PI3K-Dependent Upregulation of Mcl-1 by Human Cytomegalovirus Is Mediated by Epidermal Growth Factor Receptor and Inhibits Apoptosis in Short-Lived Monocytes

Human CMV (HCMV), a member of the Herpesviridae family, is endemic throughout the world with 50–90% of the adult population infected (1). In immunocompetent individuals, HCMV infection is generally asymptomatic, although infection can lead to monocucleosis (2) and is associated with several chronic inflammatory diseases such as atherosclerosis and inflammatory bowel disease (3, 4). In contrast, infection of immunocompromised hosts, including neonates, AIDS patients, and transplant recipients, causes significant morbidity and mortality (5–7).

HCMV pathogenesis is a direct result of systemic viral spread to and infection of multiple organ sites that occurs during either asymptomatic or symptomatic infections (8–10). Viral dissemination to different organ systems occurs via a hematogenous route because a cell-associated viremia is a prerequisite for viral spread (11, 12). Monocytes are principal in vivo targets of HCMV (12) and are the most abundant infiltrating cell type found in infected organs during primary infection (13, 14), suggesting these blood-borne immunological cells are responsible for mediating hematogenous dissemination of the virus. We have previously provided evidence that after infection with HCMV, monocytes acquire a distinctive M1 proinflammatory phenotype (15), which we propose is necessary to mediate viral spread. Our data showed unique HCMV-induced functional changes, including increased production of M1 proinflammatory chemokines alongside an increase in select M2 anti-inflammatory chemokines and enhanced motility relative to monocytes activated by alternative agents such as LPS and PMA (10, 15, 16). Moreover, HCMV infection of monocytes induced their differentiation into macrophages, which, to our knowledge, is the only identified pathogen that can directly induce the monocyte-to-macrophage differentiation process (16, 17). For the virus, monocyte-to-macrophage differentiation appears to be an essential step in the viral survival strategy, because, in contrast to monocytes, which are nonpermissive for viral replication, macrophages can support viral replication and the production of viral particles.

Although the biological changes in HCMV-infected monocytes discussed previously provide the virus with the necessary tools to mediate spread to multiple host organ systems and to establish life-long viral persistence, infection of monocytes is not without significant biological hurdles. HCMV must subvert the intrinsic biological programming of monocytes to undergo rapid cell death within 1–3 d of entering the circulation (18). In addition to counteracting the short lifespan of monocytes, HCMV must also neutralize the cellular antiviral proapoptotic cascades. In replication permissive model cell lines, such as fibroblasts, control of the apoptotic cascade is mediated by immediate-early (IE) viral proteins, which are induced within a few hours postinfection (19). In monocytes however, because viral gene expression/replication is not observed until 3–4 wk postinfection (16), HCMV must also have evolved a strategy to inhibit the proapoptotic cellular signaling pathways in the absence of de novo viral gene expression and replication. Deciphering the prosurvival mechanism(s) used by HCMV during primary infection of nonreplication permissive monocytes, to bridge the gap between short-lived monocytes and long-lived macrophages, is critical to the understanding of viral dissemination and persistence within the infected host.

The rapid HCMV-induced functional changes that take place in monocytes during the initial nonreplication permissive stages of infection occur in a temporal manner consistent with a receptor-ligand–mediated process. We have previously shown that challenge with UV-inactivated HCMV or purified glycoprotein B induced rapid functional changes in monocytes similar to that seen with replication competent virus (16, 17). Consistent with a direct role of viral binding in cellular activation, we recently showed that activation of the cellular epidermal growth factor receptor (EGFR) during viral binding (20) and the subsequent activation of the downstream PI3K/Akt signaling cascade was required for the viral induction of monocyte motility, adhesion to endothelial cells and tranendothelial migration (21, 22). Transcriptome analyses of 4 h infected monocytes also revealed the PI3K-dependent upregulation of several transcripts encoding antiapoptotic proteins (23). This regulation of multiple antiapoptotic proteins by PI3K activity also suggests a direct role for this pathway in the early survival of monocytes.

