of the
2m cDNAs was monitored by the expression of
GFP from the second cistron and detectable HLA class I molecules on the
cell surface of Daudi cells, as assessed with mAb W6/32 (data not
shown). Comparable transfection of the Hm1-C4 cDNA on Daudi double
transfectants was monitored by B22/249 mAb (data not shown). Double
transfectants of Hm1-C4 with rat or CHO2m
induced more lacZ activity in DAZ cells than the double transfectants
of Hm1-C4 with human or any of the mouse 2m
alleles (Fig. 4B). Control transfectants with
2m alone did not stimulate DAZ cells,
indicating lack of Ly-49D recognition of HLA class I molecules
associated with any of the various 2m (Fig. 4A). That the induced lacZ
activity was specifically due to Ly-49D recognition was supported by
the absence of lacZ activity in DA cells exposed to any transfectants
(Fig. 4, A and B). In addition, the lacZ activity
induced by double transfectants of Hm1-C4 and rat or
CHO2m was inhibited by mAb 4E4 (Fig. 4C). Because mAb 4E4 alone induces some weak lacZ activity
in DAZ cells, presumably due to cross-linking by the soluble
anti-Ly-49D preparation, there was no blocking observed of the
slight lacZ activity induced by Hm1-C4 and human or mouse
2m (Fig. 4C).
|
2m double transfectants to a greater
extent than they killed Hm1-C4 + murine 2m
double transfectant (Fig. 4H). Ly-49D+
and Ly-49D- B6 LAK cells did not kill Daudi
cells (Fig. 4, D and E) or transfectants with
various 2m constructs alone (Fig. 4, F and G), and the Ly-49D-
B6 LAK cells did not kill any transfectants (Fig. 4I). These
data indicate that Ly-49D recognizes the dimer of Hm1-C4 with rat or
CHO2m better than Hm1-C4 with murine
2m. Specificity of Ly-49D-transduced BALB/c LAK cells
To provide direct demonstration of the specificity of Ly-49D for
Hm1-C4 in killing assays, we used a system in which the Ly-49D receptor
derived from B6 was retrovirally transduced into bone marrow stem cells
from BALB/c, which do not express Ly-49D. The transduced bone marrow
stem cells were adoptively transferred to irradiated BALB/c hosts, and
Ly-49D+ NK cells were then isolated for killing
experiments. The expression levels of Ly-49D or DA molecules were
similar between Ly-49D+ and DA+ sorted LAK
cells from BALB/c mice (Fig. 5A). The
Ly-49D+ BALB/c LAK cells killed CHO cells and the
Hm1-C4 transfectants, but not BaF/3 cells (Fig. 5B). This
killing was completely blocked by mAb 4E4 (Fig. 5D). By
contrast, Ly-49D- BALB/c LAK cells did not kill
the targets (Fig. 5C). As a control, BALB/c LAK cells were
also prepared expressing the DA construct. Surprisingly, these NK cells
also killed CHO cells (Fig. 5E), but not BaF/3 cells or the
Hm1-C4 transfectants, and this killing was also blocked by mAb 4E4
(Fig. 5G). DA-negative BALB/c LAK cell did not kill targets
(Fig. 5F). These data suggest that the activating signal
mediated by DAP12/KARAP may not be solely responsible for the CHO cell
killing and that specific binding mediated by the DA molecule may also
contribute. Nevertheless, these gene transfer data indicate that Ly-49D
directly recognizes Hm1-C4.
|
2m-/- mice
Because 2m-deficient NK cells have a
general tolerance for autologous and allogeneic cells, it was analyzed
whether the function of Ly-49D is also inactive in
2m-/- mouse NK cells.
To analyze fresh NK cell functions, RMA transfectants were used,
because BaF/3 cells were killed by fresh splenocytes (data not shown).
