Latent CMV infection of Lymphatic endothelial cells is sufficient to drive CD8 T cell memory inflation.

CMV, a ubiquitous herpesvirus, elicits an extraordinarily large T cell response that is sustained or increases over time, a phenomenon termed 'memory inflation.' Remarkably, even latent, non-productive infection can drive memory inflation. Despite intense research on this phenomenon, the infected cell type(s) involved are unknown. To identify the responsible cell type(s), we designed a Cre-lox murine CMV (MCMV) system, where a spread-deficient (ΔgL) virus expresses recombinant SIINFEKL only in Cre+ host cells. We found that latent infection of endothelial cells (ECs), but not dendritic cells (DCs) or hepatocytes, was sufficient to drive CD8 T cell memory inflation. Infection of Lyve-1-Cre and Prox1-CreERT2 mice revealed that amongst EC subsets, infection of lymphatic ECs was sufficient. Genetic ablation of ß2m on lymphatic ECs did not prevent inflation, suggesting another unidentified cell type can also present antigen to CD8 T cells during latency. This novel system definitively shows that antigen presentation by lymphatic ECs drives robust CD8 T cell memory inflation.

Human Cytomegalovirus Infection Promotes Expansion of a Functionally Superior Cytoplasmic CD3+ NK Cell Subset with a Bcl11b-Regulated T Cell Signature.

Human CMV (HCMV) is a ubiquitous pathogen that indelibly shapes the NK cell repertoire. Using transcriptomic, epigenomic, and proteomic approaches to evaluate peripheral blood NK cells from healthy human volunteers, we find that prior HCMV infection promotes NK cells with a T cell-like gene profile, including the canonical markers CD3ε, CD5, and CD8ß, as well as the T cell lineage-commitment transcription factor Bcl11b. Although Bcl11b expression is upregulated during NK maturation from CD56bright to CD56dim, we find a Bcl11b-mediated signature at the protein level for FcεRIγ, PLZF, IL-2Rß, CD3γ, CD3δ, and CD3ε in later-stage, HCMV-induced NK cells. BCL11B is targeted by Notch signaling in T cell development, and culture of NK cells with Notch ligand increases cytoplasmic CD3ε expression. The Bcl11b-mediated gain of CD3ε, physically associated with CD16 signaling molecules Lck and CD247 in NK cells is correlated with increased Ab-dependent effector function, including against HCMV-infected cells, identifying a potential mechanism for their prevalence in HCMV-infected individuals and their prospective clinical use in Ab-based therapies.

PD-1-specific "Blocking" antibodies that deplete PD-1+ T cells present an inconvenient variable in preclinical immunotherapy experiments.

Therapeutic antibodies blocking PD-1-/PD-L1 interaction have achieved remarkable clinical success in cancer. In addition to blocking a target molecule, some isotypes of antibodies can activate complement, NK cells or phagocytes, resulting in death of the cell expressing the antibody's target. Human anti-PD-1 therapeutics use antibody isotypes designed to minimize such antibody-dependent lysis. In contrast, anti-PD-1 reagents used in mice are derived from multiple species, with different isotypes, and are not engineered to reduce target cell death: few studies analyze or discuss how antibody species and isotype may impact data interpretation. We demonstrate here that anti-PD-1 therapy to promote activation and proliferation of murine PD-1-expressing CD8 T cells sometimes led instead to a loss of antigen specific cells. This phenomenon was seen in two tumor models and a model of virus infection, and varied with the clone of anti-PD-1 antibody. Additionally, we compared competition among anti-PD-1 clones to find a combination that allows detection of PD-1-expressing cells despite the presence of blocking anti-PD1 antibodies in vivo. These data bring attention to the possibility of unintended target cell depletion with some commonly used anti-mouse PD-1 clones, and should provide a valuable resource for the design and interpretation of anti-PD-1 studies in mice.

Vaccine-induced CD8 T cells are redirected with peptide-MHC class I-IgG antibody fusion proteins to eliminate tumor cells in vivo.