Constitutive activation of the PI3K/Akt-1 pathway is known to promote the survival of cells differentiating along the monocyte/macrophage lineage through the expression of the downstream target Mcl-1 (24), a potent antiapoptotic protein of the Bcl-2 family (25). During the early phase of myelopoiesis, Mcl-1 plays an obligate role in ensuring the survival of myeloid progenitors (26). Similar to myeloid progenitors, monocytes initially express high levels of Mcl-1, which rapidly diminishes during the short lifespan of these cells (25). In vivo studies have demonstrated that monocytes have a lifespan of 1–3 d on entering the circulation (18); thus, we hypothesize that the declining Mcl-1 levels serve as a biological clock to ensure a controlled population of these proinflammatory immune cells. In contrast, Bcl-2 appears to be directly involved in both the differentiation of monocytes into and the long-term survival of macrophages on initiation of the differentiation programming in short-lived monocytes (27). These data indicate the existence of two distinct viability strategies occurring prior to and after 48–72 h of monocytes entering the peripheral blood and that Bcl-2 family members may act as temporal viability gates along the myelopoiesis differentiation continuum. In vitro, adherent-monocytes exhibit low-level activation and differentiation leading to survival through the early cell fate decision checkpoints (16, 28). However, cultured monocytes differentiating along the M1 continuum are sensitive to IL-10–induced apoptosis, via the downregulation of Bcl-2 family members, during the first 48 h of the differentiation process, but become resistant to the proapoptotic effects of IL-10 after 48 h (29), thus indicating the presence of a 48–72 h viability gate in vitro. Because HCMV-infected monocytes successfully navigate the 48–72 h viability gate in the absence of viral antiapoptotic proteins, we focused our study on the early stages of monocyte survival after infection and the induced cellular mechanism(s) responsible for this survival.

In this study, we examined the mechanism used by HCMV to ensure the short-term survival of nonreplication permissive monocytes, until the subsequent HCMV-induced macrophage differentiation reprogramming can occur. We demonstrate in this study that HCMV infection inhibited cell death of monocytes independent of antiapoptotic viral IE protein expression and that the induction of PI3K activity after infection is critical to the rapid conversion of the infected monocyte into a prosurvival state. Under normal regulatory conditions, the homeostatic levels of Mcl-1 rapidly declined; however, HCMV infection was able to slow the loss of Mcl-1, and thus decelerate the internal viability clock of short-lived monocytes. More specifically, we identified that the EGFR/PI3K signaling cascade initiates the activation of the early HCMV-infected monocyte signalosome required for the upregulation of Mcl-1 and the acquisition of apoptotic resistance during infection. Taken together, our data indicate that HCMV binding to cellular receptors during viral entry plays a pivotal role in the pathogenic induction of host antiapoptotic pathways, thus circumventing apoptosis in naturally short-lived monocytes and promoting early events in the viral dissemination and persistence strategy.

HCMV (Towne/E strain; passages 35–45) was cultured in human embryonic lung (HEL) fibroblasts (16, 17). Virus was purified on a 0.5 M sucrose cushion, resuspended in RPMI 1640 media (Cellgro, Mediatech, Herndon, VA), and used to infect monocytes at a multiplicity of infection (MOI) of 5 for each experiment. We have previously shown that infection with a MOI 5 results in 100% of monocytes being infected with HCMV at 4 h postinfection (hpi) (16). Monocytes were mock infected using equivalent volumes of RPMI 1640 media alone.

Isolation of human peripheral blood monocytes was performed as previously described (16, 17, 21, 30). Briefly, blood was drawn by venipuncture and centrifuged through a Ficoll Histopaque 1077 gradient (Sigma-Aldrich, St. Louis, MO) at 200 × g for 30 min at room temperature (RT). Mononuclear cells were collected and washed twice with PBS and 1 mM EDTA to remove platelets at 150 × g for 10 min at RT. Monocytes were then layered on top of a 45% and 52.5% iso-osmotic Percoll gradient and centrifuged for 30 min at 400 × g at RT yielding a population >90% monocyte. Cells were washed twice with saline at 150 × g for 10 min at RT to remove residual Percoll and suspended in RPMI 1640 (Cellgro) supplemented with 10% human serum (Sigma-Aldrich). University Institutional Review Board and Health Insurance Portability and Accountability Act guidelines were followed for all experimental protocols.

Isolated monocytes (30,000 cells per experimental arm) were plated onto fibronectin-coated microtiter plates for 1 h at 37°C. After incubation, cells were mock infected or HCMV infected for varying times, after which monocytes were washed with warm media and varying concentrations of LY294002 added for an additional 24 h at 37°C. MTT cleavage was performed as directed by the manufacturer’s instructions (Zymed/Intermedico, Markham, CA) to assess cell viability.

