Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

In bone-marrow and lung-transplant recipients, human cytomegalovirus (CMV) infection is poorly controlled, resulting in persistent infection of many lung cells, including endothelial cells. Although a mononuclear cell infiltrate is prevalent during CMV pneumonitis, it apparently fails to control infection (1, 2). The mechanisms by which CMV persists in the face of a chronic inflammatory response is not fully understood. We hypothesize that CMV infection of endothelium actively modulates local concentrations of chemotactic chemokines. Altering chemokine levels may thereby protect the infected cells from infiltrating mononuclear cells.

In response to cell injury, endothelial cells express a variety of chemokines, both indirectly, in response to exposure to proinflammatory cytokines that are also expressed during injury, and directly, in response to injury. Cytokines expressed during cell injury, such as the proinflammatory cytokines tumor necrosis factor- and interferon-, induce endothelial cells to express high levels of the CC chemokine regulated on activation, normal T cell expressed and secreted (RANTES) (3). Although the chemokine response of endothelial cells to CMV infection is poorly characterized, exposure of fibroblasts to the attenuated AD169 strain of CMV stimulates expression of high levels of RANTES independent of an active CMV genome (4). Hence, the local environment during infection can be predicted to be enriched in chemokines. Bronchoalveolar lavage studies of patients with CMV pneumonitis indicate that significant levels of RANTES are present during infection (5).

We have previously described a CMV-encoded chemokine receptor, CMV US28, that is expressed on the cell surface during CMV infection (6). CMV US28 functions in response to RANTES ligation to activate intracellular signaling pathways leading to cell-proliferative responses (6). In this report we propose that depletion of extracellular RANTES during CMV infection of endothelial cells is mediated by CMV-encoded receptor US28.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Recombinant chemokines RANTES, macrophage chemotactic protein (MCP)-1, macrophage inflammatory protein (MIP)-1 and interleukin (IL)-8, and Quantikine Immunoassay kits for assay of chemokine protein, were obtained from R&D Systems (Minneapolis, MN).

Cells

Human umbilical vein endothelial cells (HUVECs) were harvested from umbilical veins according to procedures described elsewhere (7), which were further modified for CMV infection (6). Human kidney epithelial cells transformed with adenovirus (type 293 cells) and which stably express CMV US28 are described elsewhere (6).

Virus

Human CMV strain 4010 has been previously described (6). Cell-free virus was prepared from supernatants of strain 4010-infected HUVEC culture. The supernatant was spun through low-speed centrifugation to remove cells and debris, and was then subjected to ultracentrifugation at 45,000 × g at 4°C in an SW28 rotor (L5-50 Ultracentrifuge; Beckman Instruments, Mountain View, CA). The pellet, containing concentrated virus particles, was resuspended in medium and stored at 80°C. For assays involving CMV infection of HUVECs, subconfluent HUVEC monolayers were infected with cell-free CMV at a multiplicity of infection of 0.1 pfu/cell. The virus was adsorbed to the cells for 90 min at 37°C, the input virus was removed, and the monolayers were fed with fresh culture medium. Mock-infected HUVEC monolayers were cultured in parallel with the actively infected monolayers. The CMV- and mock-infected HUVECs were incubated at 37°C for 5-7 d, until the cytopathic effect typical of CMV infection was observed over 80% of the monolayer.

Internalization of [¹²⁵I]RANTES

Monolayers were set up of type 293 and US28-expressing type 293 cells, HUVECs, and CMV-infected HUVECs (strain 4010) in 24-well trays until 70% confluent. Through an adaptation of methods described elsewhere (8), cells were equilibrated in Hanks' balanced salt solution/1% bovine serum albumin before incubation for 1 h on ice with 10 pM [¹²⁵I]RANTES obtained from NEN-Dupont Life Science Products (Boston, MA). The temperature was raised to 37°C, and at 0, 10, and 30 min the cells were rinsed with phosphate-buffered saline (PBS) and incubated in acid (0.2 M acetic acid/50 mM NaCl, pH 2.5) for 5 min on ice. The acid wash (acid-sensitive labeled RANTES [cell-surface]) was removed and the cells were lysed in 0.5 N NaOH (acid-resistant labeled RANTES [internalized]), with the lysate collected and its gamma emission counted on a 1282 Compugamma CS Universal Gamma Counter (Wallac, Inc., Gaithersburg, MD). Specific binding represented the percent of total added counts of [¹²⁵I]RANTES per 200,000 cells in the 293/US28 cells and 293 cells, and per 75,000 cells in the CMV-infected-HUVEC and HUVEC cells. Statistical analysis of the significance of the difference between each condition at each time point was done by comparing group mean values through two-way analysis of variance.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

