|Home | About | Journals | Submit | Contact Us | Français|
Dendritic cells (DC) are potent antigen-presenting cells that play a crucial role in antigen-specific immune responses. Thus, the targeting of exogenous antigens to DC has become a popular approach for cancer immunotherapy and vaccine development. In this report, we studied the interplay between murine cytomegalovirus (MCMV) and human monocyte-derived DC. The results showed that an enhanced green fluorescent protein (EGFP)-encoding, replication-competent MCMV vector underwent abortive infection in human DC; this was accompanied by the efficient expression of EGFP. Infection of human DC by this vector resulted in a modest increase in the expression of cell surface proteins associated with DC maturation and has no significant effect on the immunostimulatory function of the cells, as reflected by their ability to support T-cell proliferation in a mixed-lymphocyte reaction. Finally, an MCMV vector encoding the human immunodeficiency virus type 1 (HIV-1) gp120 envelope glycoprotein was constructed and used to infect cultured human DC. The infected DC were shown to be capable of stimulating the expansion of autologous, gp120-specific, class I-restricted T lymphocytes from an HIV-1-negative donor, as determined by tetramer staining and enzyme-linked immunospot analysis. Taken together, these results suggest that MCMV may have potential utility as a vector for human vaccine development.
Dendritic cells (DC) are specialized immune cells with a remarkable capacity for antigen processing and presentation. DC are especially efficient at stimulating naive T cells, and they play a critical role in the induction and regulation of antigen-specific adaptive immune responses (29, 35, 43). There is also considerable evidence that the effectiveness of vaccines against infectious disease agents and tumor antigens may be linked to their ability to target DCs (5). It may therefore be possible to develop effective cancer vaccine(s) and vaccines for important human pathogens, such as human immunodeficiency virus type 1 (HIV-1), by direct immunization with antigen-loaded DC (19) or by targeting antigens to DC in vivo.
One approach to the delivery of exogenous antigens to DC, either in vitro or in vivo, is the use of viral vectors (25). Viral vectors that have been shown to be capable of infecting DC include adeno-associated virus, adenovirus, alphaviruses (Sindbis virus and Venezuelan equine encephalitis virus), herpesviruses (herpes simplex virus type 1 [HSV-1]), lentiviruses (HIV), poxviruses (avipox virus and vaccinia virus), and rhabdoviruses (rabies virus) (8, 9, 12, 15, 17, 20, 21, 31, 32, 42, 45, 48, 52, 55, 61, 62). In many cases, however, infection of DC with viral vectors results in inhibition of cellular maturation or a reduction in immunostimulatory activity (14, 24, 51). Adenovirus and HIV (lentivirus) vectors may represent exceptions to this phenomenon (21, 46), but both vector systems also possess certain disadvantages. In the case of HIV (lentivirus) vectors, important concerns include safety and acceptability to healthy subjects who may be reluctant to become infected with any kind of HIV-1-based vector, no matter how safe. In the case of adenovirus vectors, the efficiency or selectivity of DC gene transfer is something of an issue, because DC do not express the primary receptor for type 5 adenoviruses, known as the Coxsackie and adenovirus receptor (CAR) (46). As a consequence, successful DC gene transfer by standard Ad5 vectors requires the use of a rather larger multiplicity of infection (46); furthermore, it can reasonably be expected that in vivo administration of Ad5 vectors will result in far higher levels of gene transfer into CAR-positive bystander cells than into CAR-negative DC.
In light of the considerations outlined above, it remains unclear which virally based vector system(s) may be most well suited for ex vivo gene transfer into DC (for use in the context of DC vaccine approaches) and for direct in vivo gene transfer into DC after intramuscular, intradermal, mucosal, or transcutaneous delivery. It is not unlikely that different vectors will offer specific advantages for particular applications. Thus, it may be important to fully explore the available range of viruses that could be used for DC gene transfer and vaccine delivery. With this in mind, we have focused our attention on the murine cytomegalovirus (MCMV).
Some of our reasons for wanting to explore the utility of MCMV as a possible vehicle for gene delivery to DC are as follows: (i) both MCMV and human CMV (HCMV) elicit strong and persistent cytotoxic-T-lymphocyte (CTL) responses in their natural hosts, (ii) MCMV has been shown to infect murine peripheral blood mononuclear phagocytes in vivo (56) and to productively infect murine DC both in vitro and in vivo (2), (iii) HCMV has also been shown to productively infect cultured human DC (40, 44, 49), (iv) MCMV has been shown to be capable of entering cells of human origin (28) and of expressing a vectored reporter gene in primary human brain cells (59, 60) or 293 cells (34), and (v) human 293 cells and primary human brain cells are known to be nonpermissive for MCMV replication (34, 59, 60). Collectively, these data suggested to us that MCMV might be capable of entering human DC and of expressing a vectored antigen in these cells, while the well-recognized species specificity of CMVs further suggested that MCMV might be incapable of productive replication in human DC (28, 34, 59, 60). We further hypothesized that the previously reported suppressive effects of CMVs on the function and maturation of species-matched DC (2, 40, 44, 49) might not occur in the context of a nonproductive or abortive mode of viral infection in DC from a nonpermissive species (i.e., in human DC infected with MCMV ). We report here on the results of experiments that were designed to test these hypotheses experimentally.
