| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Institute of Molecular Immunology, GSF National Research Center for the Environment and Health, Munich, Germany;
Institute of Clinical and Molecular Virology, Friedrich Alexander University of Erlangen-Nurnberg, Erlangen, Germany;
Institute of Anthropology and Human Genetics, Ludwig Maximilians University, Munich, Germany;
Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Prague, Czech Republic;
¶ Max von Pettenkofer Institute, Department of Virology, Ludwig Maximilians University, Munich, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Some HCMV immune-modulating proteins affect the core of cellular immunity through their interference with expression of MHC molecules (reviewed in Refs. 1 and 2). The impact of MHC class Ia and Ib modulation by these viral genes on NK cell-mediated responses is controversial, since many studies relied on model systems using cells transfected with individual selected genes. Four proteins encoded by the US2–11 region were found to down-regulate class I surface expression by different mechanisms, i.e., relocation of MHC heavy chains to the cytoplasm by US2 and US11 (9, 10), retention of class I molecules in the ER by US3 (11), and prevention of peptide transport through TAP by US6 (12, 13). Deletion of the entire region containing the US2–11 genes from the viral genome has previously been shown to prevent down-regulation of MHC class I molecules, indicating that virtually all MHC down-regulating genes are located within this region (14). US2–11-mediated class I modulation has been postulated to function as a viral defense mechanism, allowing infected cells to escape recognition by class I-restricted, HCMV-specific CTL (15). Thus, the role of class I down-regulation by US2–11 genes as well as the influence of additional class I modulating genes, such as gpUL40, with respect to NK cell regulation have not been evaluated under conditions of natural HCMV infection.
NK cell activity is mainly regulated by class I molecules in a negative fashion via receptor-mediated inhibition (reviewed in Refs. 16, 17, 18). Therefore, an HCMV-mediated class I down-regulation may affect NK recognition of infected cells. A high density of MHC molecules results in a turning off of signals for NK cells that are triggered by interactions between class I molecules and inhibitory receptors (IR). The specificity of each NK cell is characterized by its individual expression of one or more IR that mediate inhibition following contact with corresponding class I ligands. According to the missing self hypothesis, in situations of loss or substantial reduction of total class I expression or of individual allelic products, NK cells may no longer receive adequate inhibitory signals, and hence, they will initiate cytotoxicity (16, 17, 19, 20, 21, 22). Three receptor families are known to regulate human NK cells. Killer cell Ig-like receptors (KIR; CD158 family) bind exclusively to classical HLA-A, -B, and -C molecules, and the Ig-like transcripts (ILT), also known as leukocyte Ig-like receptors (LIR; CD85 family), bind some classical HLA molecules and the non-classical HLA-G molecule (16, 23, 24).
C-type lectin molecules form the third receptor family composed of CD94
and NKG2A heterodimers that bind non-classical HLA-E molecules
(25, 26). Stable HLA-E surface expression requires binding
of particular nonamer peptides that are often derived from leader
sequences of class Ia molecules (25, 27). In transient
transfection systems a peptide capable of stabilizing HLA-E was
identified within the leader sequence of the HCMV-encoded protein
gpUL40 (28, 29). The relevance for NK recognition of
enhanced HLA-E expression in HCMV-infected cells by a gpUL40-derived
peptide supply is discussed controversially (2). On the
one hand, HLA-E was found to be refractory to US2–11-mediated
down-regulation and thus was able to confer resistance to lysis by a
single NK line in an allogeneic infection system of human fibroblasts
(28, 30). On the other hand, blocking of the HLA-E
receptor complex CD94/NKG2A did not augment NK lysis (31).
Transient transfection of gpUL40 alone was not able to protect leukemic
cell lines from lysis by NKL, and exogenous stimulation of IFN-
was
necessary to reach protective levels of HLA-E expression
(32). Moreover, resistance or sensitivity to NK lysis is
influenced by virus strain-specific characteristics (30, 33). Most discrepancies may be explained by experimental
details. The two known HLA-E alleles, EG and
ER, differ in their capacity to inhibit
CD94/NKG2A-regulated NK cells (27). Differences in
US2–11-mediated down-regulation with respect to individual class I
alleles (34) may also account for opposing effects.
Artificial overexpression systems of single genes do not necessarily
reflect the natural HLA-E expression and, therefore, may not allow
predictions and general conclusions with respect to the complex
regulatory processes occurring during natural infection. Most infection
systems used allogeneic fibroblasts that were lysed by NK cells before
CMV infection (28, 30, 32, 33). In this setting the
natural HLA-E expression by uninfected fibroblasts was either too
low or they failed to express the correct allele to confer protection.
While an allogeneic setting represents an important model to elucidate
the role of allogeneic NK cells in bone marrow transplantation
(35), HCMV effects such as gpUL40-mediated enhancement or
US2–11-mediated down-regulation of HLA-E may be difficult to
observe.
To imitate an autologous situation we established an infection model in which uninfected fibroblasts were resistant to lysis by two well-characterized NK lines. In the fibroblasts the natural levels of expression of appropriate class I alleles provided protection from NK-mediated lysis. Within this setting we analyzed the impact of class I down-regulation vs stabilization of HLA-E on NK cell function during infection using a set of mutant viruses. The viral mutants carried deletion of gpUL40 as a single gene, deletion of the entire region between US2 and US11, or both. Viruses were analyzed for their potential to modulate class Ia and class Ib expression as well as their influence on lysis by the defined NK cell lines.