To determine the frequency of apoptosis, TUNEL was performed as previously described (31). Briefly, after acetone:methanol fixation and PBS washing, the fraction of nuclei with nicked DNA was determined by TUNEL reaction. After the reaction was terminated by adding 300 mM sodium chloride plus 30 mM sodium citrate, the cells were washed with distilled water, and endogenous peroxidase activity was neutralized by a 30-min incubation at RT with 3% H₂O₂. Nonspecific binding was blocked with 10% nonimmune goat serum at RT (Zymed/Intermedico) and a secondary biotinylated goat anti-mouse Ab and streptavidin-peroxidase conjugate (Streptavidin Biotin System, Histostain-SP Kit; Zymed/Intermedico) added according to the manufacturer's instructions. Nickel-diaminobenzidine substrate was added to culture wells to yield a dark brown precipitate localized to apoptotic nuclei. Monocytes were counterstained with hematoxylin (Sigma-Aldrich) to visualize total nuclei.

Monocytes were harvested in RIPA buffer (50 mM Tris-HCl [pH 7.5], 5 mM EDTA, 100 mM NaCl, 1% Triton X-100, 0.1% NaDodSO₄ [SDS], and 10% glycerol) containing one times protease inhibitor mixture (Sigma-Aldrich), one times phosphatase inhibitor mixture I (Sigma-Aldrich), and one times phosphatase inhibitor mixture II (Sigma-Aldrich) for 30 min on ice. The lysates were cleared by centrifugation at 4°C (2 min, 16,000 × g) and stored at −20°C until analyzed. Sample protein concentrations were determined in duplicate using the DC protein assay (Bio-Rad, Hercules, CA). Sample protein (15–20 μg) was solubilized in six times sample buffer (Sigma-Aldrich) by boiling for 5 min and stored until electrophoresis. Equal amounts of total cellular protein from each sample were separated using SDS-PAGE, followed by immunoblotting. Blots were blocked in a 5% milk–TBS–Tween 20 solution, followed by incubation overnight at 4°C with an anti–Bcl-2 Ab (Santa Cruz Biotechnology, Santa Cruz, CA), an anti–Mcl-1 Ab (Santa Cruz), an anticaspase 3 Ab (Calbiochem, San Diego, CA), an antiprocaspase 9 Ab (Calbiochem), an anticleaved caspase 9 Ab (Calbiochem), or an anti–β-actin Ab. Blots were then incubated with diluted HRP-conjugated secondary Abs (Amersham Biosciences, Piscataway, NJ) for 1 h at RT, washed extensively and developed using ECL Plus (Amersham Biosciences) according to the manufacturer’s protocol.

Ontology reanalysis of affymetrix gene data obtained from our previous studies were used to perform a global transcriptional evaluation of the HCMV-infected monocyte transcriptome at 4 (15), 24 (20), and 48 hpi (21). Briefly, the following criteria were used for compilation of the data from all experiments. ANOVA tests were performed on the HCMV-infected versus mock-infected replicates and p values were calculated for genes upregulated or downregulated. A p value of ≤0.05 was used as the criteria for statistically significant genes among replicates. These criteria allowed us to generate a pool of genes that were statistically significant in all of the replicates after infection. An average fold-change of ≥1.5 HCMV-infected versus mock-infected samples were considered to be significantly regulated by infection. For a detailed description of the experimental procedures, see the 1Materials and Methods from our previous studies (15, 20, 21). The GEO accession numbers for these data are GSE11408 (15), GSE17948 (20), and GSE19772 (21).

Monocytes (3 × 10⁶ cells/experimental arm) were resuspended in 100 μl Amaxa nucleofection solution for monocytes (Human Monocyte Nucleofector Kit; Amaxa Biosystems, Cologne, Germany) containing 300 μg Mcl-1 small interfering RNA (siRNA; Invitrogen) or control siRNA (Invitrogen) and transfected in an Amaxa nucleofector electroporator in accordance with the manufacturer’s instructions. The cells were immediately mixed with 500 μl prewarmed human monocyte nucleofector medium, transferred into 1 ml medium, and incubated at 37°C for 24 h. After incubation, TUNEL and Western blot analysis were performed as described previously.

PI3K activity is essential to the survival of cells differentiating in the monocyte/macrophage lineage (24, 32); however, the extent to which PI3K activity functions to enhance monocyte survival prior to and after the 48–72 h viability gate is unclear. To determine whether PI3K activity plays a prominent role in short-term and/or long-term viability, adherent monocytes, which exhibit low-level activation and differentiation (16, 28), were treated with LY294002 (PI3K inhibitor) before and after the 48–72 h cell fate decision checkpoint. Monocytes treated with 25 μM LY294002 at time points prior to 72 h exhibited similar hypersensitivity kinetics to the cytotoxic effects of the PI3K inhibitor (Fig. 1). However, although still sensitive to the cell death inducing properties of LY294002, monocytes treated at 72 h postisolation displayed significantly higher rates of cell survival than those treated with LY294002 at earlier time points. These data demonstrate the critical role of PI3K activity in the short-term survival of monocytes prior to the 48–72 h viability gate.