Our findings show that MCMV can indeed infect human DC, that it can express a vectored indicator gene (enhanced green fluorescent protein [EGFP]) in these cells, and that this is associated with an abortive and nonproductive mode of viral infection. We further show that transduction of human DC by MCMV vectors results in a modest increase in the cell surface expression of costimulatory molecules and markers of cellular maturation and that it has no significant effect on the T-cell stimulatory capacity of the cells, as measured in an allogeneic mixed-lymphocyte reaction (MLR). Finally, we show that human DC transduced by a MCMV vector encoding the HIV-1 gp120 envelope glycoprotein can elicit the de novo expansion of gp120-specific CD8+ T lymphocytes from autologous PBMC of a naive donor.
Human peripheral blood mononuclear cells (PBMC) were isolated from whole blood by using LSM lymphocyte separation medium (ICN, Inc., Costa Mesa, Calif.). Monocytes were then purified by using immunomagnetic microbeads conjugated to a CD14-specific monoclonal antibody (Miltenyi Biotec, Auburn, Calif.). The purity of the monocyte cell population was determined to be ca. 94%, as measured by flow cytometric staining of cells with a phycoerythrin (PE)-conjugated, CD14-specific monoclonal antibody. Purified monocytes were then incubated for 6 to 7 days in RPMI 1640 medium supplemented with 1% autologous plasma plus human recombinant granulocyte-macrophage colony-stimulating factor (R&D Systems, Minneapolis, Minn.) and interleukin-4 (IL-4; R&D Systems), as described previously (50). Culture medium was replenished every 2 to 3 days. After 6 to 7 days, immature DC (iDC) were generated and were characterized by immunophenotyping with monoclonal antibodies specific for CD11b, CD11c, CD14, CD80, CD83, and CD86, as well as class I and class II MHC antigens. Mature DC (mDC) were generated from the iDC cell population by the addition of 1 μg of lipopolysaccharide (LPS; Sigma, St. Louis, Mo.)/ml and subsequent incubation of cells for one additional day.
Phagocytosis by human DC was examined by using a fluorescein isothiocyanate (FITC)-dextran internalization assay. Briefly, DC (105) were resuspended in medium containing 0.1 mg of FITC-dextran (70 kDa; Sigma)/ml for 5 min at 37°C. The reaction was then terminated by using cold phosphate-buffered saline (PBS) containing 0.1% sodium azide, and cells were examined by flow cytometry (after an extensive washing).
MCMV-EGFP and MCMV-gp120 are replication-competent MCMV recombinants which express, respectively, a humanized derivative of the jellyfish green fluorescent protein (EGFP) or the HIV-1 surface envelope glycoprotein (gp120). The constructs were generated by homologous recombination into a molecularly cloned MCMV bacterial artificial chromosome, as previously described (3, 36), by using sequences that were derived from pEGFP-C1 (Clontech) or pJW4303-gp120 (a generous gift of Shan Lu, University of Massachusetts Medical Center). In both cases, the exogenous gene expression cassette was inserted into the nonessential immediate-early 2 (ie2) gene of MCMV (10, 33, 37). The genomic architecture of the two viral recombinants was then examined by restriction digestion and Southern blot analysis (not shown) in order to confirm that the recombinants did not contain any deletions or unexpected sequence rearrangements. After this, titered viral stocks were generated and used to infect M2-10B4 cells. Expression of the encoded exogenous gene products was then confirmed by fluorescence microscopy (MCMV-EGFP) or by immunoblot analysis of infected cell lysates with a gp120-specific goat antiserum (Biodesign; Dunn, Asbach, Germany); the latter analysis resulted in the detection of a major immunoreactive protein species of the expected size (ca. 120 kDa) in cells infected with the MCMV-gp120 construct but not in uninfected cells (data not shown).
Both the virus-permissive cell line M2-10B4 (30) and human iDC were infected with MCMV-EGFP at a multiplicity of infection (MOI) of 5. Culture medium was then collected at 2, 24, 48, 72, 96, and 120 h after virus infection, and viral titers in these culture supernatants were determined by transfer of 10-fold serially diluted supernatants to uninfected M2-10B4 cells. Five days later, 50% tissue culture infective doses were calculated on the basis of virally induced cytopathic effects and EGFP expression in the indicator cells by using previously described methods (36).