In this experimental model deletion of UL40 from the viral genome had virtually no effect on HLA-E expression, nor did it influence the lysis of infected fibroblasts by the two NK cells that were differentially regulated by class Ia or class Ib molecules. In contrast, fibroblasts infected with viral mutants carrying deletions of the entire US2–11 region maintained their high HLA-A, -B, -C, and -E expression and therefore remained resistant to NK lysis. This observation confirmed previous reports using an independently generated US2–11 mutant virus that the US2–11 region contains virtually all class I down-regulating genes (14). The comparison of the single mutants and a double mutant lacking both UL40 and the US2–11 region allowed us to demonstrate a hierarchy in which MHC class I down-regulation by US2–11 genes dominates over the HLA-E stabilization mediated by gpUL40 protein.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Virus and cell culture were performed as previously described (36, 37). Briefly, primary human foreskin fibroblasts (HF) and MRC-5 cells (BioWhittaker, Verviers, Belgium) were cultured in DMEM, supplemented with 5% FCS, 2 mM glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin sulfate (Invitrogen, Carlsbad, CA). MRC-5 cells were used for virus reconstitution, virus stock propagation, and determination of virus titers by standard plaque assays (36, 37). HF cells were used for NK cytotoxicity assays and cell phenotype studies.
Generation of virus mutants
An enhanced green fluorescence protein (EGFP)-loxP-cassette was
introduced into the bacterial artificial chromosome (BAC) clone
HB5-loxP by shuttle mutagenesis (36, 37) using flanking
homologies between BAC and US11/12 (nt 200,173–202,063). Excision of
the BAC-EGFP cassette from AD169
US2–11-EGFP-loxP with recombinase
Cre (pBRep) resulted in the mutant virus US2–11. The
US2–11UL40 mutant was generated by deleting the UL40 gene in the
US2–11-BAC by introduction of a kanamycin resistance gene
(pcp015-UL40) using linear recombination in a recombination-proficient
bacterial strain. The UL40-BAC was generated by searching a random
transposon library of the cloned AD169 genome (36, 37),
and reconstitution of the virus mutant was achieved by transfection of
the BAC in the presence of pBRep into permissive fibroblasts. Correct
tn insertion was verified by direct BAC sequencing. Additionally, the
entire gpUL40 gene was sequenced in AD169-BAC and showed complete
identity with the published HCMV sequence (38) (data not
shown). In the recombinant virus RVAD169 the US2–6 region was
reintroduced into HB5-loxP(US2–6) to reconstitute the viral
wild-type genome as described previously (36, 37). Thus,
all viral mutants represent clonal derivatives of the original clone
AD169-BAC (36, 39) that can be compared directly without
complications connected with the potential generation of viral variants
by several rounds of plaque purification of the recombinant virus from
the wild-type pool.
Growth curves of AD169-wild type (WT), RVUS2–11, RVUL40,
RVUS2–11UL40, and RVAD169 viruses
MRC-5 cells were infected with either wild-type virus AD169
(AD169-WT) or BAC reconstituted viruses RVUS2–11, RVUL40,
RVUS2–11UL40, and RVAD169 with an multiplicity of infection
(MOI) of 0.1. Virus titers from cells and supernatant were determined
in duplicate by a standard plaque assay at the time points (days)
indicated (37).
Protein expression of gpUL40 by Western blot analysis
Cell lysates of RVAD169, RVUS2–11, mock-infected, AD169-WT,
RVUL40, and RVUS2–11UL40-infected HF were separated on a
SDS-polyacrylamide gel and Western blot analysis was performed using an
anti-UL40 rabbit serum. DNA encompassing the entire reading frame
of gpUL40 (AD169) was amplified by PCR and inserted into the expression
vector pcDNA3 (Invitrogen). Integrity of the coding sequence was
confirmed by DNA sequencing. 293T cells were transfected with the
respective plasmid DNA using Lipofectamine Plus (Invitrogen) according
to the manufacturer’s suggestion, except that the transfection mixture
consisted of 1 µg DNA, 95 µl DMEM, and 6 µl Lipofectamine Plus
reagent. The mixture was added to a cell culture dish (3.5-cm diameter)
seeded with 2 x 105 cells the day before.
After 48 h cells were harvested, washed three times with PBS, and
used for immunoblot analysis. A rabbit antiserum was raised against a
fusion protein comprising aa 101–189 of gpUL40 fused to GST using the
pGEX-6P vector (Pharmacia, Stockholm, Sweden). The fusion protein was
purified on glutathione-Sepharose (Pharmacia) according to the
manufacturer’s instructions.
