|
|
||||||||

*
Centro Nacional de Biología Fundamental, Instituto de Salud Carlos III, Madrid, Spain; and
Max-von-Pettenkofer Institute, Ludwig-Maximilian University of Munich, Munich, Germany
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
We are interested in studying whether the coexpression of several Ags in the same cell compromises the efficiency of presentation to CD8+-specific T cells, particularly when they are restricted by the same MHC class I molecule. The endogenous pathway of viral Ag presentation to CTL comprises several steps, including proteolytic processing by proteasomes or other enzymes (4, 5, 6, 7), peptide transport to the endoplasmic reticulum (ER)3 by TAP, assembly of the peptide/MHC complex, and migration to the cell membrane. It is possible that different Ags may compete with each other at any of these stages. This question was addressed in a mouse model by studying the interference between two Ags recognized by CD8+ CTL, both presented by the murine MHC class I molecule Ld (8, 9). One is derived from HIV-1 strain IIIB, the immunodominant determinant 318RGPGRAFVTI327 of the envelope gp160 (ENV), which is also presented by various murine (Dd (10), H-2q, H-2u, and H-2p (11)), and human class I molecules (HLA-A2 (12), -A11 (13), -A3 (14), and -B27 (15)). The second epitope is derived from murine CMV, corresponds to the immunodominant 168YPHFMPTNL176 nonamer of the nuclear phosphoprotein pp89, and can elicit CD8+ T lymphocytes that protect against lethal murine CMV infection (16, 17, 18). This report shows that HIV ENV interfered with presentation of CMV pp89 epitope (9pp89) to CTL, and that this occurred at the level of processed peptides bound to MHC class I molecules in the coinfected cells.
| Materials and Methods |
|---|
|
|
|---|
BALB/c mice (H-2d haplotype) were bred in our colony.
Cell lines
All cell lines were cultured in IMDM supplemented with 10% FCS and 5 x 10-5 M 2-ME. Generation of site-directed Ld mutants and their transfection into murine L fibroblasts is described elsewhere (19, 20). Transfectants are named according to the new amino acid substituting for the native amino acid at the indicated residue number. Accordingly, transfectant R144K/R145H was renamed as L/Ld144K145H, AB as L/Ld107W116Y, L95 as L/Ld95I, L116 as L/Ld116Y, and A as L/Ld107W116Y155H157K. T2 cells transfected with Ld have been described (21).
Synthetic peptides
Peptides were synthesized in a peptide synthesizer (model 431A; Applied Biosystems, Foster City, CA), purified, and found homogeneous by reversed-phase HPLC. The sequences are as follows: 9pp89, YPHFMPTNL; 10ENV, RGPGRAFVTI; 9ENV, GPGRAFVTI; 9ENV-MN, GPGRAFYTT.
Recombinant vaccinia viruses (rVV)
Generation of rVV-eN-A9pp89A and rVV-HBe viruses has been described previously (22). The former encodes the murine CMV 9pp89 immunodominant epitope biterminally flanked by five alanines and inserted into the amino terminus of the secretory core protein of hepatitis B virus (HBe). rVV-eC-10env encodes the HIV-1 strain IIIB ENV 318327 epitope at the carboxyl end of HBe (7). The rVV-sN-9pp89S recombinant was generated similarly (22), and the sequence at the N-terminal insertion site is . . .mdiGYPHFMPTNLSdpy. . ., where the lower-case letters indicate the native HBe carrier protein N-terminal sequence. As opposed to the other HBe-based chimeras, this one contains the influenza virus hemagglutinin signal in place of the native HBe signal sequence, and the 9pp89 epitope is not flanked by alanines. rVV-ENV-IIIB virus (vSC25) encodes ENV envelope glycoprotein from the strain IIIB of HIV-1 (23). rVV-ENV-MN encodes ENV from strain MN of HIV-1 (10), which contains the homologous epitope 318IGPGRAFYTT327 presented by Dd (10). rVV-encoded proteins relevant to this study contain their respective signal sequences for translocation into the ER. All foreign genes cloned into the rVV used in this study are under the control of the vaccinia early-late promoter 7.5k. rVV-ENV-IIIB and rVV-ENV-MN are based on vaccinia Western reserve (WR) strain whereas the others present Copenhagen background.