M2-10B4 cells and human iDC were infected with MCMV-EGFP at various MOIs (0.5, 1, and 5). After 48 h, the cells were fixed in phosphate-buffered 2.5% glutaraldehyde, postfixed in phosphate-buffered 1.0% osmium tetroxide, dehydrated in a graded series of ethanol to 100%, and infiltrated for 4 h with Spurr epoxy resin. BEEM capsules (Electron Microscopy Sciences, Fort Washington, Pa.) filled with liquid epoxy resin were placed over the cells on the glass slide and polymerized overnight at 70°C. The next day they were dipped into liquid nitrogen five times and detached from the slide by using forceps. The capsules containing the entrapped cells were inverted and examined with a light microscope to determine the appropriate area for thin sectioning. The thin sections were placed onto copper grids, stained with 1.0% uranyl acetate and lead citrate, and examined and photographed by using a Hitachi 7100 electron microscope. Particles that exhibited morphological characteristics typical of MCMV virions were observed in both the cytoplasm and the nucleus of virally infected M2-10B4 cells (which are permissive for productive viral replication). However, no viral particles were detected in any of the sections from MCMV-EGFP-infected iDC that were examined. In addition, few or no cytopathic changes were observed in the MCMV-infected human iDC.
M2-10B4 cells (ATCC no. CRL-1972) and human iDC were infected with MCMV-EGFP at an MOI of 5 or were mock infected. Two days after infection, cell lysates were prepared, and protein was quantified by using a Bradford assay (Bio-Rad, Hercules, Calif.). Equivalent amounts of protein were then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the resulting gel was transferred to nitrocellulose (ECL; Amersham) prior to being probed with monoclonal antibodies specific for MCMV IE1 (croma 101), E1 (20/238/28) (gifts from Stipan Jonjic, Rijeka, Croatia), and a polyclonal antibody specific for the MCMV late protein, glycoprotein B (gB; a gift from John Shanley, University of Connecticut Health Center). The blots were then incubated with sheep anti-mouse horseradish peroxidase-labeled secondary antibody (Amersham), and reactive proteins were detected by using ECL chemiluminescence and X-ray film (Kodak, Rochester, N.Y.).
A total of 105 cells were stained with phycoerythrin (PE)-labeled monoclonal antibodies in fluorescence-activated cell sorting (FACS) buffer (1× PBS [pH 7.2] containing 0.09% sodium azide and 2% fetal bovine serum) for 30 min on ice. The cells were then washed twice with FACS buffer and resuspended in 0.5 ml of the same buffer prior to analysis on a FACSCalibur (BD) by using CellQuest software (BD). The following monoclonal antibodies were used (all were obtained from BD Pharmingen, San Jose, Calif.): PE-HLA-A, -B, and -C (clone G46-2.6); PE-HLA-DR (clone G46-6, L243); PE-CD14 (clone M5E2); PE-CD11c (clone B-ly6); PE-CD83 (clone HB15e); PE-CD80 (clone L307.4); PE-CD86 (clone 2331, FUN-1); PE-CD40 (clone 5C3); and PE-labeled immunoglobulin G1 (IgG1) and IgG2a isotype controls. Dead cells were excluded from the flow cytometric analysis by using propidium iodide staining.
An allogeneic MLR was performed to examine the T-cell stimulatory function of MCMV-EGFP-infected human DC. Briefly, human iDC were infected with MCMV-EGFP or mock infected (no MCMV-EGFP) at an MOI of 5. LPS-matured human DC were also mock infected or infected with MCMV-EGFP under identical conditions. At 48 h after MCMV infection, DC were subjected to a cesium-137 irradiation source for 10 min (3,000 rads) and mixed with allogeneic PBMC (3 × 105 cells/well) at serial twofold dilutions in defined ratios in 96-well U-bottom plates (Falcon), in triplicate. The mixed DC-PBMC cocultures were then maintained for 3 days and, during the final 18 h of this time period, 1 μCi of [H3]thymidine (New England Nuclear, Boston, Mass.) was added/well to each well. At the conclusion of the assay period, cells were harvested by using a 96-well automated harvester (Tomtec, Hamden, Conn.) and counted by using the Microbeta 1450 Trilux counter (Perkin-Elmer Wallac).