Infection of human fibroblasts, NK cell, and target cell lines
Primary HF derived from human foreskin (HLA typing: A30, B13, B49, Cw6, Cw7, and HLA-ER allele) were cultured in DMEM, 10% heat-inactivated FCS, and infected with an MOI of 2–4 with AD169-WT and virus mutants, respectively. MHC class I expression and susceptibility to NK lysis was analyzed as indicated at 24, 48, and 72 h postinfection (p.i.). The human NK leukemic line, NKL, was provided by M. Lopez-Botet (Madrid, Spain). Constitutive CD94/NKG2A and ILT2 (CD85j) inhibitory receptor expression was proven by flow cytometry using commercially available mAb (Beckman/Coulter, Westbrook, CA) in parallel to each functional experiment. B.3NK cells were generated by allogeneic stimulation and magnetic bead separation as previously described (40). Briefly, PBL of donor B (HLA-A*0101, B*0801, B57, Cw*0602, Cw*0702) were stimulated with irradiated allogeneic PBL derived from donor 3 (HLA-A*0201, A*0301, B*3501, B*3701, Cw*0401, Cw*0602). CD3-CD56+p85.2+CD94+NKG2A- NK cells were separated from the alloreactive line first by depletion of CD4+ T cells, followed by subsequent depletion of CD3+ T cells using magnetic bead separation according to the manufacturer’s instructions (Dynal Biotech, Oslo, Norway). Constitutive p58.2 (CD158b) and CD94 receptor expression was proven by flow cytometry using commercially available mAb (Beckman/Coulter) in parallel to each functional experiment as previously described (40, 41).
HLA-E transfectants of the HLA class I-negative leukemia line K562 and the murine mastocytoma line P815 were generated by transfecting hybrid DNA cloned into the pcDNA3 vector encoding exon 1 of HLA-A2 and exons 2–7 of HLA-E. This construct allows HLA-E surface expression through stabilization of HLA-E with the HLA-A2-derived peptide (27). Surface expression of HLA-E was analyzed by flow cytometry using the pan-HLA class I-specific mAb W6/32 and the HLA-E-specific mAbs MEM-E/04 and MEM-E/06 (data not shown). An HLA-Cw*0701 transfectant of P815 was generated by cloning of a genomic Cw7-fragment into the pHebo vector and transfection. Selection of transfected cells was performed in medium containing 0.4–0.6 mg/ml G418 (42).
mAbs, flow cytometry, and cell-mediated lympholysis
The pan-HLA class I-specific mAb W6/32 (IgG2a, ascites and hybridoma supernatant; provided by J. Johnson, Munich, Germany) was used for surface expression analyses and functional blocking of inhibition as previously described (40). A truncated F(ab')2 of W6/32 was generated from 3.0 l hybridoma supernatant using a commercially available kit (Pierce, Rockford, IL). Complete digestion was proven by gel electrophoresis (data not shown). The HLA-E specific mAb ME-E/04 (IgG1) and the HLA-E+C-specific mAb MEM-E/06 (IgG1) were generated (V. Horejsi, Prague, Czech Republic) from BALB/c mice immunized with bacterially expressed, purified, and refolded HLA-E H chain (provided by J. Strominger, Dana Faber Cancer Institute, Boston, MA). The specificities were defined using the following class I-transfected cell lines: K562-E, K562-Cw3, P815-E, P815-Cw6, P815-Cw7, and a panel of L721.221 transfectants (data not shown). mAb MEM-E/04 only stained K562-E and P815-E transfectants and thus was determined to be HLA-E specific. mAb MEM-E/06 additionally bound to K562-Cw3-, P815-Cw6 and P815-Cw7-transfected cells, indicating specificity for HLA-E and at least some HLA-C alleles (data not shown). P815 and K562 were purchased from American Type Culture Collection (Manassas, VA).
Standard chromium release experiments were performed as previously described (40). Briefly, target cells were labeled with 51Cr for 60 min and exposed to effector cells for 4 h at the E:T cell ratios indicated. The percentage of specific lysis was calculated according to standard methods (43); mean values and SD of triplicates are listed. The percent relative cytotoxic response (% RCR) was calculated using the specific lysis of K562 cells at one E:T cell ratio in each experiment as a reference value of 100%. The percent lysis of other target cells was normalized to this reference value and was expressed as the percentage of RCR at a defined E:T cell ratio (43). For blocking studies, target or effector cells, were preincubated for 30 min with the following mAb: class I-specific mAb, W6/32 (IgG2a, ascites, final dilution, 1/100), W6/32 F(ab')2 (final dilution, 5 µg/ml), MEM-E/04 and MEM-E/06 (IgG1, ascites; final dilution, 1/100), CD94-specific mAb HP-3B1 (IgG2a, ascites; final dilution, 1/100; provided by M. Lopez-Botet), NKG2A-specific mAb Z199 (IgG2b; final concentration, 10 µg/ml), CD158b (p85.2-)-specific mAb (GL183, IgG1; final concentration, 10 µg/ml; both purchased from Beckman/Coulter), and ILT2 (LIR-1-)-specific mAb (CD85j, VM55, IgG1; final concentration, 10 µg/ml; DAKO, Glostrup, Denmark). Redirected lysis experiments were performed after 30 min preincubation of 51Cr-labeled P815 and P815-Cw7 cells with the titrated mAb specific for 2B4 (C1.7, IgG1; Beckman/Coulter), CD16 (3G8, IgG1; Beckman/Coulter) and isotype control mAb (maximum concentration, 100 ng/ml). Mean values of triplicate samples are shown.