In order to determine whether extracellular concentrations of chemokines are altered during CMV infection of the endothelium, we added excess recombinant RANTES and IL-8 to cultures of CMV-infected endothelial cells (HUVECs). Subconfluent monolayers of HUVECs were infected with CMV strain 4010 (6) or were mock-infected, and at 24 h after absorption of CMV to the cells, excess RANTES (2.5 ng) and IL-8 (1.5 ng) were added to CMV- and mock-infected cultures (the chemokines were also incubated in cell-free medium in order to measure their degradation during culture). During the 6 d of culture, supernatants were collected and assayed for chemokine concentrations (Figure 1). RANTES disappeared from the cultures of CMV-infected HUVECs, but not from the supernatants of uninfected HUVECs or from medium alone (Figure 1). Moreover, there was no difference in the levels of IL-8 in the supernatants of CMV-infected and uninfected HUVECs; the concentration of IL-8 increased above the added dose in both culture systems. These results suggest that at late times of infection of HUVECs (72 h after infection), CMV regulates extracellular concentrations of RANTES.

View larger version (10K): [in this window] [in a new window]   Figure 1.   Depletion of extracellular RANTES during CMV infection of HUVECs. Three nanograms of recombinant RANTES (A) or 1 ng recombinant IL-8 (B) were added to the culture medium of monolayers of CMV(4010)-infected HUVECs (closed circles), uninfected HUVECs (open circles), or cell-free medium (open squares). The supernatants were collected during culture and assayed for chemokine protein. Each graph is representative of three replicative experiments.

Our previous work showed that CMV encodes a RANTES receptor, US28, that is expressed at 72 h after infection and which functions to stimulate intracellular signaling pathways in response to stimulation by RANTES (6). We stably expressed the full-length US28 in type 293 cells, a mammalian cell line that has no detectable RANTES receptor. The timing of expression of US28 on the infected cell surface, which occurred concomitantly with the depletion of RANTES from the extracellular milieu, suggested a link between US28 and RANTES depletion. Indeed, addition of recombinant RANTES to cultures of US28-expressing type 293 cells demonstrated that RANTES was rapidly depleted (within 60 min) in cultures of US28-expressing type 293 cells but not in cultures of non-US28-expressing type 293 cells (Figure 2), thereby indicating that US28 plays an important role in depletion of RANTES during infection.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Depletion of CC chemokines during culture of US28-expressing type 293 cells. Monolayers of type 293 cells that stably express US28 (closed circles) and non-US28-expressing type 293 cells (open circles) were incubated in culture with recombinant RANTES, MCP-1, MIP-1, and IL-8. At specific times during culture, the supernatants were collected and assayed for chemokine protein. A parallel preparation of each chemokine in conditioned medium without cells (closed triangles) was used to measure its degradation in culture medium over time. Each graph is representative of three replicative experiments.

Two other CC chemokines, MCP-1 and MIP-1, and the CXC chemokine IL-8, were added to parallel cultures of type 293 cells and type 293 cells expressing US28 (Figure 2). During the same time course as that of RANTES depletion, there was insignificant depletion of these chemokines in the cultures of type 293 cells expressing US28. These results suggested that RANTES was preferentially depleted from the extracellular milieu of type 293 cells expressing US28.

In order to determine whether the US28-specific depletion of extracellular RANTES was the result of proteolytic factors secreted by the cells that expressed US28, we added RANTES to conditioned medium from US28-expressing and non-US28-expressing cells (data not shown). The level of RANTES protein remained constant in conditioned medium without cells. RANTES was depleted from the extracellular milieu of US28-expressing type 293 cells at a similar rate in both fresh and conditioned medium. Thus, depletion of RANTES did not represent degradation of this chemokine by proteolytic factors secreted from the cells.