An experiment was performed to determine whether infection of human iDC with a gp120-encoding MCMV vector might allow for the in vitro expansion of gp120-specific CTLs from autologous, naive human PBMC. To do this, iDC were either mock infected or infected with MCMV-gp120 at an MOI of 5. Two days thereafter, DC were harvested and counted. Cells were irradiated in a cesium-137 source for 10 min (3,000 rads), in some cases after being pulsed with 10 μg of V3 peptide (RGPGRAFVTI)/ml for 2 h at 37°C.
Autologous PBMC (2 × 106 cells) and DC (1 × 105 cells) were then seeded in each well of a 24-well plate in quadruplicate in a volume of 2 ml of RPMI 1640 medium, supplemented with 10% human AB serum (Sigma), 100 μg of streptomycin/ml, 100 IU of penicillin/ml, and 10 ng of IL-7 (R&D Systems)/ml. Every 2 days, 1 ml of spent medium was aspirated and replaced with 1 ml of fresh medium. After 2 weeks of culture, cells from two wells of each group were harvested and examined by tetramer staining for V3 peptide-specific CTLs. The remaining two wells were further stimulated for two additional weeks with similarly prepared DC. The cells were then harvested and examined for antigen specific CD8 T cells by using tetramer staining and gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assay.
Antigen-presenting cell (APC)-conjugated HLA A2*01 tetrameric complexes containing the HIV-1 gp120 V3 peptide (RGPGRAFVTI) were prepared by the NIH tetramer facility (Emory University, Atlanta, Ga.). A total of 106 PBMC were then reacted with 0.1 μg of an FITC-conjugated anti-human CD8 monoclonal antibody (clone RPA-T8; BD Pharmingen) and with 0.5 μg of the APC-conjugated V3-specific tetramer in FACS buffer. Cells were then mixed well and incubated on ice for 1 h in the dark. After this incubation, the cells were washed with FACS buffer and analyzed on a FACSCalibur (BD) instrument by using CellQuest software (BD). A total of 200,000 events were collected for each sample, and lymphocytes were gated on the basis of forward and side scatter parameters.
To further examine the functional activity of the in vitro-expanded, V3-specific CTL, IFN-γ ELISPOT assay was performed. We coated 96-well plates (Millipore, Mass.) with anti-human IFN-γ monoclonal antibody (clone 1-D1K) from Mabtech (Sweden) diluted in 1× PBS. After a wash with RPMI 1640 medium and blocking with RPMI 1640 medium containing 10% human AB serum, PBMC were seeded into the culture wells in triplicate. Cells were then incubated in the presence or absence of the V3 peptide for 48 h, after which the plates were washed extensively with PBST (PBS containing 0.1% of Tween 20). A biotinylated anti-human IFN-γ monoclonal antibody (clone 7-B6-1; Mabtech) was then added to the wells, followed by incubation for 2 h at room temperature. Plates were then washed again with PBST, and streptavidin-alkaline phosphatase (Jackson Immunoresearch) was added and allowed to incubate for 1 h at room temperature. Vector Blue substrate (Vector Laboratories) was prepared according to the manufacturer's instructions and incubated for 10 min at room temperature in the dark. Plates were then washed thoroughly with distilled H2O, allowed to dry overnight, and analyzed on a blinded basis by using an automated ELISPOT plate reader (Zellnet Consulting, Inc., New York, N.Y.).
Human DC were generated from purified, immunomagnetically selected CD14+ peripheral blood monocytes, after a 7-day culture in the presence of granulocyte-macrophage colony-stimulating factor and IL-4. As expected, the iDC that were derived through the use of this protocol were characterized by their loss of cell surface CD14 expression and by their expression of high levels of HLA-DR, CD11b, and CD11c (e.g., see the data presented in Fig. Fig.4).4). mDC were then generated from this iDC population, by incubating the cells in medium containing 1 μg of LPS/ml for 24 h. These cells were distinguished by their increased expression of HLA-DR and CD83 (data not shown) and by their loss of phagocytic activity (e.g., see the data presented in Fig. Fig.22).
In order to examine the ability of MCMV to transfer exogenous DNA into primary human DC, iDC, or murine M2-10B4 cells (a positive control for virus infection) were exposed to the MCMV-EGFP recombinant virus at an MOI of 5. After 12, 24, and 48 h, the cells were examined by fluorescence microscopy in order to detect EGFP expression. As shown in Fig. Fig.1A,1A, strong EGFP expression was detected in human iDC as early as 12 h after infection. This level of EGFP expression did not, however, increase substantially over time, suggesting that the virus was not replicating or spreading to new cells (see Fig. Fig.33 for additional information on this point). In contrast, the number of EGFP-positive cells in the M2-10B4 cell culture increased rapidly from 12 to 48 h, a finding consistent with virus replication and spread in the fully permissive M2-10B4 cells.