Simulating a viral infection in tumor cells is an attractive concept to eliminate tumor cells. We previously reported the molecular design and the in vitro potency of recombinant monoclonal antibodies fused to a virus-derived peptide MHC class I complex that bypass the peptide processing and MHC loading pathway and directly displays a viral peptide in an MHC class I complex on the tumor cell surface. Here, we show that a vaccination-induced single peptide-specific CD8 T cell response was sufficient to eliminate B16 melanoma tumor cells in vivo in a fully immunocompetent, syngeneic mouse tumor model when mice were treated with mouse pMHCI-IgGs fusion proteins targeting the mouse fibroblast activation protein. Tumor growth of small, established B16 lung metastases could be controlled. The pMHCI-IgG had similar potency as an analogous pan-CD3 T-cell bispecific antibody. In contrast to growth control of small tumors, none of the compounds controlled larger solid tumors of MC38 cancer cells, despite penetration of pMHCI-IgGs into the tumor tissue and clear attraction and activation of antigen-specific CD8 T cells inside the tumor. pMHCI-IgG can have a similar potency as classical pan-T-cell recruiting molecules. The results also highlight the need to better understand immune suppression in advanced solid tumors.

Stochastic Expansions Maintain the Clonal Stability of CD8+ T Cell Populations Undergoing Memory Inflation Driven by Murine Cytomegalovirus.

CMV is an obligate and persistent intracellular pathogen that continually drives the production of highly differentiated virus-specific CD8+ T cells in an Ag-dependent manner, a phenomenon known as memory inflation. Extensive proliferation is required to generate and maintain inflationary CD8+ T cell populations, which are counterintuitively short-lived and typically exposed to limited amounts of Ag during the chronic phase of infection. An apparent discrepancy therefore exists between the magnitude of expansion and the requirement for ongoing immunogenic stimulation. To address this issue, we explored the clonal dynamics of memory inflation. First, we tracked congenically marked OT-I cell populations in recipient mice infected with murine CMV (MCMV) expressing the cognate Ag OVA. Irrespective of numerical dominance, stochastic expansions were observed in each population, such that dominant and subdominant OT-I cells were maintained at stable frequencies over time. Second, we characterized endogenous CD8+ T cell populations specific for two classic inflationary epitopes, M38 and IE3. Multiple clonotypes simultaneously underwent Ag-driven proliferation during latent infection with MCMV. In addition, the corresponding CD8+ T cell repertoires were stable over time and dominated by persistent clonotypes, many of which also occurred in more than one mouse. Collectively, these data suggest that stochastic encounters with Ag occur frequently enough to maintain oligoclonal populations of inflationary CD8+ T cells, despite intrinsic constraints on epitope display at individual sites of infection with MCMV.

Correction: Cross-presentation of a spread-defective MCMV is sufficient to prime the majority of virus-specific CD8+ T cells.

[This corrects the article DOI: 10.1371/journal.pone.0009681.].

The immune response to CMV infection and vaccination in mice, monkeys and humans: recent developments.

The immune response to CMV is characterized by extremely large T cell and antibody responses that persist for a lifetime, but do not prevent superinfection with other CMV strains. This makes generation of a vaccine against CMV very difficult, but has facilitated development of CMV-vectored vaccines, which have shown promise in mouse tumor models and in monkey models of infectious disease. The serendipitous use of a mutant rhesus CMV vector for the SIV vaccine elicited extraordinary, CD8 T cell responses restricted by MHCII and non-classical MHCI molecules which apparently provide protection against SIV. CMV-specific CD8 T cell responses in the mouse model are driven by antigen and live out their lives primarily within the intravascular compartment.

Enzymatic synthesis of core 2 O-glycans governs the tissue-trafficking potential of memory CD8+ T cells.

Trafficking of memory CD8+ T cells out of the circulation is essential to provide protective immunity against intracellular pathogens in nonlymphoid tissues. However, the molecular mechanisms that dictate the trafficking potential of diverse memory CD8+ T cell populations are not completely defined. We show that after infection or inflammatory challenge, central memory (TCM) CD8+ T cells rapidly traffic into nonlymphoid tissues, whereas most effector memory cells remain in the circulation. Furthermore, we demonstrate that cellular migration of memory CD8+ T cells into nonlymphoid tissues is driven by interleukin-15 (IL-15)-stimulated enzymatic synthesis of core 2 O-glycans, which generates functional ligands for E- and P-selectins. Given that IL-15-stimulated expression of glycosyltransferase enzymes is largely a feature of TCM CD8+ T cells, this allows TCM to selectively migrate out of the circulation and into nonlymphoid tissues. Collectively, our data indicate that entry of memory CD8+ T cells into inflamed, nonlymphoid tissues is primarily restricted to TCM cells that have the capacity to synthesize core 2 O-glycans.