Surface expression analyses were performed using the pan class I-specific mAb (W6/32, IgG2a), HLA-E specific mAb (MEM-E/04, IgG1), or isotype control mAbs (UPC10, IgG2a and MOPC21, IgG1) at concentrations of 10 µg/ml or ascites (dilution, 1/500). Following incubation at 4°C for 60 min and two washing steps, secondary PE-labeled goat-anti-mouse mAb (F(ab')2; 115–116-146; Dianova, Hamburg, Germany) was added for 30 min at a dilution of 1/100. Flow cytometry was performed with the FACSCalibur (BD Biosciences, Mountain View, CA) and CellQuest software.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
An attractive approach to investigate the simultaneous roles of
all US2–11 genes in MHC regulation during natural infection is to
employ virus mutants that carry deletions of this region. The HCMV
genome of strain AD169 was recently cloned as a BAC in
Escherichia coli (36, 39, 44). In the current
study the entire region between US2 and US11 was deleted to generate a
virus mutant in which all known MHC down-regulatory genes were removed
(Fig. 1
A). A second loxP site
together with an EGFP reporter gene was introduced into HB5-loxP by
shuttle mutagenesis, yielding the recombinant viral genome designated
ADUS2–11-EGFP-loxP (37). Virus reconstitution in the
presence of pBRe excised the entire BAC-EGFP cassette and yielded
the recombinant virus RVAD169US2–11 (designated RVUS2–11). A
BAC reconstituted virus reflecting AD169-WT characteristics was
previously generated by reinsertion of the US2–6 region into HB5-loxP
(designated RVAD169) (37). The BglII
restriction analysis of DNA derived from mutagenized BAC clones is
depicted in Fig. 1B.
|
UL40) with a tn
insertion at nt 53,700 according to the published HCMV sequence (Fig. 1
B) (38). In addition, a double deletion mutant
(designated RVUS2–11UL40) was generated based on US2–11-BAC
in which the UL40 gene was subsequently deleted by introducing a
kanR gene between nt 53,421 and 53,928 of the
HCMV sequence (Fig. 1B) (38). The
BglII restriction analysis of DNA derived from mutagenized
BAC clones of UL40 and US2–11UL40 (EGFP-loxP) is depicted in
Fig. 1B (lanes 5 and 6;
for details, see Materials and Methods). Low MOI growth
curves comparing growth characteristics of AD169-WT, RVUS2–11,
RVUL40, RVUS2–11UL40, and RVAD169 confirmed the WT growth
characteristics of the virus mutants (Fig. 1C). Expression
of gpUL40 was controlled by Western blot analyses of HF at 48 h
p.i. using anti-gpUL40 rabbit serum (Fig. 1D). The
gpUL40 was detected as a 29-kDa protein in lysates from HF cells
infected with RVAD169, RVUS2–11, and AD169-WT (lanes
1, 2, and 4), but not in mock control cells
or cells infected with RVUL40 or RVUS2–11UL40, respectively
(lanes 3, 5, and 6). The band
of 29 kDa also represented the largest protein detected in lysates of
293 cells transiently expressing gpUL40, although some degradation was
noticed by isolated expression of recombinant gpUL40 (lane
7). The identity of the gpUL40 gene to the published sequence was
proven by sequencing AD169-BAC (data not shown). HCMV US2–11 genes down-regulate MHC class I during infection and govern the cytotoxic activity of NK cells
To evaluate the consequences of infection with AD169-WT or viral
mutants on MHC expression, immunofluorescence staining of class I
molecules was compared in HF mock control cells and HF infected with
AD169-WT or RVUL40 and with RVUS2–11 or RVUS2–11UL40,
respectively, at an immediate early (24 h), early (48 h), and late
phase (72 h). Mock control cells showed bright class I expression by
staining with the pan class I-specific mAb W6/32. Down-regulation of
class I molecules was already initiated at 24 h p.i. with AD169-WT
and RVUL40 mutant (Fig. 2A,
left). In contrast, no class I modulation was observed in HF
cells infected with RVUS2–11 or RVUS2–11UL40 mutant viruses
(Fig. 2A, right). During early (48 h) and late
(72 h) phases, class I expression was significantly down-regulated in
AD169-WT- and RVUL40-infected cells (Fig. 2, B and
C, left) and was most detectable at 72 h
p.i. Both viruses in which the US2–11 region was deleted (Fig. 2, B and C, right) had lost the ability
to down-regulate class I molecules, indicating that all MHC
down-regulating genes are localized within this region of the viral
genome. This observation confirms previous reports localizing the
down-regulating activity of AD169 in the region between US2 and US11 in
a set of independently generated mutant viruses (14).
Moreover, an increase in class I expression was observed following
infection with the two mutants lacking US2–11 genes. This reproducible
effect was independent of gpUL40 and was restricted to class Ia
molecules (data not shown). Down-regulation was specific for class I
molecules, since no alteration was observed in expression of the
adhesion molecule, ICAM-1 (CD54) (data not shown). Recently, some
immediate early HCMV genes have been described to enhance class I
expression in transfected human fibroblasts (45).
|
, D–F, left). The somewhat higher background
levels of lysis of mock control cells seen with B.3NK cells over time
may be explained by the low levels of HLA-C expression (Fig. 2, D–F, right), as detected with an HLA-C-specific
mAb (data not shown). Susceptibility of HF cells infected with the
mutant viruses correlated well with the modulation of MHC expression.
At the immediate early phase (24 h), all infected cells remained
resistant to lysis (Fig. 2D), corresponding to the
substantial levels of class I. However, at early (48 h) and late (72 h)
phases, HF cells infected with AD169-WT and RVUL40 became
susceptible to lysis by both NK lines as a consequence of class I
down-regulation (Fig. 2, E and F). Importantly,
HF cells infected with RVUS2–11 and RVUS2–11UL40 mutant
viruses maintained their resistance to lysis. Thus, deletion of US2–11
genes from the HCMV genome was sufficient to reconstitute the natural
resistant status of uninfected cells against lysis by NK cells.