The experiment shown in Fig. Fig.1A1A suggests that ca. 30 to 50% of iDC from this donor could be infected with MCMV-EGFP under the conditions used (MOI of 5; samples analyzed at 12 to 48 h after virus infection). Other experiments confirmed that a dose-response relationship existed between the number of EGFP-positive DC and the MOI of MCMV-EGFP used to infect the cells (data not shown), and flow cytometric analysis confirmed the overall estimate of infection efficiency that was obtained by using the fluorescence microscopic analysis. Specifically, Fig. Fig.1B1B reveals that roughly 30% of the cells from this particular batch of human iDC were successfully transduced by the MCMV-EGFP recombinant under the conditions used (MOI = 5; time = 24 h after virus infection).
In addition to presenting an analysis of GFP expression in human DC at various times after exposure to MCMV-EGFP, the results shown in Fig. Fig.1A1A also include a fluorescence microscopic analysis of GFP expression in iDCs that were incubated with MCMV-GFP in the presence or absence of the protein synthesis inhibitor, cycloheximide. The data show that no GFP could be detected in the DC if cycloheximide was present, indicating that the presence of GFP is dependent upon new protein synthesis and not upon pinocytic or phagocytic uptake by the DC.
We attempted to examine further the possible contribution of phagocytosis to GFP expression in MCMV-EGFP exposed DC, by performing the virus infection in the presence of phagocytosis inhibitors such as cytochalasin B, cytochalasin D, or nocodazole. However, these drugs proved uniformly toxic to our DC cultures when applied at the concentrations necessary to exert a physiologic effect on phagocytosis. We therefore decided to take a different approach and to directly compare the phagocytic activity of human DC (as measured by uptake of FITC-dextran particles) with their susceptibility to infection with MCMV-EGFP. The results (Fig. (Fig.2)2) reveal that (as expected) mDC were weakly phagocytic compared to iDC. However, the two DC populations (mDC and iDC) were essentially equal in their susceptibility to infection with MCMV-EGFP (as revealed by flow cytometric analysis of GFP expression). Thus, the phagocytic activity of human DC does not correlate with their susceptibility to infection with MCMV-EGFP or with their expression of a MCMV-vectored antigen (GFP).
The data presented in Fig. Fig.22 also reveal that our initial estimate of the transfection efficiency for MCMV-EGFP in human DC (30 to 50%; Fig. Fig.1)1) is rather conservative. In the experiment shown in Fig. Fig.2,2, up to 85% of iDC and 70% of mDC were found to be positive for EGFP expression (as measured by flow cytometric analysis) at 24 h after exposure to MCMV-EGFP at an MOI of 5. Subsequent experiments yielded results somewhere in between these values, with most experiments giving ca. 50% transduction efficiency under standard conditions (MOI = 5, time of analysis = 24 h after infection; data not shown). This suggests that there may be some individual-to-individual or batch-to-batch variation in the susceptibility of human DC to infection with MCMV for reasons that are at present uncertain.
One of the characteristic features of CMVs is their highly species-specific host range (28). To examine whether MCMV might be undergoing a nonproductive cycle of infection in human iDC, we compared the one-step viral growth curves for MCMV-EGFP-infected human iDC versus MCMV-infected M2-10B4 cells. No virus progeny was detected in supernatants collected from MCMV-EGFP-infected human iDC, whereas MCMV titers increased steadily in the supernatant of MCMV-EGFP-infected M2-10B4 cells, reaching a peak at 3 days after exposure to the virus and declining thereafter as the cultures underwent virally induced cytopathic effects and cytolysis (Fig. (Fig.3A).3A). Importantly, the transduction of human DC by MCMV was not associated with any significant loss of cellular viability relative to control cultures over the period of this experiment (data not shown).
MCMV replication in human iDC was also examined by electron microscopic (EM) examination of infected cells. EM analysis was performed both for MCMV-EGFP-infected M2-10B4 cells (as a positive control) and for MCMV-EGFP-infected human iDC (as well as for mock-infected iDC, as a negative control). These experiments were performed by using a range of viral MOIs (0.5, 1, and 5) and, 2 days later, the cells were fixed with 4% glutaraldehyde and processed for EM analysis. The results of this experiment revealed the presence of numerous virus particles, at different stages of assembly, in virally infected M2-10B4-infected cells, but not in iDC (Fig. (Fig.3B).3B). In addition, the MCMV-transduced DC exhibited no signs of any virally induced cytopathic effects (not shown).