We have previously shown that monocyte PI3K activity is stimulated within 10 min after infection with HCMV (20) and now show that the increased kinase activity is maintained through 96 hpi (Fig. 2A). This rapid and chronic induction of PI3K activity suggested to us that PI3K activity may an important role in the protection of HCMV-infected monocytes from cell death prior the 48–72 h viability gate. To examine this possibility, cell viability analysis of mock-infected and HCMV-infected monocytes treated with varying concentrations of LY294002 at 24 hpi revealed a hypersensitivity of mock-infected monocytes to the inhibition of PI3K activity (Fig. 2B). Uninfected and infected monocytes were resistant to the cell death inducing effects of LY294002 at 5 μM. In contrast, at 25 μM and 100 μM, HCMV-infected cells were protected from LY294002 treatment, whereas the viability of mock-infected cells decreased by 40% and 55% after treatment, respectively. Next, we examined the length of time that HCMV-infected monocytes exhibited a prosurvival state after infection. By 1 hpi HCMV-infected monocytes were partially resistance to LY294002 at 25 μM. A 20% decrease in cell viability was observed when compared with mock-infected cells which showed a 33% decrease (Fig. 2C). The resistance of HCMV-infected monocytes to LY294002 increased with time until complete resistant was reached at 24 hpi. Infected cells exhibited resistance through 48 hpi. However, at 72 hpi HCMV-infected monocytes did not display enhanced protection from the effects of LY294002 treatment (Fig. 2C), despite the high levels of PI3K activity at this time point (Fig. 2A).

Congruent with the MTT viability assays (Fig. 2C), examination of DNA fragmentation associated with apoptosis using TUNEL analysis found that HCMV-infected monocytes exhibited resistance to apoptosis induced by LY294002 (Fig. 3A). Partial resistance from LY294002-induced apoptosis was observed at 4 hpi and complete resistance was observed by 24 hpi, which was then lost by 72 hpi. It should also be noted that no significant decrease in cell survival of HCMV-infected/LY-treated monocytes were observed at 4 hpi (Fig. 2C), despite the induction of apoptosis (Fig. 3A). Based on our analyses, in mock-infected monocytes, a 6-fold increase in the frequency of apoptosis was required to induce a 30% decrease in cell survival 24 h after LY294002 treatment. Because HCMV infection of monocytes decreased the rate of LY294002-induced apoptosis to 2.5-fold, the threshold frequency required to detect any concurrent decrease beyond the SD in monocyte survival is unlikely to be achieved. Monocytes challenged with UV-inactivated viral particles (UV-HCMV) displayed sensitivity kinetics to LY294002 treatment similar to that observed for live HCMV-infected cells, indicating newly synthesized viral gene products were not responsible for the induction of the monocyte prosurvival state after infection (Figs. 2B, 2C, 3A). The lack of long-term protection of HCMV-infected monocytes to LY294002 treatment, despite the high levels of PI3K activity, suggest a complex signaling profile involving the simultaneous activation of multiple pathways during HCMV bind/entry is required for the induction of survival factors. Nevertheless, it appears that the rapid activation of the PI3K pathway is central to the favorable outcome during the early critical cell fate decision period that infected monocytes must navigate.

To expand our understanding of the early resistance to LY294002-induced apoptosis in monocytes after infection with HCMV, we next examined the cleavage and subsequent activation of caspase 3 and 9, which occur early in the apoptotic process, prior to DNA fragmentation. Caspase 3 and 9 are synthesized as inactive procaspase precursors of 32 kDa and 46 kDa, respectively. Caspase 3 is proteolytically cleaved into a 20-kDa and a 12-kDa subunit that together forms an intermediary protease with partial activity, which then undergoes a second cleavage event to generate the fully active 17-kDa/12-kDa protease (33). Procaspase 9 is processed into an active heterodimer consisting of a 35-kDa and a 10-kDa subunit. At 1 and 72 , no significant difference in caspase 3 and 9 cleavage was observed between mock-infected and HCMV-infected monocytes after LY294002 treatment (Fig. 3B). In contrast, when LY294002 was added at 24 hpi, there was no stimulation of the cleavage of procaspase 3 or procaspase 9 in HCMV-infected monocytes, whereas both caspases were efficiently proteolytically cleaved in mock-infected cells. In accord, on Western blots where HCMV-infected and mock-infected samples were run simultaneously to allow for a direct comparison, the basal levels of cleaved caspase 9 in HCMV-infected/LY294002-treated cells were similar to the basal levels observed in mock-infected/mock-treated monocytes (data not shown). At 24 hpi, the presence of the fully active 17-kDa activated subunit was not observed in HCMV-infected monocytes after treatment with LY294002, although the presence of the intermediary 20-kDa subunit of caspase 3 was observed. Our data indicate that PI3K is rapidly activated in monocytes after infection with HCMV and that this activated signaling pathway results in the initiation of a cellular antiapoptotic reprogramming of short-lived monocytes.