CMV immune evasion and manipulation of the immune system with aging.

Human cytomegalovirus (HCMV) encodes numerous proteins and microRNAs that function to evade the immune response and allow the virus to replicate and disseminate in the face of a competent innate and acquired immune system. The establishment of a latent infection by CMV, which if completely quiescent at the level of viral gene expression would represent an ultimate in immune evasion strategies, is not sufficient for lifelong persistence and dissemination of the virus. CMV needs to reactivate and replicate in a lytic cycle of infection in order to disseminate further, which occurs in the face of a fully primed secondary immune response. Without reactivation, latency itself would be redundant for the virus. It is also becoming clear that latency is not a totally quiescent state, but is characterized by limited viral gene expression. Therefore, the virus also needs immune evasion strategies during latency. An effective immune response to CMV is required or viral replication will cause morbidity and ultimately mortality in the host. There is clearly a complex balance between virus immune evasion and host immune recognition over a lifetime. This poses the important question of whether long-term evasion or manipulation of the immune response driven by CMV is detrimental to health. In this meeting report, three groups used the murine model of CMV (MCMV) to examine if the contribution of the virus to immune senescence is set by the (i) initial viral inoculum, (ii) inflation of T cell responses, (iii) or the balance between functionally distinct effector CD4+ T cells. The work of other groups studying the CMV response in humans is discussed. Their work asks whether the ability to make immune responses to new antigens is compromised by (i) age and HCMV carriage, (ii) long-term exposure to HCMV giving rise to an overall immunosuppressive environment and increased levels of latent virus, or (iii) adapted virus mutants (used as potential vaccines) that have the capacity to elicit conventional and unconventional T cell responses.

Fibroblast-adapted human CMV vaccines elicit predominantly conventional CD8 T cell responses in humans.

Cytomegalovirus (CMV)-based vaccines have shown remarkable efficacy in the rhesus macaque model of acquired immune deficiency syndrome, enabling 50% of vaccinated monkeys to clear a subsequent virulent simian immunodeficiency virus challenge. The protective vaccine elicited unconventional CD8 T cell responses that were entirely restricted by MHC II or the nonclassical MHC I molecule, MHC-E. These unconventional responses were only elicited by a fibroblast-adapted rhesus CMV vector with limited tissue tropism; a repaired vector with normal tropism elicited conventional responses. Testing whether these unusual protective CD8 T responses could be elicited in humans requires vaccinating human subjects with a fibroblast-adapted mutant of human CMV (HCMV). In this study, we describe the CD8 T cell responses of human subjects vaccinated with two fibroblast-adapted HCMV vaccines. Most responses were identified as conventional classically MHC I restricted, and we found no evidence for MHC II or HLA-E restriction. These results indicate that fibroblast adaptation alone is unlikely to explain the unconventional responses observed in macaques.

Blocking Virus Replication during Acute Murine Cytomegalovirus Infection Paradoxically Prolongs Antigen Presentation and Increases the CD8+ T Cell Response by Preventing Type I IFN-Dependent Depletion of Dendritic Cells.

Increasing amounts of pathogen replication usually lead to a proportionate increase in size and effector differentiation of the CD8+ T cell response, which is attributed to increased Ag and inflammation. Using a murine CMV that is highly sensitive to the antiviral drug famciclovir to modulate virus replication, we found that increased virus replication drove increased effector CD8+ T cell differentiation, as expected. Paradoxically, however, increased virus replication dramatically decreased the size of the CD8+ T cell response to two immunodominant epitopes. The decreased response was due to type I IFN-dependent depletion of conventional dendritic cells and could be reproduced by specific depletion of dendritic cells from day 2 postinfection or by sterile induction of type I IFN. Increased virus replication and type I IFN specifically inhibited the response to two immunodominant epitopes that are known to be dependent on Ag cross-presented by DCs, but they did not inhibit the response to "inflationary" epitopes whose responses can be sustained by infected nonhematopoietic cells. Our results show that type I IFN can suppress CD8+ T cell responses to cross-presented Ag by depleting cross-presenting conventional dendritic cells.

A Phase 1 Study of 4 Live, Recombinant Human Cytomegalovirus Towne/Toledo Chimera Vaccines in Cytomegalovirus-Seronegative Men.