The absence of detectable MCMV replication in human iDC, combined with the efficient expression of virally vectored EGFP, suggested to us that MCMV might undergo abortive infection in human DC. To verify this, we examined the expression of viral immediate-early (immediate-early gene 1 [IE1]), early (early gene 1 [E1]) and late (gB) genes in human iDC that had been infected with MCMV-EGFP. Permissive M2-10B4 cells infected with the same virus recombinant were used a positive control for this experiment, and mock-infected M2-10B4 and iDC were used as negative controls. As shown in Fig. Fig.3C,3C, immunoblot analysis of infected cell lysates with virus antigen-specific antibodies revealed that the IE1, E1, and gB proteins were all strongly expressed in the M2-10B4 cells, while only the immediate-early and early proteins could be detected in lysates from virally infected iDC (gB was not detected, even after prolonged exposure of the immunoblots; data not shown). It is also readily apparent that the level of IE1 protein expression in iDC was much reduced, in comparison to the permissive M2-10B4 cells. This suggests that MCMV replication in human iDC may be blocked both at the level of early gene expression and at the level of late gene expression. These findings also confirm that MCMV does indeed undergo abortive infection in human DC.
Human iDC were infected with MCMV-EGFP, and the expression of selected cell surface molecules associated with DC maturation and/or DC costimulatory activity (CD11b, CD11c, CD40, CD80, CD83, CD86, and MHC class I and II) was examined by flow cytometry at 24 h after virus infection. The flow cytometric results were then analyzed by gating the DC into two populations: EGFP-positive cells (MCMV infected) and EGFP-negative cells (bystander cells in the same culture dish). Immunophenotypic analysis of the two gated cell populations is presented in Fig. Fig.44 (note that, in this experiment, ca. 50% of the iDC were EGFP positive; data not shown).
The data presented in Fig. Fig.44 show that expression of several markers was upregulated in the MCMV-infected, EGFP-positive iDC population compared to uninfected, EGFP-negative bystander cells. Among these markers were such maturation-associated molecules as CD40, CD86, and HLA-DR (see Fig. Fig.4).4). However, neither the infected cells nor the bystander iDCs expressed CD83 (data not shown), suggesting that the generally modest effects of MCMV infection stopped well short of inducing full DC maturation. Furthermore, similar experiments performed in mDC revealed that MCMV infection had only a minimal effect on the expression of these same cell surface proteins in mDC (data not shown).
In light of the modest maturation-enhancing effect of MCMV infection on the expression of cell surface proteins by human iDC, we wondered whether MCMV infection might exert any effect on the functional activity of human DC. To address this question, we measured the ability of MCMV-EGFP-infected iDC to promote T-cell proliferation in an allogeneic MLR. The results showed a very modest increase in cellular proliferation in the MLR assay, when MCMV-infected DC were compared to DC that were mock infected (Fig. (Fig.5).5). This difference was found in all three of the experiments performed, but the extent of the effect observed was small enough to be statistically insignificant (see error bars, which denote the standard deviation of triplicate counts-per-minute [cpm] values).
To examine the effect of MCMV infection on the ability of human DC to stimulate antigen-specific T-cell proliferation, we conducted an in vitro experiment by using the MCMV-gp120 vector. Human iDC generated from a HLA A2-positive, HIV-naive donor were infected with this virus (or mock infected), and aliquots of these cells were then pulsed with a HLA A2-restricted peptide derived from gp120 (this peptide is designated V3 and corresponds to the sequence RGPGRAFVTI ). Two days later, the DC were washed, irradiated, and added to autologous PBMC at a ratio of 20:1 (PBMC to DC). After 2 weeks, some PBMC were harvested for use in tetramer staining assay, while the remaining cells were restimulated in the presence of new DC for an additional 2 weeks and then analyzed by using both tetramer staining and ELISPOT assays.
Tetrameric complexes of HLA-A2, refolded in the presence of the V3 peptide (RGPGRAFVTI), were used to perform flow cytometric analysis of the in vitro stimulated PBMC at both the 2- and 4-week time points. This analysis revealed that the frequency of V3 tetramer-reactive, CD8-positive T lymphocytes was rather modest in all experimental conditions at the 2-week time point (Fig. (Fig.6A6A).
However, after 4 weeks of in vitro cultivation, significantly higher frequencies of such cells could be detected in PBMC cultures that had been incubated in the presence of V3 peptide-pulsed iDC, MCMV-gp120-infected iDC, and iDC that were both V3 peptide pulsed and infected with MCMV-gp120. The highest frequency of tetramer-reactive CD8+ T cells was detected in cultures which had been expanded in the presence of iDC that were infected with MCMV-gp120 and also pulsed with the V3 peptide. Slightly lower levels of tetramer-positive CD8+ lymphocytes were detected in cultures that were expanded in the presence of iDC that were only infected with MCMV-gp120 or only pulsed with the V3 peptide but, in all cases, the frequency of tetramer-reactive cells greatly exceeded the level that was detected in cells that were expanded in the presence of untreated iDC (Fig. (Fig.6A6A).