Viral IE antiapoptotic genes, including IE1–72, IE2–86, UL36 (viral inhibitor of caspase-8 activation), and UL37 (viral mitochondria-localized inhibitor of apoptosis) are some of the first gene products produced during productive HCMV infection (34). We have previously shown that viral gene expression and replication does not occur in infected monocytes until ∼3 wk postinfection (16), thus the involvement of these viral antiapoptotic proteins in the acquisition of a prosurvival state within 24 hpi in HCMV-infected monocytes is not likely. We confirmed that IE1-72, IE2-86, UL36, and UL37 were not transcribed in monocytes during the first 72 h of infection with HCMV (Fig. 4), despite internalization of the viral particle within 1 hpi (16, 20). Infection of a control HEL fibroblast cell line showed IE1-72, IE2-86, and UL36 gene expression by 24 hpi and UL37 transcription by 48 hpi. The absence of these IE gene transcripts during the initial 72 h of monocyte infection indicates that new viral gene products are unlikely to regulate the early events in HCMV-infected monocytes.

Because our cell viability assays provided evidence that a cellular antiapoptotic reprogramming of HCMV-infected monocytes occurred within the first 48 hpi, we analyzed our transcriptome databases from our previously published studies to create a temporal analysis of HCMV-infected monocytes at 4, 24, and 48 hpi (15, 20, 21). Ontology examination of cellular antiapoptotic genes revealed 35, 18, and 7 antiapoptotic genes were upregulated ≥1.5-fold at 4, 24, and 48 hpi, respectively (Table I), indicating that HCMV infection transcriptionally induces a prosurvival cellular environment in monocytes after infection. The induction of multiple antiapoptotic genes likely counteracts the proapoptotic antiviral processes activated on viral entry and the intrinsic programming of monocytes to undergo cell death within 3 d of entering the circulation (18, 35). The decline in the number of upregulated antiapoptotic genes over 48 h hints that the remaining elevated transcripts are critical to the long-term survival of infected monocytes. We identified two members of the Bcl-2 family, Mcl-1 and Bcl-2, which were upregulated at 24 and 48 hpi, respectively. These factors have been shown to be involved in the survival of macrophages (24, 27) and to regulate mitochondrial-mediated apoptosis. Because our data showed the inhibition of LY294002-induced caspase-9 cleavage, which occurs on activation of the mitochondrial stress pathway, in HCMV-infected monocytes, we focused our examination on the potential role mitochondrial membrane permeabilization regulators, Mcl-1 and Bcl-2, had in the survival of infected monocytes.

Analysis of protein expression levels confirmed the upregulation of Mcl-1 and Bcl-2 protein in HCMV-infected monocytes at 24 and 48 hpi, respectively (Fig. 5A). By 72 hpi, Mcl-1 protein expression in HCMV-infected monocytes had returned to mock levels, whereas Bcl-2 protein levels remained elevated. High levels of Mcl-1 protein expression were initially observed in newly isolated mock-infected monocytes at 1 h postisolation and the levels rapidly dissipated over 72 h. In contrast, elevated levels of Mcl-1 was maintained in infected monocytes over the same time course, demonstrating that HCMV infection extended the period in which monocytes maximally express Mcl-1. Moreover, a detailed analysis revealed that, after isolation of peripheral blood monocytes, an induction of Mcl-1 protein in both mock-infected and HCMV-infected monocytes had occurred at 2 hpi. However, Mcl-1 expression declined by 4 h in mock-infected cells, whereas a sustained high level of Mcl-1 protein remained through 24 h in HCMV-infected cells (Fig. 5B). Our data suggests that the decrease in the rate of Mcl-1 loss in HCMV-infected monocytes could be a central mechanism by which infected monocytes survive through the 48–72 h viability gate.