Ly-49D was equally expressed on B6 and
2m-/- spleen NK cells,
as previously reported (36, 37) (Fig. 6A). Surprisingly, fresh NK
cells from 2m-/- mice
specifically killed RMA transfectants with Hm1-C4; albeit this
cytotoxicity was somewhat lower than that mediated by B6 NK cells (Fig. 6, B and C), suggesting that Ly-49D function is
not tolerized in 2m-/-
NK cells.
|
2m-/- NK cells for
MHC class I-deficient cells (13). The expression levels of
Ly-49D were similar between Ly-49D-sorted LAK cells from B6 and
2m-/- mice (Fig. 6E). Although Ly-49D-positive
2m-/- LAK cells killed
BaF/3 transfectants with Hm1-C4, but not BaF/3 cells, the cytotoxicity
was slightly lower than that mediated by Ly-49D-positive B6 LAK cells
(Fig. 6, F and G). This killing of Hm1-C4
transfectants was specific because Ly-49D-negative B6 and
2m-/- LAK cells did
not kill either the transfectants or BaF/3 cells (Fig. 6, F
and G), and the killing was blocked by mAb 4E4 (Fig. 6H). Similarly, Ly-49D-positive
2m-/- LAK cells also
specifically killed RMA transfectants with Hm1-C4, but to a lesser
degree than Ly-49D-positive B6 LAK cells (Fig. 6J).
Therefore, Ly-49D function is still active in
2m-/- LAK cells, but
the lower specific killing activity of
2m-/- NK cells using
Ly-49D-Hm1-C4 interaction is not completely normalized after in vitro
culture, in the absence of host environmental milieu.
In lung clearance assays of NK cell function in vivo,
2m-/- mice were able
to reject i.v. inoculated RMA cells transfected with Hm1-C4, but they
retained significantly higher levels of radioactivity than B6 mice
(Fig. 6L, p = 0.005). The capacity to reject
RMA transfected with Hm1-C4 is specific because
2m-/- and B6 mice did
not eliminate inoculated RMA cells. Treatment of either
2m-/- or B6 mice with
anti-NK1.1 mAb 3 days before the assay resulted in an increased
retention of RMA transfected with Hm1-C4, to the level of parental RMA,
consistent with a role of NK cells in specific tumor clearance from the
lung. These data demonstrate that the in vivo function of Ly-49D is
also active, but altered in
2m-/- mice.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2m.
We showed this receptor-ligand specificity using four different
effector cells (DAZ cells, Ly-49D+ B6 LAK cells,
Ly-49D-transduced BALB/c LAK cells, and fresh NK cells) and four
separate targets (BaF/3, C1498, Daudi, and RMA cells) to avoid
potential artifacts. Although we were unable to demonstrate Ly-49D
specificity for H-2Dd, as previously reported
(31), these studies with a xenogeneic MHC class I ligand
strongly suggest that Ly-49D recognizes a
2m-dependent epitope on an MHC class I
molecule.
With respect to ligand specificity of Ly-49D, it is well known that
Ly-49-inhibitory NK receptors recognize classical MHC class I
molecules. Similar to Ly-49A recognition of H-2Dd
with mouse 2m, but not with human
2m (34, 35), Ly-49D also
recognizes the dimer of Hm1-C4 and CHO2m.
However, Ly-49D preferentially recognizes a xenogeneic MHC class I
molecule in the context of xenogeneic 2m,
either hamster or rat, which displays only 80% amino acid homology.
Although Ly-49D can recognize Hm1-C4 with murine
2m, we could not find any difference in the
recognition of Hm1-C4 with different mouse 2m
alleles having only one amino acid difference between alleles. Double
transfectants of human 2m and Hm1-C4 in Daudi
cells were not recognized well by DAZ cell, but were killed well by
Ly-49D+ B6 LAK cells (Fig. 4, B and
H). This discrepancy suggests the existence of the other NK
receptors on B6 LAK cells, which may recognize Hm1-C4 and human
2m. Although these data could be interpreted
to suggest that mouse or human 2m do not allow
appropriate folding of Hm1-C4 H chain, we prefer the explanation that
Ly-49D recognizes both Hm1-C4 and 2m,
analogous to Ly-49A recognition of both H-2Dd and
2m, even though Ly-49D is thought to recognize
the 1 and 2 domains of H-2Dd
(38). Ly-49D may bind a dimer of Hm1-C4 and hamster or rat
2m with higher affinity than Hm1-C4 and mouse
2m, consistent with structural analysis of
Ly-49A complexed with its MHC class I ligand.