Since the tetramer analysis did not reveal whether the in vitro-expanded, V3-specific T cells might possess functional activity specific for V3-epitope positive, autologous target cells, we performed IFN-γ ELISPOT analysis on the in vitro-expanded PBMC cultures. As shown in Fig. Fig.6B,6B, this analysis revealed results that were consistent with the tetramer staining data. Thus, the frequency of V3-epitope specific, IFN-γ-secreting T cells are highest in those cultures that had been expanded in the presence of V3-pulsed and MCMV-gp120-infected iDC and much lower in cultures that were expanded in the presence of iDC that received only one of these modes of antigen exposure (Fig. (Fig.6B).6B). The results suggested that there might be some synergistic action in cells stimulated with both V3-pulsed and MCMV-gp120-infected iDC.
Taken together, the results shown in Fig. Fig.66 show that infection with MCMV does not interfere with the ability of human DC to stimulate antigen-specific T-cell proliferation. Furthermore, this experiment also establishes that a MCMV vector can successfully elicit the functional presentation of an exogenous encoded antigen by human DC (in this case, HIV-1 gp120).
In the present study, we demonstrated that MCMV undergoes an abortive infection cycle in human monocyte-derived DC, which is characterized by the efficient expression of viral immediate-early genes, but only low levels of early gene products and the absence either of viral replication or of late viral proteins. We also demonstrated that MCMV can efficiently express vectored proteins such as EGFP and gp120 in human DC, when such genes are placed under the transcriptional control of the HCMV immediate-early promoter (IEP) and inserted into the ie2 gene locus of MCMV. These results are broadly consistent with findings previously reported by Manning et al., who showed that recombinant MCMV vectors can abortively infect human 293 cells, leading to expression of a HCMV-IEP-driven transgene (vesicular stomatitis virus G protein) and the production of MCMV immediate-early genes (but no other MCMV gene products) (34).
In our experiments, as in those reported by Manning et al. (34), the infection of human cells with MCMV vectors resulted in persistent transgene expression in the absence of detectable cytopathic effects. In addition, MCMV infection of immature human DC resulted in a modest upregulation in the cell surface expression levels of costimulatory molecules and/or markers of cellular maturation but no significant change in the ability of these cells to stimulate T-cell proliferation, as measured in an allogeneic MLR. These findings suggested that it might be possible to use MCMV vectors to express desired antigens in human DC in order to then elicit immune responses against the vectored antigens. We therefore conducted initial proof-of-principle experiments that were designed to test this hypothesis by using a MCMV vector that encoded HIV-1 gp120. Human DC were infected with the MCMV-gp120 vector, and their ability to support the in vitro expansion of autologous, gp120-specific T cells from a HIV-naive, HLA-A2-positive donor was examined. This experiment confirmed that MCMV-gp120-infected human DC could indeed stimulate the expansion of functionally active, gp120-specific T cells, as assessed by both flow cytometric analysis by using a gp120-specific HLA-A2 tetramer and an IFN-γ ELISPOT assay. These findings therefore suggest that MCMV vectors could have future potential for use as vaccine delivery vehicles for immunization against infectious disease agents or tumor antigens.
In light of the strict host-species specificity of CMVs (28), it is not surprising that MCMV undergoes abortive infection in human DC. Our results are also consistent with findings previously reported by van Den Pol et al. (59, 60) and Manning et al. (34). These investigators examined CMV cell tropism for human cells and showed that MCMV-based vectors could be used for successful gene transfer into 293 cells and into brain cells derived from a wide range of mammalian species, including humans. These researchers also showed that MCMV did not replicate in these cells and that MCMV replication was arrested at the transition from immediate-early to early, at least in 293 cells (34); the precise stage at which MCMV replication was arrested in human brain cells was not examined (59, 60).
In the present experiments, we did consider the possibility that the detection of MCMV antigens in human DC might be a consequence not of virus infection and subsequent gene expression but rather a reflection of phagocytic or pinocytic uptake of virus particles. However, this explanation was rejected for two major reasons. First, the fact that cycloheximide prevents GFP expression in MCMV-EGFP-exposed DC (Fig. (Fig.1A)1A) provides strong evidence that new protein synthesis (and not phagocytosis) is required for antigen detection. Second, if phagocytosis were making a dominant contribution to the detection of MCMV antigens in human DC, one would expect to see high levels of the late protein product, gB, in the DC because gB is the major constituent of the human CMV envelope and thus a major component of the CMV particle (7, 18). However, we could not detect any gB in the MCMV-exposed human DC. In contrast, we did detect high levels of the IE1 protein, which (at least in the case of human CMV) is not considered to be a virion component (38).