Viability assays indicated that the expression of the cellular factor responsible for the increase in apoptotic resistance at 24 hpi was dependent on the PI3K signaling pathway (Fig. 2A); thus, we examined whether Mcl-1 and/or Bcl-2 expression were regulated by PI3K activity. Monocytes infected for 24 h exhibited a dose-dependent decrease in Mcl-1 levels in response to LY294002 treatment, whereas Bcl-2 expression was unaffected by LY294002 treatment (Fig. 5C). Concurrent with data indicating that maximum protection from LY294002-induced apoptosis in HCMV-infected monocytes occurs at 24 and 48 hpi, Mcl-1 protein expression remained high at these time points after LY294002 treatment (Fig. 5D). In accord with the temporal apoptotic resistant state of HCMV-infected monocytes, our data suggest that the early signaling events after infection lead to the upregulation of Mcl-1 and that the chronic activation of PI3K alone is unable to sustain the increase in Mcl-1 expression. This correlation between Mcl-1 protein levels and resistance to LY294002-induced apoptosis suggest that elevated Mcl-1 expression may be the molecular link to the rapid generation of the antiapoptotic state during the early stages of HCMV infection.

To investigate if Mcl-1 is critical to the early survival of monocytes, we used siRNA specifically targeted against Mcl-1 and examined the relative rates of apoptosis. At 24 hpi, mock-infected and HCMV-infected monocytes were transfected with Mcl-1 siRNA for 24 h and examined for Mcl-1 expression. The knockdown of Mcl-1 levels was efficient as demonstrated by the ≥90% downregulation of Mcl-1 protein expression in transfected mock-infected (Fig. 6A) and HCMV-infected monocytes (Fig. 6B). Bcl-2 protein levels were unaffected by transfection of the Mcl-1 siRNA and transfection of the nonspecific control siRNA had no effect on Mcl-1 expression levels. We found a 3- and 5-fold increase in the percentage of Mcl-1 deficient mock-infected and HCMV-infected monocytes undergoing apoptosis, respectively (Fig. 6C). The elevated levels of apoptosis observed with Mcl-1 deficient HCMV-infected monocytes suggests an initiation of proapoptotic signaling pathways after infection and that Mcl-1 plays a central role in counteracting this early cellular antiviral response. Moreover, the rapid induction of apoptosis in mock-infected monocytes after knockdown with Mcl-1 siRNA provide the first documented evidence of a direct role that Mcl-1 plays in normal monocyte survival. Overall, these data imply that, despite the increased expression of a multitude of cellular antiapoptotic factors in monocytes after HCMV infection (Table I), Mcl-1 is essential for the early steps of monocyte survival.

To investigate the cellular trigger responsible for the activation of the PI3K signaling branch within the HCMV-infected monocyte signalosome and the subsequent upregulation of Mcl-1, we examined the potential role of EGFR engagement during viral binding. We have new data showing that EGFR is expressed on human peripheral blood monocytes and that activation of EGFR kinase activity after infection results in a rapid induction of PI3K activity (20). Consequently, we next examined if the PI3K-dependent upregulation of Mcl-1 in HCMV-infected monocytes required the activation of EGFR. Reanalysis of our previous transcriptome study (20), with a focus on Mcl-1 gene expression, revealed a 50% reduction in Mcl-1 expression in HCMV-infected monocytes pretreated with functional blocking anti-EGFR Ab, whereas the presence of AG1478 (a pharmalogical inhibitor of EGFR) completely abrogated Mcl-1 upregulation in HCMV-infected monocytes (Fig. 7A). Similarly, analysis of Mcl-1 protein levels from HCMV-infected monocytes pretreated with anti-EGFR Ab or AG1478 paralleled the transcriptional data (Fig. 7B). Densitometic analysis revealed a 50% and 95% decrease in Mcl-1 protein expression in anti-EGFR Ab and AG1478 pretreated HCMV-infected monocytes, respectively. Treatment with anti-EGFR Ab or AG1478 did not affect Mcl-1 protein levels in (Fig. 7C), virion binding to (20) or cell viability of (data not shown) mock-infected cells, suggesting that inhibition EGFR signal does not affect basal Mcl-1 expression. These data indicated that Mcl-1 is rapidly upregulated after infection with HCMV at the transcriptional level in a manner dependent on the activation of EGFR during viral binding.