The contention that resistance to NK lysis was directly mediated by HLA
class I molecules was confirmed by masking MHC expression with the
class I-specific mAb, W6/32 and the corresponding W6/32
F(ab')2. In addition, mAbs MEM-E/04 and MEM-E/06
were used to block HLA-E and HLA-E+C molecules, respectively (Table I). Masking of class I molecules with
W6/32 and the W6/32 F(ab')2 significantly
increased lysis of HF cells by both NK cells, indicating that natural
resistance was mediated by HLA class I molecules. As expected,
F(ab')2 and intact Ab showed identical effects,
and thus an involvement of FcR-mediated Ab-dependent cell-mediated
cytotoxicity (ADCC) could be excluded. HLA-E molecules were responsible
for inhibition of NKL, because masking with both HLA-E-specific mAb
reversed inhibition. For B.3 cells only the HLA-E+C-specific mAb
MEM-E/06 was able to increase lysis, underlining the inhibitory role of
HLA-C and not -E for these NK cells. NKL- as well as B.3NK-mediated
lysis of HF cells 48 h p.i. with those viruses that strongly
down-regulated class I expression, i.e., AD169-WT, RVAD169, and
RVUL40, was further increased by masking remaining class I
molecules.
|
US2–11- and RVUS2–11UL40-infected HF cells was conferred
by HLA-E- and C-mediated inhibition of NK cells. Increased lysis by
B.3NK cells following masking with MEM-E/06, but not MEM-E/04, mAb
supported this contention. The observation that the HLA-E-specific mAbs
were not able to reverse inhibition of NKL cells can be explained by
the weak affinity of these mAb to HLA-E compared with W6/32, which was
proven by increased mAb concentrations.
An interference ADCC could be further excluded, because NKL and B.3NK
cells expressed very low levels of the ADCC receptor CD16, (FcRIII)
(47). To functionally exclude ADCC, redirected lysis
experiments were performed by loading P815 target cells with mAb
specific for CD16 or 2B4, an activating coreceptor for NK cells
(48) (Fig. 3A).
Strong cytotoxic activation was induced in both NK lines by triggering
2B4, and the intensity of the activating signal correlated with the
amount of specific mAb. In contrast, only minimal background lysis of
P815 cells was detected by CD16-mediated signals, demonstrating that
this naturally low CD16 expression was only minimally involved in the
cytotoxicity of NKL (Fig. 3A, left) and B.3NK
(Fig. 3A, right).
|
B). Masking of NKL cells with CD94- and
NKG2A-specific mAb resulted in significantly increased lysis of HF
cells (Fig. 3B, left). Importantly, addition of
ILT2 (LIR-1)-specific mAb had no influence on the natural resistance of
HF cells. Therefore, the CD94/NKG2A inhibitory receptor complex, but
not ILT2, was responsible for HF cell resistance to lysis by NKL cells.
This correlated well with the influence of HLA-E on NKL activity shown
in Table I. The dominant regulation of NKL cells by CD94/NKG2A
interactions with HLA-E was confirmed by masking of CD94, NKG2A, and
class I in an HLA-E transfectant of K562 (Fig. 3C,
left). The almost complete inhibition of K562 lysis by HLA-E
could be reversed by masking the CD94/NKG2A complex as well as HLA-E
expression. With B.3NK cells, only masking with CD158b
(p58.2-)-specific mAb substantially increased lysis of HF cells (Fig. 3B, right), indicating that this receptor, but
not CD94, accounted for resistance of HF cells.
B.3NK cells were also found to express CD94 (40), but in
contrast to the negative regulation of NKL by CD94/NKG2A heterodimers,
the expression of CD94 on B.3NK cells was associated with an increase
in cytotoxicity. This was demonstrated by the significantly enhanced
lysis of P815 cells coated with CD94-specific mAb via FcR-dependent,
redirected activation (Fig. 3C, middle).
Nevertheless, baseline activity as well as CD94-triggered lysis of P815
were strongly inhibited by Cw7 expression in transfected P815 cells,
confirming the specific inhibition of B.3NK cells by HLA-Cw7 molecules,
which correlated with CD158b expression (40). CD94 serves
as a coreceptor for B.3NK activity, because HLA-E expression in K562
and P815 cells substantially enhanced the lysis of transfectant cells
(Fig. 3C, right), although primary induction of
cytotoxicity itself is probably mediated by the major activating
receptors, such as natural cytotoxicity receptors, NKG2D, 2B4, and
others (see Fig. 3A, right). Costimulation of
B.3NK cells through the triggering of CD94 molecules demonstrated that
CD94 expression alone could not predict HLA-E-mediated inhibition, as
negative or positive signal transduction is determined by the
expression of the NKG2A or NKG2C coreceptors, respectively (26, 40). NKG2C expression in B.3NK cells could only be detected by
RT-PCR analysis due to the lack of commercially available
NKG2C-specific mAb (data not shown).