Interestingly, gene-transferred BALB/c LAK cells with the DA construct could kill CHO cells, but not Hm1-C4 transfectant, and the killing was blocked by anti-Ly-49D mAb. This was a surprising result because DA molecules should not be able to interact with the DAP12/KARAP signaling molecule and transduce activating signals. Yet, our previous work demonstrated that the Ly-49D receptor accounted for the vast majority of the killing activity against a CHO cell line, Pro-5 (7). These data suggest that binding of Ly-49D and its ligand may facilitate the interactions of other activating receptors and ligands with lower affinities (39). However, it is also possible that the DA construct can form heterodimers with other activating Ly-49 receptors in LAK cells or that DA constructs can associate with some unknown immunoreceptor tyrosine-based activation motif-bearing adaptor molecules capable of transducing activation signals in BALB/c LAK cells. DAP12-/- NK cells kill CHO cells (40), but the killing is not blocked by mAb 4E4 (41), and the DAP12-/- mice were on a mixed genetic background that may result in epigenetic effects or there may be different Ly-49 molecules involved. Indeed, the killing of another CHO cell line, CHO-K1, was not completely blocked by anti-Ly-49D mAb (data not shown). Taken together, there may be other ligands for Ly-49D in CHO cells; the current findings indicate that CHO cell killing by B6 LAK cells is mediated by a specific Ly-49D-Hm1-C4 interaction that leads to direct signaling; and specific binding may also contribute to killing independent of direct signaling.
It was reported that CHO cells were recognized by the Ly-49A or G
inhibitory receptors as well as Ly-49D (42). To evaluate
these possibilities, we made transfectants with an AAZ construct
(Ly-49A extracellular and transmembrane domains and CD3 cytoplasmic
domain chimeric construct) or a GAZ construct (Ly-49G extracellular,
Ly-49A transmembrane, and CD3 cytoplasmic domain chimeric construct)
in BWZ.36 cell. The AAZ cells autoreacted with each other (data not
shown), suggesting that the AAZ cell recognizes
H-2Dk on BWZ.36 cell, consistent with known
specificity of Ly-49A for H-2Dk. Evaluation of
CHO reactivity by Ly-49A, therefore, could not be done by the system
reported in this study. In contrast, the GAZ cell did not react to CHO
cells (data not shown). Therefore, it remains controversial whether
Ly-49G recognizes CHO cells, but it is unlikely to influence the other
results reported in this study.
A number of functional studies suggest that Ly-49D recognizes the mouse
MHC class I molecule, H-2Dd. In allogeneic bone
marrow transplantation assays, the Bennett group determined that
rejection of transferred bone marrow from H-2Dd
mice was abrogated by administration of mAb 12A8, specific for Ly-49A
and Ly-49D (43). In addition, using mAb 4E5 (monospecific
for Ly-49D)-sorted NK cells in killing assays against Con A blasts from
different mouse strains, George et al. (44, 45)
demonstrated that sorted
Ly-49D+G2- NK cells from
B6 mice could kill Con A blasts from the H-2Dd
transgenic mouse line, D8, but not from B6 (H-2b)
or B10.BR (H-2k). Mason et al. (46)
showed that the stimulation of transfected H-2Dd
expressed on rat YB2/0 target cell led to DAP12/KARAP phosphorylation
and production of IFN-
in sorted
Ly-49D+A-G2-
NK cells from B6 mice. Nakamura et al. (32)
transfected the rat NK cell tumor line, RNK16, with the Ly-49D cDNA and
obtained stable transfectants (RNK.mly-49D) that gained the capacity to
kill CHO cell (32). RNK.mLy-49D was used to demonstrate
that Ly-49D recognizes transfected H-2Dd
expressed on YB2/0, rat cell line (31). Furthermore,
RNK.mLy-49D gained the capacity to kill Con A blasts from B10.D2
(H-2d) and BALB/c, but not B10.BR
(H-2k), B6 (H-2b), or
BALB.B (H-2b). Thus, in functional assays, Ly-49D
appears to recognize H-2Dd, but not
H-2b- or H-2k-encoded
molecules.