Because HCMV-infected monocytes pretreated with EGFR inhibitors displayed lower levels of Mcl-1, we next asked if these cells were hypersensitive to LY294002-induced apoptosis. Mock-infected monocytes treated with LY294002 for 24 h exhibited a 5-fold induction of apoptosis, whereas monocytes infected for 24 h prior to treatment with LY294002 exhibited similar levels of apoptosis to that of the untreated control cells (Fig. 7D). However, inhibition of EGFR signaling prior to 24 h infection significantly reduced the HCMV mediated protection from LY294002-induced apoptosis. HCMV-infected monocytes pretreated with anti-EGFR Ab and AG1478 exhibited a 3-fold and 6-fold increase in apoptosis in response to LY294002 treatment, respectively. The heightened sensitivity of AG1478 pretreated HCMV-infected monocytes to LY294002 treatment is consistent with the lower levels of Mcl-1 expression observed in AG1478 pretreated infected cells (Fig. 7B). Overall, we propose that engagement and activation of EGFR by HCMV during viral binding/entry rapidly upregulates Mcl-1, which generates an antiapoptotic microcellular environment. Biologically enhanced survival of infected monocytes would support viral spread by allowing the infected monocyte to survive until differentiation into virus replication permissive macrophages can occur.

We have previously shown that HCMV infection of monocytes stimulates the polarization of infected cells toward a unique M1 proinflammatory phenotype, exhibiting secretion of M1 and M2 chemokines, hyper cell motility and monocyte-to-macrophage differentiation (15, 16). This HCMV-directed differentiation of infected monocytes into M1 macrophages is likely a critical step for viral persistence because HCMV requires macrophages for viral replication (16). However, for the monocyte-to-macrophage differentiation process to occur, HCMV-infected monocytes must be reprogrammed to survive past the 48–72 h viability gate. The lack of antiapoptotic viral IE proteins during this critical early cell fate decision period [(16), Fig. 3] suggests that HCMV has evolved a specific mechanism to use available cellular factors to promote the survival of monocytes during the early stages of infection. In the current study, we provide functional evidence that HCMV usurps the EGFR/PI3K signaling pathway to rapidly upregulate the antiapoptotic Mcl-1 protein after infection, thereby stimulating the acquisition of a prosurvival phenotype, a necessary step for the virally induced differentiation of monocytes to take place.

We speculate that Mcl-1 acts as a rapidly inducible, short-term effector of cell viability, allowing for the initiation of monocyte-to-macrophage differentiation, at least when maximally expressed. Indeed, Mcl-1 is rapidly and transiently upregulated in human myeloblastic leukemia cells on exposure to several differentiation-inducing agents and is believed to prevent cell death, thereby allowing cellular differentiation (36). Similarly, the transient induction of Mcl-1 in HCMV-infected monocytes suggest that once the intrinsic monocyte programming to undergo cell death has been circumvented by Mcl-1, high levels are no longer required for the survival of the differentiating monocyte. Our data indicate that an apoptotic regulatory switch occurs at ∼48–72 hpi when another antiapoptotic factor, such as Bcl-2, becomes the predominant survival factor (over Mcl-1) in the HCMV-infected differentiating monocyte. The effects of Mcl-1 on cell viability are not as prolonged as the effects of Bcl-2 (37), and Bcl-2 expression is critical to the long-term survival of cells differentiating along the myeloid lineage (27). We found the PI3K-independent induction of Bcl-2 protein expression in HCMV-infected monocytes occurs after 48 hpi. Studies are currently underway in our laboratory to determine whether Bcl-2 activity is critical for the long-term survival and differentiation of HCMV-infected monocytes. Together, these data suggest that HCMV may exploit the temporal differences in Bcl-2 family members to control short-term and long-term cell survival.

Although our data suggests that Mcl-1 and Bcl-2 play an essential role in the early and late phases of HCMV-infected monocyte survival, temporal transcriptome analysis of the infected monocyte provided evidence that several antiapoptotic factors are involved in the complex regulation of apoptosis during HCMV-induce monocyte differentiation. Ontology analysis revealed that 35 (0.27% of the total genes examined), 18 (0.14%), and 7 (0.06%) antiapoptotic genes were upregulated ≥1.5-fold at 4, 24, and 48 hpi, respectively (Table I), indicating a potentially complex regulation of the induction of a prosurvival state in HCMV-infected monocytes. The gradual decline in the number of upregulated antiapoptotic genes after initial infection indicated that those remaining elevated at later time postinfection, including Bcl-2 are likely involved in the long-term survival of the infected differentiating monocyte. Overall, the overwhelming upregulation of antiapoptotic gene expression during early infection suggest that HCMV evolved a mechanism to coerce the cellular signaling pathways into initiating a rapid conversion to an apoptotic resistant state, which we advocate is necessary for directing a favorable outcome when the critical cell fate decision is determined early in the monocyte-to-macrophage differentiation process.