Reintroduction of the US2–6 region reconstitutes AD169-WT characteristics
On the background of the HB5-loxP-BAC, in which the region between
US2 and US6 has been replaced by the BAC cassette (see Fig. 1A) (36), a reconstituted virus designated
RVAD169 was produced by reintroducing the PCR cloned US2-US6 fragment.
Detailed description and analyses of these viruses were provided
previously (36, 37). RVAD169 was included in our analysis
to demonstrate that cloning procedures did not alter viral wild-type
characteristics of the AD169-WT strain. RVAD169 and RVHB5(US2–6)
were compared for their ability to modulate class I expression and the
influence on NK cytotoxicity 72 h p.i. of HF cells.
RVAD169-infected HF cells displayed substantial class I down-regulation
(Fig. 4A) to an extent similar
to the AD169-WT-infected cells shown in Fig. 2C. As
expected, this down-regulation of class I molecules correlated with
increased lysis by both NK lines (Fig. 4B and Table I).
Deletion of the US2–6 region in RVHB5-infected HF cells resulted in a
weak modulation of class I expression (Fig. 5A), which is probably
mediated by the remaining US11 protein. In correlation with this
intermediate alteration in class I expression, cytotoxicity of B.3NK
cells was enhanced over that of mock-infected control cells (Fig. 4B). Deletion of US2–6 genes had only intermediate effects
on class I expression and susceptibility to NK lysis compared with
mock- and RVAD169-infected cells, because class I down-modulation was
incomplete.
|
|
The observation that infection of HF cells with AD169-WT rendered
them susceptible to lysis by both NK lines raised the question of how
gpUL40 contributed to NK cell recognition. The mutants RVUL40 and
RVUS2–11UL40 are both defective in gpUL40 protein expression, as
shown in Fig. 1D. With these virus mutants, the specific
influence of gpUL40 on HLA-E molecules was assessed in the absence or
the presence of the class I down-modulating genes
US2–11. Staining with the HLA-E-specific mAb,
MEM-E/04, demonstrated that HLA-E molecules were almost completely
down-regulated 72 h p.i. without differences between AD169-WT and
RVUL40 viruses (Fig. 5A). As expected, infection with
RVUS2–11 and RVUS2–11UL40 had no significant effect on HLA-E
expression (Fig. 5B). The slightly diminished staining of
RVUS2–11UL40-infected cells was not significant. Moreover, HF
cells infected with AD169-WT or RVUL40 virus showed no significant
differences in susceptibility to lysis by NKL (Fig. 5C). In
contrast, infection with RVUS2–11 or RVUS2–11UL40 conferred
complete resistance to NKL-mediated lysis. A similar pattern was
observed with B.3NK cells, although the B.3NK-mediated lysis of HF
cells infected with RVUL40 was lower than lysis of AD169-WT-infected
cells (Fig. 5D). The loss of gpUL40-derived peptides to
stabilize HLA-E molecules that can costimulate B.3NK cells via their
CD94/NKG2C complexes may be responsible for this minor effect, which
was not significant (see also Fig. 2). HF cells infected with
RVUS2–11 or RVUS2–11UL40 showed similar levels of resistance
as B.3NK cells. Thus, expression of gpUL40 in AD169-WT-infected cells
was not sufficient to prevent HLA-E down-regulation by US2–11 genes.
In contrast, the absence of US2–11 genes in RVUS2–11- and
RVUS2–11UL40-infected cells allowed the expression of class I
and HLA-E molecules at levels similar to those in mock control
cells.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Within the body, nucleated cells normally express sufficient amounts of
HLA class I molecules to neutralize NK cells via their inhibitory
receptors, protecting them from NK attack (50, 51). To
simulate this normal situation, HF cells were chosen to express the
class I ligands that interact with the IR responsible for inhibition of
the defined NK lines used in our system. Specifically, human
fibroblasts expressing high levels of HLA-E molecules and substantial
amounts of HLA-Cw7 molecules were used, since these ligands were
defined as the inhibitory specificities for NKL and B.3NK cells,
respectively (40, 47, 52). The bidirectional role of HLA-E
in these two NK cells was underlined by its inhibitory function for
CD94/NKG2A+ NK cells (NKL) in contrast to its
activating capacity for CD94/NKG2C+ NK cells
(B.3NK). In this model the almost complete loss of class I expression
48 and 72 h p.i. with AD169-WT allowed lysis by both NK lines.
Therefore, the principle of NK activation in the absence of inhibitory
class I molecules was seen with HLA-E-regulated NK cells as well as
with NK cells regulated by class Ia (HLA-Cw7) molecules. These two NK
lines follow the rule of the "missing self" hypothesis by which
every NK cell expresses at least one IR recognizing a self class I
molecule to avoid autoreactivity (19, 20, 50, 51, 53). In
such a setting, MHC class I molecules dominate NK regulation in
response to HCMV-infected cells. This contention was confirmed by the
complete protection of RVUS2–11-infected cells from lysis by the
two NK lines. Moreover, comparison of RVUL40 and the double mutant
RVUS2–11UL40 revealed that the function of class I modulating
proteins was organized in a hierarchy. In principle, since
gpUL40-derived peptides can stabilize HLA-E expression, they should be
able to confer protection to lysis by CD94/NKG2A+
NK cells (28, 29). However, if US2–11 proteins are
expressed simultaneously, their strong capacity to down-regulate MHC
substantially reduces HLA-E surface expression. Therefore, even in the
presence of stabilizing gpUL40 leader peptide, HLA-E is missing from
the cell surface along with HLA class Ia molecules, and thus inhibition
of CD94/NKG2A+ NK cells does not occur. The
discrepancy of our results to those reporting that HLA-E expression is
refractory to US2, -3, -6, and 11-mediated down-regulation may be due
to variations between AD169 strains propagated in various laboratories,
allelic HLA-E differences (27), or an influence of US7–10
genes.