BACKGROUND: Human cytomegalovirus (HCMV) infection causes disease in newborns and transplant recipients. A HCMV vaccine (Towne) protects transplant recipients. METHODS: The genomes of Towne and the nonattenuated Toledo strain were recombined, yielding 4 Towne/Toledo chimera vaccines. Each of 36 HCMV-seronegative men received 1 subcutaneous dose of 10, 100, or 1000 plaque-forming units (PFU) in cohorts of 3. Safety and immunogenicity were evaluated over 12 weeks after immunization and for 52 weeks for those who seroconverted. RESULTS: There were no serious local or systemic reactions. No subject had HCMV in urine or saliva. For chimera 3, none of 9 subjects seroconverted. For chimera 1, 1 of 9 seroconverted (the seroconverter received 100 PFU). For chimera 2, 3 subjects seroconverted (1 received 100 PFU, and 2 received 1000 PFU). For chimera 4, 7 subjects seroconverted (1 received 10 PFU, 3 received 100 PFU, and 3 received 1000 PFU). All 11 seroconverters developed low but detectable levels of neutralizing activity. CD4+ T-cell responses were detectable in 1 subject (who received 100 PFU of chimera 4). Seven subjects receiving chimera 2 or 4 had detectable CD8+ T-cell responses to IE1; 3 responded to 1-2 additional antigens. CONCLUSIONS: The Towne/Toledo chimera vaccine candidates were well tolerated and were not excreted. Additional human trials of chimeras 2 and 4 are appropriate. CLINICAL TRIALS REGISTRATION: NCT01195571.

Intratumoral Infection with Murine Cytomegalovirus Synergizes with PD-L1 Blockade to Clear Melanoma Lesions and Induce Long-term Immunity.

Cytomegalovirus is an attractive cancer vaccine platform because it induces strong, functional CD8(+) T-cell responses that accumulate over time and migrate into most tissues. To explore this, we used murine cytomegalovirus expressing a modified gp100 melanoma antigen. Therapeutic vaccination by the intraperitoneal and intradermal routes induced tumor infiltrating gp100-specific CD8(+) T-cells, but provided minimal benefit for subcutaneous lesions. In contrast, intratumoral infection of established tumor nodules greatly inhibited tumor growth and improved overall survival in a CD8(+) T-cell-dependent manner, even in mice previously infected with murine cytomegalovirus. Although murine cytomegalovirus could infect and kill B16F0s in vitro, infection was restricted to tumor-associated macrophages in vivo. Surprisingly, the presence of a tumor antigen in the virus only slightly increased the efficacy of intratumoral infection and tumor-specific CD8(+) T-cells in the tumor remained dysfunctional. Importantly, combining intratumoral murine cytomegalovirus infection with anti-PD-L1 therapy was synergistic, resulting in tumor clearance from over half of the mice and subsequent protection against tumor challenge. Thus, while a murine cytomegalovirus-based vaccine was poorly effective against established subcutaneous tumors, direct infection of tumor nodules unexpectedly delayed tumor growth and synergized with immune checkpoint blockade to promote tumor clearance and long-term protection.

Strength in Numbers: Visualizing CTL-Mediated Killing In Vivo.

Cytotoxic CD8+ T lymphocytes (CTLs) have long been believed to be extremely efficient killers. Forster and colleagues (Halle et al., 2016) used in vivo imaging to tell a different story, in which each CTL killed only 2-16 targets a day, and several CTLs per target were needed to get the job done.

Cytokine-Mediated Activation of NK Cells during Viral Infection.