T cell lines and cytolytic assays
Polyclonal pp89- or ENV-IIIB-monospecific CTL were generated by
immunization of mice with murine CMV or with rVV-ENV-IIIB,
respectively, as described elsewhere (5, 6), and they were
used as effector cells in standard cytolytic assays after overnight
infection of transfectants with rVV as described (9). The
percentage of specific inhibition was calculated as:
![]() |
Western blot
Cell pellets corresponding to 4 x 106 infected L/Ld cells were loaded onto SDS-PAGE gels, transferred to Immobilon-P membranes (Millipore, Bedford, MA), and developed either with a rabbit antiserum to HBc/HBe (7), or with mAb 96.13.48 specific for the HIV-1 strain IIIB ENV epitope, kindly provided by Dr. A. Toraño (Instituto de Salud Carlos III, Madrid, Spain; unpublished observation). Standard ECL Plus (Amersham, Arlington Heights, IL) procedures were followed.
Cold target inhibition assays
51Cr-labeled P815 cells were prepulsed for 30 min with 1 x 10-6 M 9pp89. Cold targets were unlabeled P815, pulsed with 1 x 10-6 M 9ENV or 1 x 10-6 M 9pp89. All target cells were then washed and mixed for 30 min before adding pp89-specific CTL in a standard 3-h cytolytic assay. The cold:hot target ratio was 30:1.
Isolation of naturally processed peptides
Transfectants were infected in parallel with the different mixtures of rVV at the indicated multiplicity of infection (moi), and 16 h later, naturally processed peptides were extracted from whole cells with trifluoroacetic acid and were purified essentially as previously described (22). Instead of gel filtration, Macrosep 3K centrifugal concentrators (Pall Filtron, Nortborough, MA) were used for the isolation of low-molecular weight peptides. Reversed-phase HPLC fractions were tested in triplicate with 51Cr-labeled P815 cells and pp89-specific CTL for their content in antigenic peptides. Serial dilutions of positive fraction 31 of HPLC runs, which corresponds to 9pp89 immunodominant nonamer, were tested in cytolytic assays performed always in parallel.
MHC/peptide stability assays
T2/Ld cells were incubated overnight at 26°C, washed, and incubated for 2 h at 37°C in PBS containing 0.1% BSA and 500 µM of the different peptides. A control without peptide was also included. After washing, the cells were resuspended in the same buffer (time point 0 h) and were further incubated at 37°C. Aliquots removed at different time points were stained with mAb 30-5-7, which recognizes Ld bound to peptide (24), and FITC-labeled goat anti-mouse Ab, or with the second Ab alone. Analysis of fixed cells was performed in a FACScan (BD Biosciences, Mountain View, CA). Fluorescence index was calculated at each time point as the ratio of mean linear fluorescence of the sample to that of the control incubated without peptide.
| Results |
|---|
|
|
|---|
Coinfection experiments were performed with rVV that express
either HIV-1 ENV or CMV pp89 optimal 9 mer, named 9pp89, in an optimal
chimeric protein context (22). Coinfection of
rVV-eN-A9pp89A together with an excess of the control virus rVV-HBe did
not affect the recognition of infected L/Ld
fibroblasts by CMV-specific CTL (Fig. 1
A). Vaccinia codes for some
unidentified Ags presented by Ld
(9), which by their nature and amount apparently do not
prevent 9pp89 Ag processing and presentation by
Ld (18), nor that of many other Ags
(25). These results show that this is also true even when
vaccinia Ags are expressed in 5-fold excess. Coinfection experiments
with rVV-ENV-IIIB, however, practically abolished recognition of 9pp89
(Fig. 1
A), with an average inhibition of 73 ± 13%
(n = 10). Coinfection experiments in the macrophagic
cell line J774 led to similar specific inhibition of 9pp89 presentation
caused by coexpression of ENV (83 ± 6%, n = 5),
whereas the negative control WR did not cause any interference (3%).
These data showed that although not the rule, in certain combinations
there is interference between different Ags expressed in the same cell
and presented by the same MHC class I allele.
|
|
As a control, the effect of coinfection with the different viruses on
the level of expression of the antigenic protein eN-A9pp89A was checked
by Western blot. As shown in Fig. 3
(upper panel), although coinfection with a 5-fold excess of
another rVV somehow decreased the level of eN-A9pp89A protein, this
happened to a similar degree with the interfering ENV-IIIB and with
wild-type WR. In addition, although coinfection with rVV-ENV-MN
strongly affected eN-A9pp89A protein expression (lane
4), this was not sufficient to affect specific recognition of
infected cells by CTL (Fig. 1
C). A similar reduction in
total protein level after coinfection was detected when protein
expression was analyzed with an ENV-IIIB-specific mAb (Fig. 3
, lower panel). Thus, in conclusion, ENV interfered with CMV
9pp89 Ag presentation and this could not be explained by a selective
reduction in intracellular levels of the antigenic protein.
|
The findings described above are compatible with two situations.