In contrast, the specificity of Ly-49D for H-2Dd
was not confirmed by our functional studies presented in this study,
even though H-2Dd was coexpressed with rat
2m (data not shown), as it was expressed on
rat cell lines in Nakamura’s experiments. In addition, the
Ly-49D+ BALB/c LAK cells did not kill
H-2Dd transfectants in gene transfer experiments
(data not shown). Hm1-C4 is not any more or less homologous to
H-2Dd than the other nonreactive Hm1 molecules.
In addition, H-2Dd tetramers have previously been
shown not to bind Ly-49D transfectants (29, 47).
H-2Dd was transfected in
TAP1-/- fibroblast MEF1TAP1-27 (gift from C.
Harding, Case Western Reserve University, Cleveland, OH)
(48). The susceptibility of the transfectant to
Ly-49A-D+- or
Ly-49A+D--sorted B6 LAK
cells was analyzed in the presence of
H-2Dd-specific peptide to express
H-2Dd on the cell surface. The transfectant was
not killed by Ly-49A+D-
LAK cells in the presence of H-2Dd-specific
peptide, but was killed in the absence of the peptide, indicating that
Ly-49A recognizes H-2Dd and inhibits the killing
(data not shown). However, the transfectant was equally killed by
Ly-49A-D+ LAK cells in the
presence or absence of H-2Dd-specific peptide
(data not shown). All of these data suggest that
H-2Dd is not recognized by Ly-49D.
Perhaps these discrepancies may be explained by several possibilities. First, anti-Ly-49D mAbs are used to sort NK cells for in vitro study, or depletion of an NK cell subpopulation, and the expression pattern of Ly-49D may overlap with unknown NK receptors with related specificities. Although gene transfer experiments may avoid this potential problem, the use of a rodent NK cell line may permit expression of other NK cell receptors. For example, Ly-49A can be expressed on interspecies hybridoma cells between Ly-49A-negative mouse and rat cells (49). In addition, transfection of Ly-49P and W into RNK16 resulted in gaining capacity to kill the transfectant of H-2Dd in YB2/0 cell (50, 51). These findings suggest the possibility that the transfection of Ly-49D, P, or W in the rat cell line RNK16 caused the coexpression of other rat NK receptors recognizing H-2Dd. Furthermore, the RNK16.mLy-49D transfectant could not kill H-2Dd-transfected C1498 or RMA targets and killed only the H-2Dd transfectant in YB2/0 (52). To avoid such difficulties, the receptor-ligand interaction in our system was shown using four effector cells and four target cells. Finally, a potential explanation unifying all of these data is the possibility that Ly-49D recognition of H-2Dd is dependent on the rat cell-derived peptide bound to H-2Dd, a possibility that is inconsistent with peptide-independent recognition of H-2Dd by the Ly-49A inhibitory receptor. Thus, it is still controversial that Ly-49D recognizes H-2Dd.
Nevertheless, the capacity to detect functional recognition of a
specific ligand for an NK cell activation receptor also provided a
means to evaluate NK cell activities that heretofore has been less than
complete. In particular, we explored the activity of this activation
receptor in the context of
2m-/- mice that
display NK cell tolerance, which heretofore has not been evaluated with
respect to potential dysfunction or the lower expression levels of
activation receptors. In the current study, we determined that NK cells
from 2m-/- mice kill
Hm1-C4 transfectants in cytotoxicity assays in vitro and in vivo,
indicating that the function of Ly-49D on
2m-/- NK cells for
ligand recognition and activation is still active.
On closer inspection, however, the killing by
2m-/- NK cells was
reproducibly somewhat lower than that mediated by wild-type NK cells,
and this may be important to consider in more generalized tolerance of
2m-/- NK cells. The
alteration persisted after in vitro culture with IL-2, indicating that
the function of Ly-49D on
2m-/- NK cells appears
not to be normalized by in vitro culture. Consistent with previous
reports, we showed that the Ly-49D expression levels and percentages of
Ly-49D+ NK cells are unaltered in
2m-/- mice (36, 37). Because the expression pattern of Ly-49D cannot explain the
slightly impaired function of Ly-49D in NK cells from
2m-/- mice, our data
suggest that the threshold for killing mediated by activating NK
receptors may be set somewhat higher in NK cells from
2m-/- mice as opposed
to wild-type B6 mice.