Although the foregoing results indicated that depletion of RANTES during CMV infection was mediated by CMV-encoded US28, we wished to determine whether the depletion occurred through binding per se, or, since US28 is a G-protein-coupled receptor, through binding followed by receptor-mediated internalization. Our initial experiments addressed the relative binding of RANTES to the two cell cultures expressing US28. We determined that RANTES had a higher specific binding index per 20,000 cells for CMV-infected HUVECs (26.2 ± 2.5%) than for US28-expressing type 293 cells (8.8 ± 0.4%), whereas uninfected HUVECs showed minimal binding (4.4 ± 0.7%). In order to determine the relative efficiency of internalization of RANTES, we incubated CMV-infected HUVECs (at Day 4 of infection) and US28-expressing type 293 cells with [¹²⁵I]-labeled-RANTES at 4°C for 1 h, and then rinsed the cells with PBS. The temperature was raised to 37°C for 0, 10, and 30 min to allow internalization of bound RANTES. At each time point, the cells were incubated with acid (0.2 M acetic acid/50 mM NaCl, pH 2.5) for 5 min on ice to remove bound extracellular RANTES. The acid rinses were removed and cells were lysed in 0.5 N NaOH to recover intracellular RANTES. Acid rinses and cell lysates were collected and their gamma emission was counted. Following absorption at 4°C, labeled RANTES was detected only in the acid rinses of CMV-infected HUVECs and US28-expressing type 293 cells. Labeled RANTES was not detected in either the cells or acid rinses from uninfected HUVECs or non-US28-expressing type 293 cells (Figure 3). As the CMV-infected HUVECs and US28-expressing type 293 cells were warmed at 37°C, less labeled RANTES was detected in the acid rinses and, conversely, the amount of labeled RANTES in the intracellular fraction increased, indicating internalization (Figure 3).

View larger version (13K): [in this window] [in a new window]   Figure 3.   Internalization of [125I]RANTES in US28-expressing cells. (a) US28-expressing type 293 cells, (b) type 293 cells, and (c) CMV-infected HUVECs and uninfected HUVECs. Cell-surface [125I]RANTES: closed circles: CMV-HUVECs or US28-expressing type 293 cells; open circles: HUVECs or type 293 cells. Internalized [125I]RANTES: closed squares: CMV-infected HUVECs or US28-expressing type 293 cells; open squares: HUVECs or type 293 cells. Specific binding represents the percent of total added counts of [125I]RANTES per 200,000 cells in the US28-expressing type 293 cells and type 293 cells, and per 75,000 cells in the CMV-infected HUVECs and HUVECs. The differences between the cell-surface RANTES and the internalized RANTES over time in the US28-expressing type 293 cells were significant (P < 0.0001). The values for CMV-infected HUVECs represent a triplicate experiment that showed a similar pattern of internalization of RANTES to that shown by the US28-expressing type 293 cells.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Using an in vitro model of CMV infection in primary endothelial cells, we found that during CMV infection, extracellular concentrations of the chemokine RANTES are depleted. Furthermore, we found that the mechanism of chemokine depletion involves receptor-mediated internalization of RANTES by the CMV-encoded chemokine receptor US28.

Our initial experiments addressed the question of whether there might be virally induced mechanisms that regulate extracellular concentrations of RANTES and IL-8. We were interested in RANTES not only because of its presence during CMV infection in vitro and in vivo (4, 5), but also because RANTES recruits monocytes, macrophages, and T cells. We found that excess RANTES was depleted from cultures of CMV-infected HUVECs, but that there was no significant loss of RANTES from the supernatants of uninfected HUVECs or medium alone (Figure 1). Moreover, there was no difference in the levels of IL-8 in the supernatants of CMV-infected and uninfected HUVECs; in fact, the concentration of IL-8 increased above the added dose in both culture systems. Thus, at the later time of 72 h after infection, CMV infection regulates extracellular concentrations of RANTES in HUVECs.