The two full-length Ags or peptides derived from them may compete
intracellularly in the processing and presentation pathway.
Alternatively, peptides derived from both epitopes may bind separately
to MHC, and ENV complexes may antagonize (26)
pp89-specific CTL at the cell surface. The latter possibility was
simulated with synthetic peptides. Uninfected cells loaded with 9ENV
were used as cold inhibitors of 51Cr-labeled
9pp89-loaded target cells, and were found not to block CTL recognition
of 9pp89 (Fig. 4
). As a positive control,
when both cold and labeled cells were preloaded with 9pp89, inhibition
of lysis was observed. Altogether, these results indicated absence of
antagonism between both epitopes, suggesting that interference occurred
within the infected cell.
|
To confirm that the interference between both Ags occurred
intracellularly, the amount of naturally processed peptides extracted
from infected cells was assessed. Peptides that coelute in fraction 31
with the synthetic 9pp89 and that constitute the major antigenic
activity of the CMV protein in reversed-phase HPLC runs were analyzed.
Serial dilution of this material allowed quantitation of 9pp89 obtained
from coinfection experiments with the control rVV-HBe or with the rVV
that expressed the ENV epitope (rVV-eC-10env) (Fig. 5
). The data showed 3- to 10-fold less
rVV-eN-A9pp89A-derived 9pp89 obtained from cells coinfected with ENV
rVV than with control rVV-HBe. Such a decrease in the yield of
endogenously processed 9pp89 peptide was previously found to severely
impair induction of protective immunity to murine CMV
(22). No antigenicity was recovered from infected cells
lacking Ld (data not shown). Thus, inefficient Ag
presentation in ENV epitope-expressing cells correlated with a lower
amount of processed 9pp89 peptide bound to Ld.
Collectively, these results suggested that Ag competition took place
intracellularly before or at the step of peptide binding to
Ld.
|
Because intracellular competition for Ag presentation correlated with lower amounts of naturally processed 9pp89 peptide, constructs that yield less processed 9pp89 should be easier to compete with.
rVV-eN-A9pp89A and rVV-sN-9pp89S viruses express the immunodominant CMV
epitope with different local flanking regions. Quantitation of
naturally processed peptides from both chimeric proteins yielded 5- to
10-fold less 9pp89 peptide from rVV-sN-9pp89S (E. Luderer, M. Del Val,
and U. H. Koszinowski, unpublished data), which correlated
with target cell lysis. Fig. 6
shows
coinfection followed by cytolysis using rVV-eC-10env as competing
virus. The specific inhibition by rVV-eC-10env of rVV-sN-9pp89S
(78 ± 14%, n = 3) was more prominent than that
of rVV-eN-A9pp89A (35 ± 5%) at the same moi ratios (1:5,
Ag:competitor). Thus, inhibition is stronger under conditions where
less antigenic peptide is generated.
|
Point mutations in the MHC class I molecule affect the competition pattern
If peptides would compete for binding to MHC molecules, then MHC
mutants that affect the peptide groove and that selectively modulate
binding of the antigenic peptide (20) or of the competitor
should differ in their capacity to resist competition. Coinfection and
cytolysis assays were performed using L cells transfected with
different MHC class I molecule Ld mutants. Six
transfectants, including L/Ld, from a panel of 11
MHC class I mutant cells (19, 20), were able to present
rVV-eN-A9pp89A to CTL and were tested for competition (Fig. 7
). The results show that rVV-ENV-IIIB
efficiently inhibited 9pp89 presentation in all cases, whereas
inhibition by rVV-eC-10env was only partial, as described above for
L/Ld. Transfectants
L/Ld107W116Y155H157K,
L/Ld95I, and L/Ld144K145H
(the last two not shown) exhibited a similar reaction pattern as
Ld-expressing target cells. In these
transfectants, the degree of inhibition by rVV-ENV was on average
2-fold higher than the inhibition caused by rVV-eC-10env. In contrast,
the two transfectants L/Ld116Y and
L/Ld107W116Y differed from the previous pattern.