The rapid degradation of Mcl-1 suggests that the antiapoptotic activity of this prosurvival factor is tightly controlled and crucial to the early regulation of the short lifespan of monocytes (38, 39). In accord with other studies, we show that the homeostatic expression of Mcl-1 rapidly declines during the initial 24–72 h of monocytes entering the peripheral circulation, suggesting to us that Mcl-1 may act as a biological clock to regulate the lifespan of unstimulated monocytes. We now show for the first time that Mcl-1 is directly involved in the survival of primary human monocytes and that HCMV infection is able to transcriptionally induce Mcl-1 expression, thus decreasing the rate of Mcl-1 loss and slowing the intrinsic monocyte proapoptotic clock. Similarly, transformation of normal monocytes into long-lived leukemia cell lines has been shown to require Mcl-1 (24, 40, 41). In addition, our data also indicate that posttranscriptional regulation of Mcl-1 occurs in monocytes after HCMV infection, because infected monocytes lacked elevated levels of Mcl-1 transcript at later times postinfection (Table I), but continued to exhibit high levels of Mcl-1 protein expression (Fig. 5A). Our data indicate that the aberrant induction of Mcl-1 in HCMV-infected monocytes provides the rapid prosurvival signal necessary for the positive outcome of the early 48–72 h cell fate decision period that infected monocytes must navigate prior to differentiation into long-lived macrophages.

Several monocyte/macrophage tropic pathogens appear to use the antiapoptotic properties of Mcl-1 to enhance host cell survival by inducing apoptotic resistance during infection (42–44). Mycobacterium tuberculosis required the expression of a bacterial gene product to induce Mcl-1 upregulation after infection (42), and HIV indirectly upregulated Mcl-1 expression via the secretion of M-CSF after infection (44). This convergent evolution displayed by biologically distinct pathogens in the regulation of Mcl-1 demonstrates the essential function that this antiapoptotic protein must have in regulating the viability of cells in the myeloid lineage. We now show for the first time a pathogen that has evolved a strategy to directly regulate the cellular signaling pathways responsible for modulating Mcl-1 expression. The unique EGFR-activating properties of HCMV provide the virus with a distinct survival advantage by immediately inducing Mcl-1 expression without the need of an intermediary signaling pathway or molecule. Unlike the aforementioned pathogens, the lack of HCMV replication in monocytes necessitates the rapid regulation of the host cellular anti-apoptotic pathways to ensure differentiation into viral replication-permissive macrophages.

In addition to the EGFR/PI3K signaling cascade, our data also suggest the involvement of other receptors and signaling pathways in the HCMV-mediated induction of Mcl-1. First, inhibition of EGFR by anti-EGFR Ab was 50% less effective than inhibition with AG1478. Integrins are activated during HCMV entry (45, 46) and are able to phosphorylate the cytoplasmic kinase domain of EGFR (47); thus, unlike anti-EGFR Ab, AG1478 is able to block crosstalk activation of EGFR by other receptors (Fig. 8). Second, although we observed increasing levels of PI3K through 96 hpi (Fig. 1A), the elevated levels of Mcl-1 message and protein dissipated by 72 hpi, suggesting a distinct combination of signaling events during HCMV infection is necessary to modulate Mcl-1 expression and function (Fig. 8). Nonetheless, our data indicates that the EGFR/PI3K signaling cascade is a necessary component of the early monocyte signalosome required for the increased production of Mcl-1 and the acquisition of a prosurvival phenotype after infection with HCMV.

HCMV engagement of EGFR can mediate multiple early events necessary for hematogenous dissemination, including viral entry (independent of PI3K activity) and cellular motility (dependent on PI3K activity) (20). We now provide evidence of an additional biological function occurring in HCMV-infected monocytes that is mediated by the EGFR/PI3K cascade: the rapid acquisition of a prosurvival state. Our current study indicates that HCMV activates the EGFR signaling pathway and the downstream PI3K activity during viral binding/entry to induce Mcl-1. Overall, we suggest that a unique combination of receptor-ligand events during binding/viral entry is responsible for mediating the early antiapoptotic state necessary for the survival monocytes through the 48–78 h viability gate; thus, allowing for the virally induced monocyte-to-macrophage differentiation and viral spread to occur. By deciphering how HCMV modulates these signal transduction pathways to promote viral dissemination and persistence, we hope to provide insight to the immunopathogenesis of HCMV and, thus, potentially identify novel therapeutic targets.

Disclosures The authors have no financial conflicts of interest.