UNLABELLED: Natural killer (NK) cells provide a first line of defense against infection via the production of antiviral cytokines and direct lysis of target cells. Cytokines such as interleukin 12 (IL-12) and IL-18 are critical regulators of NK cell activation, but much remains to be learned about how cytokines interact to regulate NK cell function. Here, we have examined cytokine-mediated activation of NK cells during infection with two natural mouse pathogens, lymphocytic choriomeningitis virus (LCMV) and murine cytomegalovirus (MCMV). Using a systematic screen of 1,849 cytokine pairs, we identified the most potent combinations capable of eliciting gamma interferon (IFN-γ) production in NK cells. We observed that NK cell responses to cytokine stimulation were reduced 8 days after acute LCMV infection but recovered to preinfection levels by 60 days postinfection. In contrast, during MCMV infection, NK cell responses to cytokines remained robust at all time points examined. Ly49H-positive (Ly49H+) NK cells recognizing viral ligand m157 showed preferential proliferation during early MCMV infection. A population of these cells was still detected beyond 60 days postinfection, but these divided cells did not demonstrate enhanced IFN-γ production in response to innate cytokine stimulation. Instead, the maturation state of the NK cells (as determined by CD11b or CD27 surface phenotype) was predictive of responsiveness to cytokines, regardless of Ly49H expression. These results help define cytokine interactions that regulate NK cell activation and highlight variations in NK cell function during two unrelated viral infections. IMPORTANCE: Natural killer cells play an important role in immunity to many viral infections. From an initial screen of 1,849 cytokine pairs, we identified the most stimulatory cytokine combinations capable of inducing IFN-γ production by NK cells. Ly49H+ NK cells, which can be directly activated by MCMV protein m157, preferentially proliferated during MCMV infection but did not show enhanced IFN-γ production following direct ex vivo cytokine stimulation. Instead, mature CD11b+ and/or CD27+ NK cells responded similarly to innate cytokine stimulation regardless of Ly49H expression. Collectively, our data provide a better foundation for understanding cytokine-mediated NK cell activation during viral infection.

Cytomegalovirus-based cancer vaccines expressing TRP2 induce rejection of melanoma in mice.

Cytomegalovirus (CMV) induces strong and long-lasting immune responses, which make it an attractive candidate for a cancer vaccine vector. In this study, we tested whether a tumor antigen expressed in CMV can induce a strong anti-tumor effect. We expressed an unmodified melanoma antigen, mouse tyrosinase-related protein 2 (TRP2), in mouse cytomegalovirus (MCMV). Prophylactic vaccination of the mice with a single dose of MCMV-TRP2 induced rejection of B16 melanoma challenge; therapeutic vaccination with MCMV-TRP2 prolonged the survival of the mice challenged with B16 cells. Additionally, vaccination with MCMV-TRP2 five months before tumor challenge still induced tumor rejection, which indicated that the vaccine induced long-term protection. Furthermore, MCMV-TRP2 protected mice against B16 melanoma challenge regardless of the pre-existing CMV infection. We found that vaccination with MCMV-TRP2 induced long-lasting TRP2 specific antibodies but not CD8 T cells. In addition, depletion of CD4 and CD8 T cells did not compromise the antitumor effect by MCMV-TRP2; while in B cell deficient (µMT) mice, the vaccine lost its antitumor effect. These results indicate that antibodies, not T cells, are important in mediating the antitumor effect during the effector phase by the vaccine. We also made a spread deficient MCMV-TRP2 lacking the essential glycoprotein gL, which showed a similar antitumor effect. In conclusion, our study indicates that tumor antigen (TRP2) expressed in MCMV induces a strong and long-lasting anti-melanoma effect through an antibody-dependent mechanism. Our findings demonstrate that CMV might be a promising vector for the development of cancer vaccines.

Competition for antigen at the level of the APC is a major determinant of immunodominance during memory inflation in murine cytomegalovirus infection.

The unique ability of CMV to drive the expansion of virus-specific T cell populations during the course of a lifelong, persistent infection has generated interest in the virus as a potential vaccine strategy. When designing CMV-based vaccine vectors to direct immune responses against HIV or tumor Ags, it becomes important to understand how and why certain CMV-specific populations are chosen to inflate over time. To investigate this, we designed recombinant murine CMVs (MCMVs) encoding a SIINFEKL-enhanced GFP fusion protein under the control of endogenous immediate early promoters. When mice were infected with these viruses, T cells specific for the SIINFEKL epitope inflated and profoundly dominated T cells specific for nonrecombinant (i.e., MCMV-derived) Ags. Moreover, when the virus encoded SIINFEKL, T cells specific for nonrecombinant Ags displayed a phenotype indicative of less frequent exposure to Ag. The immunodominance of SIINFEKL-specific T cells could not be altered by decreasing the number of SIINFEKL-specific cells available to respond, or by increasing the number of cells specific for endogenous MCMV Ags. In contrast, coinfection with viruses expressing and lacking SIINFEKL enabled coinflation of T cells specific for both SIINFEKL and nonrecombinant Ags. Because coinfection allows presentation of SIINFEKL and MCMV-derived Ags by different cells within the same animal, these data reveal that competition for, or availability of, Ag at the level of the APC determines the composition of the inflationary response to MCMV. SIINFEKL's strong affinity for H-2K(b), as well as its early and abundant expression, may provide this epitope's competitive advantage.