Both displayed a significantly lower percentage of specific inhibition
with rVV-eC-10env virus as compared with rVV-ENV, by a factor of around
4 (Fig. 7
). These data show that mutation of MHC class I molecules
alters the competition hierarchy in cells that otherwise share the
whole Ag processing and presentation machinery. Hence, interference
appears to be the result of competition between different peptides for
binding to Ld or mutants thereof. Therefore,
competition probably operates at the step of MHC-peptide complex
assembly or stabilization.
|
|
| Discussion |
|---|
|
|
|---|
The interference by the full-length ENV protein is higher than that by the Ld-restricted 318327 epitope construct. Thus, beside the action of the epitope itself, a contribution of other unknown Ld-restricted epitopes or other properties of the ENV protein to inhibition appear likely. Interestingly, TAP-independent presentation of several HLA-restricted ENV epitopes has been described (27, 28), which can act simultaneously with the TAP-dependent classical presentation pathway. In our murine system, presentation of the 318327 epitope from full-length ENV is TAP dependent (6), while it can follow both a TAP-dependent and a TAP-independent pathway from the chimeric protein (6, 7). Thus, additional ENV-derived peptides processed also in the secretory pathway may constitute an extra source of peptides that may compete with 9pp89 in our experiments. Notably, preferential and specific proteolytic cleavage of ENV protein occurs between residues 315 and 316, adjacent to the Ld-restricted epitope, as detected in several cell lines (29, 30) including the L cells used in our study (31). The long-time residence of ENV in the ER (28) would favor this cleavage. This may also explain the higher efficiency of recognition of the full-length ENV by ENV epitope-specific CTL. The described O-glycosylation of ENV at this epitope sequence (32) may also contribute to the diversity of peptides with potential for competition with 9pp89. A third source of competing ENV-related peptides could be the extracellular space or the cell membrane (33). Thus, the efficient processing of ENV at several relevant subcellular sites may make it a good source of competing peptides of sufficient affinity for Ld. Differences in the epitope itself, as shown in this study for strain MN, or in these other properties, might contribute to the differential pathogenicity of HIV strains.
The experiments with Ld mutants showed different relative competition strength of the two ENV-expressing competing constructs in transfectants that possess identical endogenous processing pathway, and only differ in point-mutated MHC class I molecules. In addition, quantitation of naturally processed 9pp89 peptide in cells expressing the competitor showed lower amounts of antigenic peptide. Also, the amount of processed peptide that was produced under different experimental conditions correlated inversely with the susceptibility to competition by ENV. This set of data therefore suggests that competition is not due to preferential cleavage by the proteasome or to differential transport by TAP of any 9pp89-related precursor peptide (34). Rather, we place the step of competition at a level where MHC class I molecules are involved, particularly at the stage of MHC/peptide complex assembly and stability. Interference would thus be explained by competition between different peptides for binding to Ld. Complex formation occurs in the ER (35), which is where competition may primarily occur. Peptide exchange at a later step in the secretory pathway (36), associated with further points of ENV peptide generation as discussed above, may contribute as well.
Competition should occur under conditions of limited resources. For
example, only strongly binding peptides should complex with limited MHC
class I molecules, whereas under excess amounts of MHC, weaker peptides
would also have the chance to bind. It has been estimated that TAP
translocates 20,000 peptides every minute per cell into the ER, whereas
only
100 peptide-receptive MHC class I molecules are synthesized
during this time (37). In such a situation of MHC
shortage, it is feasible that a potentially higher production in a
relevant site of efficiently binding ENV peptides would lead to a
disadvantage for CMV epitope binding.
Deficient endogenous presentation of an Ag may arise from inefficiencies in any of the sequential processes of degradation, TAP-dependent transport, and MHC class I binding. In this study we showed that the presentation of an immunodominant epitope is also conditioned by the presence of another dominant epitope. Two studies on CTL responses to HIV (2) and lymphocytic choriomeningitis virus (1), showing that weak epitopes are suppressed in the presence of stronger epitopes, although these weak epitopes are able to generate a competent response if expressed in isolation, could be interpreted accordingly. Also, the reverse situation of what is reported in this study has been described; namely, competition between MHC class I alleles for presentation of overlapping viral epitopes that resulted in suboptimal loading of one of the alleles in vitro (38). Altogether, competition for the same MHC class I molecule appears to represent a mechanism that can modulate presentation, whereas it may not operate for MHC class II (39).