The observation that class I and, in particular, HLA-E down-regulation was already initiated at immediate early time points (24 h), but became significant during early (48 h) and late (72 h) phases of infection, underlines the importance of the kinetics of viral gene expression. Thus, NK cells may be crucial for the elimination of infected cells in the early and late phases, while peptide-specific CTL may preferentially recognize infected cells during the immediate early phase (49, 54, 55). The similar effects seen with NK cells expressing different IR indicate that this is a general principle of NK regulation during natural HCMV infection.
The recent identification of a murine activating NK receptor (Ly49H) associated with inbred strain-specific resistance to murine CMV infection and identification of the corresponding MCMV-encoded ligand (m157) has focused the attention on the activating component of NK regulation (56, 57, 58, 59, 60, 61). In MCMV-resistant mouse strains, the MHC homologue m157 binds to the activating Ly49H receptor that confers NK-mediated protection to viral replication. Recently, another MCMV gene, m152, was shown to down-regulate the H-60 molecule, one ligand for the activating NK receptor NKG2D (62) in addition to the well-known down-regulation of MHC class I molecules (3). In this situation, susceptibility to MCMV infection was conferred by diminished NK activation due to the m152-mediated reduction of H-60 expression. Thus, some viral proteins may play a dual role in evasion of CTL and NK responses simultaneously. The bivalence of a viral proteins that regulate resistance and susceptibility via modulations of activating and inhibitory receptors supports the hypothesis of a close coevolution between virus proteins and NK cell receptors (60, 63). Moreover, NK cells may undergo two different activation phases during MCMV infection: a first line of nonspecific response mediated by cytokines, and a second specific response triggered by defined NK receptors (57). In both situations, however, the balance between positive and negative signals will determine the final NK activity.
Several human activating NK receptors have been identified, such as natural cytotoxicity receptors (64) and coreceptors such as NKG2D (65). Similar to MCMV, HCMV seems to influence the expression of activating ligands for the coreceptor NKG2D either by induction of MHC class I chain-related molecule A (66) or by enhancement of ULBP expression (67, 68). However, the consequences of these modulations for NK cell activity have not yet been evaluated in the context of MHC class I regulation (60). Recent studies have demonstrated substantial heterogeneity among human NK cells with respect to the contribution of activating receptors to lysis of individual target cells (69). The separation of NK cells into various groups may help to explain some of the controversial observations regarding NK activity directed against HCMV-infected cells. If target cells express class I ligands that bind to the corresponding IR, as occurs in an autologous setting, then they should be protected from lysis even if they express NK-activating ligands. On the other hand, if target cells do not express class I ligands corresponding to the IR, as may occur in unselected allogeneic NK populations, then lysis will proceed. However, target cells not expressing ligands for activating receptors will escape NK attack independently of their inhibitory receptors.
In addition to the specific characteristics of target cells and NK
cells, the virus strain may also be important for tuning of the immune
response. The strain AD169 represents the reference laboratory strain
whose complete genome sequence is the only one available to date
(38). The achievement of cloning the genome of AD169 as a
BAC in E. coli (36) for the first time made
genetically stable virus reference material available and guaranteed
the generation of a genetically comparable set of virus mutants to
study NK cell function. Strain variations of AD169 derived from several
sources and cultured in many laboratories for decades may account for
some discrepancies observed using different experimental systems
(30). Compared with the Toledo strain and clinical
isolates of HCMV, AD169 is missing at least one region, termed UL/b',
which comprises 
13.5 kb (70). This region, which was
lost in AD169 due to fibroblast adaptation, may encode additional genes
involved in immune response and tissue tropism of HCMV. Therefore,
cloning and mutagenesis of a clinical isolate of HCMV are important
challenges, and their success will be crucial for analyzing additional
influences on immune function (71). The detailed kinetics
evaluated for the US2–11 genes as a paradigm for viral gene expression
may also determine the appropriate effector cell type for particular
phases during infection.
To date, most immunotherapies to combat HCMV complications following bone marrow transplantation are based on CMV-specific cytotoxic T cells displaying variable success rates (72, 73). An incorporation of NK cells into defense strategies may enhance the efficiency of antiviral immunotherapies (49). The complex and sometimes controversial picture of the various interactions between HCMV-infected cells and NK cells will profit substantially from the complete identification of ligands and receptors governing NK interactions with target cells. Novel viral mutants that allow the analysis of single genes or gene clusters in the context of the entire viral genome will eventually allow the role of NK cells in HCMV immunity to be deciphered.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Gabriele Hahn, Max von Pettenkofer Institute, Department of Virology, Ludwig Maximilians University, Pettenkoferstrasse 9a, D-80336 Munich, Germany. E-mail address: ghahn{at}m3401.mpk.med.uni-muenchen.de
3 Abbreviations used in this paper: HCMV, human CMV; ADCC, Ab-dependent cell-mediated cytotoxicity; BAC, bacterial artificial chromosome; EGFP, enhanced green fluorescence protein; HF, human foreskin fibroblast; IR, inhibitory receptor; KIR, killer cell Ig-like receptor; LIR, leukocyte Ig-like receptor; MOI, multiplicity of infection; p.i., postinfection; tn, transposon; RCR, relative cytotoxic response; WT, wild type.