Our previous work indicated that CMV encodes a RANTES receptor, US28, that not only binds RANTES but also functions in response to ligand binding to activate intracellular signaling pathways linked to cell-proliferative responses (6). Our data determined that RANTES bound efficiently to CMV-infected cells no earlier than 4 d (96 h) after infection, thereby suggesting that active infection was necessary for expression of the functional receptor US28 on the cell surface. Using type 293 cells that stably express US28, we determined that depletion of RANTES during CMV infection of HUVECs is associated with expression of US28. Extracellular RANTES was rapidly depleted (within 60 min) from cultures of type 293 cells expressing US28, but not from cultures of non-US28-expressing type 293 cells (Figure 2). The slope of RANTES depletion may represent several factors, including the rate of de novo expression of functional US28 receptors on the cell surface, and the rate of regeneration of functional US28 receptors. In contrast to RANTES depletion, other CC chemokines (MIP-1, MCP-1) and the CXC chemokine IL-8 were not depleted from the extracellular milieu of type 293 cells expressing US28. Further studies indicated that US28-specific depletion of extracellular RANTES was not the result of proteolytic factors secreted by the cells expressing US28. The level of RANTES protein remained constant in conditioned medium without cells, and US28-specific depletion was unaffected by conditioned medium.

We determined the mechanism of US28-mediated depletion of RANTES by tracking the presence of radiolabeled RANTES on both infected HUVECs and US28- expressing type 293 cells. The loss of labeled RANTES on the cell surface concomitantly with an increase in labeled RANTES in the intracellular fraction indicated that cell-surface-bound RANTES became internalized in the CMV-infected HUVECs and US28-expressing type 293 cells. These results suggested that RANTES is internalized after binding to CMV-infected HUVECs, and that this internalization is mediated by the chemokine receptor US28. The rate of RANTES internalization in the CMV-infected HUVECs did not appear to differ from that in the US28-expressing type 293 cells, suggesting that expression of US28 is sufficient to permit internalization of RANTES. Further studies are underway to define the mechanism of internalization of RANTES and US28, and whether US28 is recycled to the cell surface. That the curve in Figure 2 depicting RANTES depletion continues to demonstrate a depletion of RANTES over time suggests that the receptor is being recycled to the cell surface.

These findings complement the recently reported studies by Bodaghi and colleagues of extracellular CC chemokine depletion during CMV infection of fibroblasts, which occurred independently of the CMV induction of RANTES expression found earlier by these investigators (4, 9). Bodaghi and colleagues invoked a role for the CMV- encoded chemokine receptor US28 by infecting cultures of fibroblast cells with mutant strains of CMV that do not express US28. These cells did not show depletion of extracellular CC chemokines (9). Thus, Bodaghi and colleagues' studies demonstrated that US28 expression was necessary for the depletion of RANTES, but the mechanism was not elucidated. In our study, the stable expression of US28 by type 293 cells provided a means of determining that US28, independently of other CMV proteins, was sufficient for depletion of RANTES from the extracellular milieu. Furthermore, our study provides data indicating that bound RANTES is internalized. The results of our study extend and confirm the findings by Bodaghi and colleagues that the CC chemokine RANTES bound on the surface of CMV-infected cells is indeed internalized by a specific receptor-mediated mechanism involving CMV-encoded US28. It should be noted, however, that in the CMV-infected fibroblast system, internalization of RANTES did not strictly correlate with US28 expression (9), whereas in our system of clinical isolate CMV-infected HUVECs, RANTES internalization was dependent on US28 expression. This discordance between the findings of our study and that of Bodaghi and colleagues may reflect the different assay procedures used for investigating internalization of RANTES, but more likely reflects the different phenotypes of the two strains of CMV and the different host cells used in the respective study systems.

Moreover, it has been well documented that CC chemokine receptors (CCR5, CCR3) other than US28 function as coreceptors for the intracellular entry of human immunodeficiency virus (HIV) and that binding of RANTES and other CC chemokines to these receptors blocks their ability to function as coreceptors (10, 11). The relationship between CMV infection and the ability to regulate chemokine concentrations, combined with the recognized ability of chemokines to modify HIV infectivity, argues for further work in this complex area.