Sustained CD8+ T cell memory inflation after infection with a single-cycle cytomegalovirus.

Cytomegalovirus (CMV) is a ß-herpesvirus that establishes a lifelong latent or persistent infection. A hallmark of chronic CMV infection is the lifelong persistence of large numbers of virus-specific CD8+ effector/effector memory T cells, a phenomenon called "memory inflation". How the virus continuously stimulates these T cells without being eradicated remains an enigma. The prevailing view is that CMV establishes a low grade "smoldering" infection characterized by tiny bursts of productive infection which are rapidly extinguished, leaving no detectable virus but replenishing the latent pool and leaving the immune system in a highly charged state. However, since abortive reactivation with limited viral gene expression is known to occur commonly, we investigated the necessity for virus reproduction in maintaining the inflationary T cell pool. We inhibited viral replication or spread in vivo using two different mutants of murine CMV (MCMV). First, famcyclovir blocked the replication of MCMV encoding the HSV Thymidine Kinase gene, but had no impact on the CD8+ T cell memory inflation once the infection was established. Second, MCMV that lacks the essential glycoprotein L, and thus is completely unable to spread from cell to cell, also drove memory inflation if the virus was administered systemically. Our data suggest that CMV which cannot spread from the cells it initially infects can repeatedly generate viral antigens to drive memory inflation without suffering eradication of the latent genome pool.

Comparing the kinetics of NK cells, CD4, and CD8 T cells in murine cytomegalovirus infection.

NK cells recognize virus-infected cells with germline-encoded activating and inhibitory receptors that do not undergo genetic recombination or mutation. Accordingly, NK cells are often considered part of the innate immune response. The innate response comprises rapid early defenders that do not form immune memory. However, there is increasing evidence that experienced NK cells provide increased protection to secondary infection, a hallmark of the adaptive response. In this study, we compare the dynamics of the innate and adaptive immune responses by examining the kinetic profiles of the NK and T cell response to murine CMV infection. We find that, unexpectedly, the kinetics of NK cell proliferation is neither earlier nor faster than the CD4 or CD8 T cell response. Furthermore, early NK cell contraction after the peak of the response is slower than that of T cells. Finally, unlike T cells, experienced NK cells do not experience biphasic decay after the response peak, a trait associated with memory formation. Rather, NK cell contraction is continuous, constant, and returns to below endogenous preinfection levels. This indicates that the reason why Ag-experienced NK cells remain detectable for a prolonged period after adoptive transfer and infection is in part due to the high precursor frequency, slow decay rate, and low background levels of Ly49H(+) NK cells in recipient DAP12-deficient mice. Thus, the quantitative contribution of Ag-experienced NK cells in an endogenous secondary response, with higher background levels of Ly49H(+) NK cells, may be not be as robust as the secondary response observed in T cells.

Cross-presentation of a spread-defective MCMV is sufficient to prime the majority of virus-specific CD8+ T cells.

CD8+ T cells can be primed by peptides derived from endogenous proteins (direct presentation), or exogenously acquired protein (cross-presentation). However, the relative ability of these two pathways to prime CD8+ T cells during a viral infection remains controversial. Cytomegaloviruses (CMVs) can infect professional antigen presenting cells (APCs), including dendritic cells, thus providing peptides for direct presentation. However, the viral immune evasion genes profoundly impair recognition of infected cells by CD8+ T cells. Nevertheless, CMV infection elicits a very strong CD8+ T cell response, prompting its recent use as a vaccine vector. We have shown previously that deleting the immune evasion genes from murine cytomegalovirus (MCMV) that target class I MHC presentation, has no impact on the size or breadth of the CD8+ T cell response elicited by infection, suggesting that the majority of MCMV-specific CD8+ T cells in vivo are not directly primed by infected professional APCs. Here we use a novel spread-defective mutant of MCMV, lacking the essential glycoprotein gL, to show that cross-presentation alone can account for the majority of MCMV-specific CD8+ T cell responses to the virus. Our data support the conclusion that cross-presentation is the primary mode of antigen presentation by which CD8+ T cells are primed during MCMV infection.