Received for publication December 14, 2001. Accepted for publication July 11, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and human cytomegalovirus protein UL40 in the HLA-E-dependent protection from NK cell-mediated cytotoxicity. Eur. J. Immunol. 31:2926.[Medline]
8+ T cell receptor genes in the H-2Ld-restricted cytotoxic T cell response against the immediate-early protein pp89 of the murine cytomegalovirus. J. Immunol. 143:4238.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  L. Mandic, M. S. Miller, C. Coulter, B. Munshaw, and L. Hertel Human cytomegalovirus US9 protein contains an N-terminal signal sequence and a C-terminal mitochondrial localization domain, and does not alter cellular sensitivity to apoptosis J. Gen. Virol., May 1, 2009; 90(5): 1172 - 1182. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Mersseman, K. Besold, M. J. Reddehase, U. Wolfrum, D. Strand, B. Plachter, and S. Reyda Exogenous introduction of an immunodominant peptide from the non-structural IE1 protein of human cytomegalovirus into the MHC class I presentation pathway by recombinant dense bodies J. Gen. Virol., February 1, 2008; 89(2): 369 - 379. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Besold, N. Frankenberg, S. Pepperl-Klindworth, J. Kuball, M. Theobald, G. Hahn, and B. Plachter Processing and MHC class I presentation of human cytomegalovirus pp65-derived peptides persist despite gpUS2-11-mediated immune evasion J. Gen. Virol., May 1, 2007; 88(5): 1429 - 1439. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Shin, B. Park, S. Lee, Y. Kim, B. J. Biegalke, S. Kang, and K. Ahn A Short Isoform of Human Cytomegalovirus US3 Functions as a Dominant Negative Inhibitor of the Full-Length Form. J. Virol., June 1, 2006; 80(11): 5397 - 5404. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Guma, M. Budt, A. Saez, T. Brckalo, H. Hengel, A. Angulo, and M. Lopez-Botet Expansion of CD94/NKG2C+ NK cells in response to human cytomegalovirus-infected fibroblasts Blood, May 1, 2006; 107(9): 3624 - 3631. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Barel, N. Pizzato, P. Le Bouteiller, E. J. H. J. Wiertz, and F. Lenfant Subtle sequence variation among MHC class I locus products greatly influences sensitivity to HCMV US2- and US11-mediated degradation Int. Immunol., January 1, 2006; 18(1): 173 - 182. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Wills, O. Ashiru, M. B. Reeves, G. Okecha, J. Trowsdale, P. Tomasec, G. W. G. Wilkinson, J. Sinclair, and J. G. P. Sissons Human Cytomegalovirus Encodes an MHC Class I-Like Molecule (UL142) That Functions to Inhibit NK Cell Lysis J. Immunol., December 1, 2005; 175(11): 7457 - 7465. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. N. Loveland, C.-K. Chan, E. J. Brignole, and W. Gibson Cleavage of Human Cytomegalovirus Protease pUL80a at Internal and Cryptic Sites Is Not Essential but Enhances Infectivity J. Virol., October 15, 2005; 79(20): 12961 - 12968. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Nattermann, H. D. Nischalke, V. Hofmeister, G. Ahlenstiel, H. Zimmermann, L. Leifeld, E. H. Weiss, T. Sauerbruch, and U. Spengler The HLA-A2 Restricted T Cell Epitope HCV Core35-44 Stabilizes HLA-E Expression and Inhibits Cytolysis Mediated by Natural Killer Cells Am. J. Pathol., February 1, 2005; 166(2): 443 - 453. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. T. Barel, M. Ressing, N. Pizzato, D. van Leeuwen, P. Le Bouteiller, F. Lenfant, and E. J. H. J. Wiertz Human Cytomegalovirus-Encoded US2 Differentially Affects Surface Expression of MHC Class I Locus Products and Targets Membrane-Bound, but Not Soluble HLA-G1 for Degradation J. Immunol., December 15, 2003; 171(12): 6757 - 6765. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Herzer, C. S. Falk, J. Encke, S. T. Eichhorst, A. Ulsenheimer, B. Seliger, and P. H. Krammer Upregulation of Major Histocompatibility Complex Class I on Liver Cells by Hepatitis C Virus Core Protein via p53 and TAP1 Impairs Natural Killer Cell Cytotoxicity J. Virol., August 1, 2003; 77(15): 8299 - 8309. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) is
compared with the BAC reconstituted viruses RV
), RV
), the double mutant RV
), respectively. The growth
rates were detected using standard plaque assays in duplicate samples.
D, Western blot analysis of HF infected with RVAD169
(lane 1), RV
; W6/32, IgG2a; 1/100 diluted ascites). For
masking of the inhibitory receptor, NKL cells (E:T cell ratio, 20:1)
were preincubated with CD94 mAb (
; HP-3B1, IgG2a; 10 µg/ml) or
NKG2A mAb (
; Z199, IgG2b; 10 µg/ml; left). The
specificity of B.3NK cells was shown by HLA-Cw7-mediated inhibition
(middle). The Fc