It is interesting that previous studies have shown that MCMV was even able to mediate gene transfer into central nervous system cells from nonmammalian species, including chickens and turtles (60). This observation strongly suggests that the virus must be capable of attaching to some highly conserved entry receptor that is expressed in many different species. The nature of this receptor remains uncertain. It is also possible that cell type-specific mechanisms of virus entry may contribute to MCMV entry into human DC. For example, human CMV has recently been shown to target DC via DC-SIGN (22). In addition, infectious and noninfectious simian immunodeficiency virus particles have been shown to enter human DC after internalization via clathrin-coated pits (16). The question of MCMV receptor usage in human DC and in cells from nonmurine species will require further investigation in future experiments.
A variety of signals, including bacterial cell wall components, immunostimulatory DNA sequences, and cytokines, can all lead to the activation and functional maturation of DCs. It is believed that this may occur, at least in part, via the activation of the transcription factor NF-κB (4, 27, 41, 47). A similar mechanism has also been invoked to explain the strong maturation-enhancing effect that recombinant adenovirus vectors exert on human DC (39). Thus, it is possible that MCMV may exert its effects on DC maturation through a similar mechanism. Alternatively, MCMV may mediate its effects in a manner more akin to canarypox virus, which has been shown to induce the maturation of human DC as a consequence of cellular phagocytosis of the viral particles (23). It is also possible that the initial cell surface binding events involved in MCMV infection and entry may play a role in the maturation-enhancing effects that we have observed. This would be consistent with recent findings that have shown that HCMV gB can mediate the induction of a number of cellular genes, including IFN response genes (53). Future experiments will be required in order to resolve these various possibilities.
The ability of MCMV-gp120-infected DC to stimulate the generation of an antigen-specific T-cell response was examined by using an in vitro system. This experiment revealed that MCMV-gp120-transduced DC were able to elicit the expansion of gp120-specific CD8+ T lymphocytes from the PBMC of a naive subject. As expected (8), the efficiency of this T-cell expansion exceeded that observed in PBMC cultures that were incubated in the presence of peptide-pulsed DC. Furthermore, the viral vector system exhibited an even more potent effect in terms of T-cell expansion, when it was used in combination with peptide pulsing. The mechanistic basis for this positive interaction is not directly addressed here, but we speculate that it may reflect the virally induced upregulation of costimulatory molecules on DC, including CD40, which has previously been shown to be central to the expansion of antigen-specific CD8+ T lymphocytes from naive CD8 T cells (26).
In summary, our results suggest that MCMV may have utility as a vaccine delivery vehicle in humans. Important advantages of this vector system would include the efficiency of gene transfer into human DC, as well as the maturation and function-enhancing consequences of MCMV infection of iDC. Other potential advantages include the strong safety profile of MCMV (MCMV is rated as a biosafety level one agent, in contrast to most other viral vector systems that are being developed for immunization purposes) and the abortive nature of its interaction with human cells. We have shown here that MCMV early proteins are expressed only very weakly in human cells and that viral late proteins are not expressed at all. This means that potentially immunomodulatory viral gene products, or proteins that may interfere with DC function, are unlikely to be expressed in human DC (2). Furthermore, even in the unlikely event that some immunomodulatory genes were to be expressed in human DC (e.g., as IE genes), they would not be expected to exert any effect because they are quite likely optimized for interfering with murine immune molecules. This would certainly be consistent with the data presented here. Finally, the abortive nature of the interaction of MCMV with human cells suggests that it may be instructive to consider MCMV-based immunization vectors as being in some senses analogous to certain poxvirus vectors, including fowlpox virus and MVA, which also undergo abortive infection in human cells (6, 11, 13, 54, 57, 58). It may therefore be of interest to conduct future studies in nonhuman primates, with a view to determining whether MCMV-based vectors can be successfully used to elicit cellular and humoral immune responses to vectored antigens.
We thank Fernando Ontiveros, as well as Peter Keng and Alexandra Livingstone, for advice and assistance with the flow cytometry analysis; Andrea Reus for technical help with the construction of the viral genomes; and Karen L. Bentley at the Electron Microscopy Research Core for help with electron microscopy. We are also grateful to John Shanley and Stipan Jonjic for their generous gifts of MCMV-specific antibodies and to the NIH/NIAID Tetramer Facility for providing the HLA A2-V3 tetramer that was used in the present study. We also thank the volunteers who donated their blood for use in the present study.
This work was supported by NIH grants to S.D. and M.M. (R21 AI46259), to Tom Evans and S.D. (R21 AI49804), and to K.S. (F31 54 330).