Even though the effects on
2m-/- NK cells are
small, it is important to recognize that NK cells may use several
different activation receptors simultaneously to kill MHC class
I-deficient cells because individual NK cells do express multiple
activation receptors. If several are used for killing of MHC class
I-deficient cells, then killing will be due to the sum of the
activities of the involved receptors. Therefore, if the activity of
several of these activation receptors is somewhat impaired, perhaps
some more than others, due to the "threshold" phenomenon, then the
combined effect of impaired signals from activating receptors might
explain the general tolerance of
2m-/- NK cells.
This threshold effect could be due to several potential mechanisms.
Perhaps there is a failure of tolerant NK cells to express
costimulatory receptors. It is also possible that the expressed Ly-49D
on NK cells from 2m-/-
mice is functionally impaired for conformational reasons. This could be
akin to T cell anergy by antagonist peptides, or absence of
costimulatory signals for T cell activation. Also, it remains possible
that there is a compensatory up-regulation in the function of the
inhibitory receptors in the
2m-/- host environment
that results in a nonspecific inhibitory influence. Thus, the
mechanisms to induce the tolerance of NK cells in
2m-/- mice remain
somewhat ambiguous, but, based on our studies of Ly-49D, there appear
not to be global defects on activation receptor expression or profound
effects on activation receptor function.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Wayne M. Yokoyama, Howard Hughes Medical Institute, Rheumatology Division, Barnes-Jewish Hospital and Washington University School of Medicine, 660 South Euclid Avenue, Box 8045, St. Louis, MO 63110-1093. E-mail address: yokoyama{at}imgate.wustl.edu
3 Abbreviations used in this paper: ITIM, immunoreceptor tyrosine-based inhibition motif; 2m, 2-microglobulin; CHO, Chinese hamster ovary; DAP12, DNAX-activating protein of 12 kDa; KARAP, killer cell activating receptor-associated protein; GFP, green fluorescent protein; hCD4, human CD4; LAK, lymphokine-activated killer; IRES, internal ribosomal entry site; lacZ, bacterial--galactosidase.
Received for publication March 15, 2002. Accepted for publication May 1, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-microglobulin-negative (2m-) T-cell blasts by natural killer cells from normal but not from 2m- mice: nonresponsiveness controlled by 2m- bone marrow in chimeric mice. Proc. Natl. Acad. Sci. USA 88:10332.2-microglobulin-deficient NK cells show increased sensitivity to MHC class I-mediated inhibition, but self tolerance does not depend upon target cell expression of H-2Kb and Db heavy chains. Eur. J. Immunol. 28:370.[Medline]
-2 microglobulin allele in mice defined by DNA sequencing. Immunogenetics 22:413.[Medline]
chain deficient mice by retroviral-mediated gene transfer. Blood 95:3071.2-microglobulin subunit. J. Immunol. 166:7327.
) and untransfected BaF/3 cells (
) by
B6 Ly-49D+ LAK cells. Vertical bars show the SD in
triplicate wells; SD bars smaller than the height of the symbols are
not shown. B, Cytotoxicity against CHO cell (
), C3 (
),
C4 (
), C5 (
), or pI4 empty control
vector (•). D, Cytotoxicity by B6 Ly-49D-
LAK cells against the same transfectants in C.
E, To block Ly-49D-mediated recognition of the Hm1-C4
transfectant (square) or CHO cell (circle) by Ly-49D+ LAK
cells, mAb 4E4 (anti-Ly-49D, open symbols) or isotype control (mAb
PK136, anti-NK1.1, filled symbols) was added to the cell culture at
the final concentration of 50 µg/ml. F, B6
Ly-49D+ LAK cell cytotoxicity against BaF/3 double
transfectants expressing Hm1-C4 and CHO