These data highlight potential limitations in the design of recombinant vaccines expressing mixtures of immunodominant epitopes from HIV ENV and other pathogens. In multiepitope vaccine studies not involving ENV, observations compatible with partial interference have not been studied further (40, 41). When the ENV epitope was included, a substantial role of the relative location of two epitopes in the efficiency of class I presentation both in vitro and in vivo was seen (41). Whether or not Ag competition by ENV or other Ags is a frequent event cannot be decided yet, although different HIV ENV strains already differ in these properties. It is also not clear whether this would apply to all MHC molecules, because the results reported in this study also stress the importance of MHC sequence differences for the outcome of competition. From what is known about immunodominance of epitopes within a single pathogen (2, 3, 42), Ag competition might not be a rare event. The real situation may lie somewhere between strong competition and complete liberty of presentation. Previously, only intrinsic features of the epitope (MHC affinity and peptide liberation) and its necessary partners (T cell repertoire and APC) have been implicated as influencing immunodominance. Now we describe in detail the mechanism that underlies the further qualitative level that contributes to immunodominance, which is the presence and properties of fellow epitopes. Thus, we show a way in which circumstances external to the epitope itself can preclude its presentation.
Moreover, coinfection of the same cells by HIV and CMV also occurs in vitro (43) as well as naturally, interestingly involving APC, particularly in the CNS (44, 45). In patients with AIDS, CMV causes severe retinitis, is associated with neurological damage, and represents a recurrent primary pulmonary pathogen (46). If ENV would interfere with Ag presentation of human CMV or of other pathogens in AIDS patients as it does with mouse CMV Ags, our finding may have relevance in CTL responses in early stages of asymptomatic HIV-infected individuals if they are infected with CMV or other pathogens.
In summary, in addition to the described properties of ENV to interact with the CD4 molecule on helper T lymphocytes (47) and to induce increased susceptibility to apoptosis in T cells (48), the phenomenon described in this study would contribute to impaired viral Ag presentation to CTL and to a strain-specific HIV-associated progressive loss of immune competence.
| Acknowledgments |
|---|
| Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Margarita Del Val, Centro Nacional de Biología Fundamental, Instituto de Salud Carlos III, Ctra. Pozuelo, Km 2, E-28220 Majadahonda, Madrid, Spain. E-mail address: mdval{at}isciii.es ![]()
3 Abbreviations used in this paper: ER, endoplasmic reticulum; 9pp89, murine CMV pp89 CTL epitope of sequence YPHFMPTNL; ENV, envelope gp160; rVV, recombinant vaccinia; HBe, hepatitis B virus secretory core protein; WR, Western reserve; moi, multiplicity of infection. ![]()
Received for publication June 1, 2001. Accepted for publication August 13, 2001.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A.-H. Hovav, M. Santosuosso, M. Bivas-Benita, A. Plair, A. Cheng, M. Elnekave, E. Righi, T. Chen, S. Kashiwagi, M. W. Panas, et al. X4 Human Immunodeficiency Virus Type 1 gp120 Down-Modulates Expression and Immunogenicity of Codelivered Antigens J. Virol., November 1, 2009; 83(21): 10941 - 10950. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Stevceva, V. Yoon, A. Carville, B. Pacheco, M. Santosuosso, B. Korioth-Schmitz, K. Mansfield, and M. C. Poznansky The Efficacy of T Cell-Mediated Immune Responses Is Reduced by the Envelope Protein of the Chimeric HIV-1/SIV-KB9 Virus In Vivo J. Immunol., October 15, 2008; 181(8): 5510 - 5521. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Guil, M. Rodriguez-Castro, F. Aguilar, E. M. Villasevil, L. C. Anton, and M. Del Val Need for Tripeptidyl-peptidase II in Major Histocompatibility Complex Class I Viral Antigen Processing when Proteasomes are Detrimental J. Biol. Chem., December 29, 2006; 281(52): 39925 - 39934. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Samino, D. Lopez, S. Guil, P. de Leon, and M. Del Val An Endogenous HIV Envelope-derived Peptide without the Terminal NH3+ Group Anchor Is Physiologically Presented by Major Histocompatibility Complex Class I Molecules J. Biol. Chem., January 9, 2004; 279(2): 1151 - 1160. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |