Human Cytomegalovirus-Induced Degradation of CYTIP Modulates Dendritic Cell Adhesion and Migration

As potent antigen-presenting cells, dendritic cells (DCs) are essential for the initiation of effective antiviral immune responses. Viruses and especially herpesviruses, which are able to establish lifelong persistence, exploit several immune evasion mechanisms targeting DC biology. Our group has previously shown that the α-herpesvirus herpes simplex virus type 1 inhibits mature DC (mDC) migration by inducing adhesion via degrading the cellular protein CYTIP (cytohesin-1 interacting protein), an important negative regulator of β2-integrin activity. In the present study, we extended our analysis to the β-herpesvirus human cytomegalovirus (HCMV), to investigate whether other herpesviridae also induce such modulations. Indeed, HCMV impairs mDC transwell migration capability following a CCL19-chemokine gradient, despite equivalent expression levels of the cognate chemokine receptor CCR7 at the corresponding time points post-infection. Remarkably, HCMV infection potently induced β2-integrin activity on mDCs. Furthermore, directly HCMV-infected mDCs, exhibiting viral gene expression, strongly adhere to fibronectin and ICAM-1, in contrast to mDCs lacking infection or viral gene expression. Interestingly, HCMV-positive mDCs display a proteasome-dependent degradation of CYTIP. Contrasting the migration toward CCL19, elevated expression levels of the chemokine receptor CXCR4 in HCMV-infected mDCs were associated with functional CXCL12-chemotaxis under the herein used conditions. In summary, our results show that HCMV shapes mDC adhesion to compromise migration toward CCL19, but retaining CXCL12 responsiveness. Thus, we hypothesize that a preferred migration pattern toward the bone marrow, but not to secondary lymphoid organs, could ultimately cause a failure in the induction of potent antiviral immune responses.

The β-herpesvirus human cytomegalovirus (HCMV) exhibits seroprevalences of 45% up to almost 100% depending on age, gender and socioeconomic situation (35). While the primary infection of healthy adult individuals is usually subclinical, infections of immune-immature fetus or neonates as well as immunocompromised patients frequently leads to severe symptoms with high morbidity and mortality (36). Permissive target cells for HCMV replication are fibroblasts, smooth muscle cells, epithelial cells, and endothelial cells (37,38). Apart from the productive lytic replication, herpesviruses establish lifelong latency in the host, despite the presence of an immune response. Regarding HCMV, CD34 + hematopoietic progenitor cells (39)(40)(41) and CD14 + monocytes (42,43) are widely accepted reservoirs for latency.
Since DCs are important stimulators of antiviral immune responses, herpesviruses evolved several strategies to interfere with DC functions. One example is the herpes simplex virus type 1 (HSV-1)-induced inhibition of mDC migration. This is mediated via the downregulation of CCR7 and CXCR4 surface expression as well as the activation of LFA-1 via the degradation of CYTIP leading to significantly induced adhesion of mDCs (44,45). However, it was unknown whether this represents a unique immune evasion mechanism of HSV-1 or if other members of the herpesviridae family including HCMV also regulate mDC adhesion and migration, which would imply an analogous or even evolutionary conserved immune evasion mechanism.
In the present study, we provide evidence that HCMV modulates mDC adhesion and migration. We show a differential regulation of CCR7 and CXCR4 upon HCMV infection of mDCs. Moreover, HCMV induces activation of β2-integrins via proteasomal degradation of CYTIP in mDCs in which HCMV established viral gene expression. This leads to increased adhesion of infected mDCs to ICAM-1 and fibronectin, which in turn influences the mDC migration capability. Hence, we suggest that this immune evasion mechanism has been evolved by at least one prototypic member of the αas well as β-herpesviridae family, to hamper the induction of potent antiviral immune responses.

cells and Viruses
The HSV-1/17+/CMV-EGFP/UL43 strain, used in this study, contains the EGFP reporter cassette under control of the CMV promoter inserted into the UL43 locus of HSV-1 strain 17+. Virus stock preparation and titer determination were performed as described previously (49). The HSV-1 strain was obtained from BioVex.
For HCMV infection, the HCMV TB40E/IE2-EYFP strain (50) expressing IE2-EYFP fusion protein (kindly provided by Thomas Stamminger, Institute for Clinical and Molecular Virology, Erlangen) or HCMV-TB40E-green fluorescence protein (GFP) was used. HCMV stocks were prepared on ARPE-19 cells, followed by propagation on human fibroblasts (HFF or MRC-5 cells), cultured in Dulbecco's Modified Eagel's Medium (DMEM; Lonza) supplemented with 10% (v/v) FCS (PAA), 2 mM l-glutamine (Lonza), 100 U penicillin ml −1 (Lonza) and 100 mg streptomycin ml −1 (Lonza). Virus-containing supernatants were harvested when all cells showed cytopathic effects. Remaining adherent cells were scraped from the flask and added to the supernatant. Cellular debris as well as intact cells were collected by centrifugation and homogenized using a tissue grinder (VWR) to release intracellular virions. Cellular debris in the virus suspensions were finally removed by centrifugation, and virus-containing supernatant was stored at −80°C. Titration of the respective HCMV strains was carried out as described elsewhere (51). Ultraviolet (UV)-inactivated virus (HCMV-UV) was generated by irradiation of HCMV-containing suspensions with defined doses of UV light (1,200 J/cm 2 ).

hsV-1 infection of mDcs
Mature DCs were harvested and washed in PBS. Cells (1 × 10 6 ) were resuspended in 300 µl RPMI 1640 (Lonza) supple mented with 20 mM HEPES (N-2-hydroxyethylpiperazine-N-2ethanesulfonic acid, Lonza). Cells were then incubated with HSV-1 at a multiplicity of infection (MOI) of 2 for 1 h at 37°C and shaking at 300 rpm. After collecting the cells by centrifugation, they were cultured in DC medium supplemented with 40 U GM-CSF ml −1 and 250 U IL-4 ml −1 until further use. For proteasomal inhibition, 10 µM of MG-132 (Enzo Life Science) or dimethyl sulfoxide (Sigma-Aldrich) as control was added 1 hour post-infection (hpi).

hcMV infection of mDcs
Mature DCs were collected and washed once in PBS. Cells (5 × 10 5 ) were resuspended in the respective volume of HCMV virus stock at an MOI of 3. For mock condition, mDCs were resuspended in the respective volume of DMEM supplemented with 10% FCS, 2 mM l-glutamine, 100 U penicillin ml −1 and 100 mg streptomycin ml −1 . For infection, mDCs were centrifuged at 1,500 × g at room temperature for 1 h. Afterward, cells were cultured in DC medium containing 40 U GM-CSF ml −1 and 250 U IL-4 ml −1 in a final volume of 1 ml.

Fibronectin adhesion assay
Fibronectin coating of 96-well Maxisorp plates (Nunc) was performed using 20 µg fibronectin ml −1 (Sigma-Aldrich) in PBS with incubation at 4°C overnight. Subsequently, wells were blocked with 0.01% (w/v) gelatin (Sigma-Aldrich) in PBS for 2 h at room temperature. Directly before adhesion wells were washed in PBS. After harvesting 24 hpi, mDCs were washed once in PBS and adjusted to 4 × 10 5 cells per ml in RPMI 1640. Per well, 100 µl of the DC suspension was added, and cells were allowed to adhere for 45 min at 37°C. Subsequently, supernatant was removed, and non-adherent mDCs were washed off with RPMI 1640. Adherent fraction was rinsed with PBS and fixed with 2% (w/v) PFA. To determine input conditions the respective volume of the cell suspension was preserved before initiation of the adhesion assay. Each experimental condition was performed with a total cell number of 2 × 10 5 cells. Infection rates before and after adhesion were determined using flow cytometry. To calculate the number of adherent cells, their β-glucuronidase activity was measured as described previously (44). Briefly, cells were resuspended in 25 µl PBS and 5 µl 1% TritonX-100 in PBS was added. After 20-min incubation at room temperature, cell debris was removed by centrifugation at 18,400 × g for 1 min, and lysates were incubated for 6 h at 37°C with 75 µl 10 mM 4-nitrophenyl-β-d-glucuronide (Sigma-Aldrich) solved in 100 mM sodium acetate, pH 4. Subsequently, 100 µl 0.4 M glycine, pH 10, was added, and the absorbance at 405 nm was measured using a Victor 2 multilabel counter (PerkinElmer, Jugesheim, Germany). The number of adherent cells was calculated via a separate standard curve for each sample using cells of the input condition.

icaM adhesion assay
For ICAM-1 coating of 96-well Maxisorp plates (Nunc), wells were coated with 4 µg goat anti-human IgG (Fc specific) per ml (Jackson ImmunoResearch Laboratories) in PBS at 4°C overnight. After blocking with 1% (w/v) BSA in PBS at 37°C for 30-min, wells were washed with PBS and subsequently incubated with 500 ng ICAM-1-Fc ml −1 in PBS for 1 h at 37°C. Adhesion of mDCs and quantification of infection rates were performed as described above for the fibronectin adhesion assay.

Transwell Migration assay
Membranes of transwell inserts with a pore size of 5 µm (Corning Costar) were coated with 20 µg fibronectin ml −1 at 4°C overnight. Transwell migration assays were performed as described elsewhere (44). In brief, mDCs were harvested 24 hpi, washed once in PBS and adjusted to 2 × 10 6 cells per ml in DC medium. To eliminate cell clusters, cells were passed with a syringe three times through a 20 gage needle. Transwell inserts were equilibrated with 100 µl DC medium, while 600 µl of medium was added to the wells of a 24-well plate. For migration assay, 2 × 10 5 cells were seeded into the inserts and allowed to migrate at 37°C for 2 h toward chemokine gradients of CCL19 or CXCL12 (100 ng ml −1 , Tebu-Bio GmbH), added to the lower compartment. Additionally, the respective amount of cells was preserved as starting condition. Afterward, migrated DCs were harvested from the lower chamber. Cell numbers and infection rates before and after migration were determined using flow cytometry.

statistical analysis
For determination of significance, data sets of two conditions were analyzed using two-tailed Student's t-test. For multiple data sets, one-way analysis of variance was applied. Results are displayed as mean ± SD. Significance was accepted for p-values less than 0.05. **** indicates p ≤ 0.0001; ***p ≤ 0.001; **p ≤ 0.01; *p ≤ 0.05; and ns, not significant. resUlTs hcMV reduces mDc Transwell Migration toward ccl19, but not cXcl12 As migration of DCs to SLOs is pivotal for induction of immune responses by activation of naïve T cells, DCs constitute potential targets for viral immune evasion strategies. Previously, we showed that HSV-1 efficiently infects monocyte-derived mDCs and inhibits transwell migration toward the SLO-expressed chemokines CCL19 and CXCL12 (44). In order to investigate whether this modulation of migration is a unique mechanism of HSV-1 or also exist in other herpesviridae, we analyzed the influence of HCMV, which represents the prototype of the β-herpesvirus family.
For all infection experiments performed in the present study, the endotheliotropic strain HCMV-TB40E expressing GFP or an IE2-EYFP (enhanced yellow fluorescent protein) fusion protein was used. The use of the GFP/YFP-tagged HCMV-TB40E strain allows the discrimination of HCMV-positive mDCs that express viral-encoded genes and HCMV-negative mDCs, which were exposed to viral particles and antigens but did not express HCMV-encoded genes. Figure 1A shows a time course analysis of HCMV-infected monocyte-derived mDCs. At the indicated time points, cells were analyzed via flow cytometry for GFP/YFP fluorescence, and approximately 15% of mDCs expressed EGFP as undeniable indicator of HCMV-induced gene expression 24 hpi, with increasing percentages at later time points (Figure 1A). Furthermore, mDCs were infected with UV-inactivated HCMV (HCMV-UV) and analyzed 24 hpi to ensure UV-irradiation efficiency.
To assess the influence of HCMV on mDC chemotaxis, we subjected mock-, HCMV-or HCMV-UV-infected mDCs 24 hpi to transwell migration assays using fibronectin-coated polycarbonate membranes (Figure 1B). After 2 h of migration without chemokine as well as toward CCL19 or CXCL12, cells in the lower well were counted and infection efficiency was determined before and after migration. No change in CXCL12-mediated chemotaxis was observed comparing mock-(white bar), HCMVpositive (green bar), HCMV-negative (black bar), and HCMV-UV-infected (gray bar) mDCs, indicating that HCMV does not compromise the general ability of mDCs to migrate to chemotactic gradients. However, spontaneous migration without chemokines (w/o) was significantly inhibited for HCMV-positive mDCs compared to mock condition. Interestingly, we observed a significantly reduced migration capability for HCMV-positive mDCs in CCL19-induced chemotaxis compared to mock cells, HCMV-negative as well as HCMV-UV-infected mDCs ( Figure 1B).

surface expression of ccr7 and cXcr4 in hcMV-infected mDcs Does not explain impaired Migration
Chemotaxis in a transwell system relies on the expression of the respective receptors on the cell surface. Our group previously reported that HSV-1 induces significant downregulation of surface expression of the receptors CCR7 and CXCR4, which are required to recognize CCL19 and CXCL12, respectively (44). To determine alterations in surface expression of both chemokine receptors on HCMV-infected mDCs, we performed time course experiments, with subsequent flow cytometric analyses (Figure 2). The results of one representative experiment of mock, HCMV-positive, HCMV-negative, and HCMV-UV-infected mDCs 24 hpi are shown in Figures 2A,C, while the results of at least three independent experiments in a time kinetic are summarized in Figures 2B,D. We demonstrate that CCR7 surface expression is unaffected until 24 hpi-the time frame in which the transwell assay was conducted-followed by a weak downregulation at later time points only in HCMVpositive mDCs (Figure 2B, left panel). Remarkably, in contrast to the downregulation of CCR7, HCMV even induced CXCR4 expression already 12 hpi, peaking at 24 hpi with an induction of approximately 150% only observed in HCMV-positive mDCs (Figure 2D, left panel). No changes in CCR7 and CXCR4 surface expression were detected on HCMV-negative (Figures 2B,D, right panels) as well as on HCMV-UV-infected mDCs ( Figure 2E).
The surface disposition of these two chemokine receptors was altered on HCMV-positive mDCs but failed to explain their impaired transwell migration observed within the first 24 h of mDC infection. Hence, additional mechanism(s), apart from meaningful downregulation of CCR7 surface expression, were considered as cause for the observed inhibition of transwell migration toward CCL19.

hcMV induces mDc adhesion by increasing β2-integrin activity leading to impaired ccl19-Directed Transwell Migration
Having demonstrated that HCMV does not downregulate surface expression of CCR7 until 24 hpi, we intended to elucidate the mechanism by which migration in this 2D system is inhibited. Apart from chemokine receptor expression, mDC adhesion on the fibronectin-coated transwell polycarbonate-membrane could be modulated by HCMV to inhibit migration. While fibronectin represents a prototypic ligand for a variety of cell adhesion molecules, belonging to β1-, β2-, β3and β7-integrin families (53, 54), ICAM-1 is a specific ligand for the β2-integrins LFA-1 and Mac-1 (55). Therefore, we next analyzed adhesion of HCMV-infected mDCs on ICAM-1 or fibronectin-coated wells 24 hpi (Figure 3A). Input condition (light green bar) represents the infection rate before adhesion and was set to 100%. ICAM-1 as well as fibronectin increased the ratio of HCMV-positive cells among the adherent fractions (second and third green bar), demonstrating that HCMV-positive mDCs adhere stronger to both ligands.
Since we previously demonstrated that increased ICAM-1 and fibronectin adhesion of HSV-1-infected mDCs is based on the induced β2-integrin activity (45), we next examined the activation status of β2-integrins after HCMV infection using an antibody which recognizes a conformational epitope only present on activated β2-integrins (56-58). As a positive control we treated the cells with Mg/EGTA to induce β2-integrin activity via alteration of divalent cation concentration, since it is known that Mg 2+ promotes transition from the inactive into the active conformational state (52,59,60). Flow cytometric analyses of untreated mock (white bar) and HCMV-positive (green bar) mDCs revealed a significant induction of β2-integrin activity of approximately 2.5-fold, while HCMV-negative (black bar) mDCs show a less pronounced induction ( Figure 3C). No differences in β2-integrin activity were observed comparing mock and HCMV-UV-infected mDCs (gray bar). As expected, treatment with Mg/EGTA showed a strong increase in β2-integrin activity in mock, HCMV-negative, and HCMV-UV condition. Interestingly, Mg/EGTA treatment of HCMV-infected cells further increased the activation status of β2-integrins on HCMV-positive mDCs compared to their untreated counterparts.
In order to analyze whether the observed increase in β2-integrin activity shown in Figure 3C correlates with induced adhesion, a fibronectin adhesion assay was conducted ( Figure 3D). In agreement with the activation status of β2-integrins on HCMVpositive mDCs, we observed a significant induction in fibronectin adhesion compared to mock cells, while this was not detected for HCMV-negative and HCMV-UV-infected mDCs. Moreover, Mg/EGTA-treated cells adhered stronger to fibronectin compared to their respective untreated counterparts.
In order to evaluate whether the induced adhesion of HCMVpositive mDCs accounts for the reduced CCL19 transwell migration (Figure 1B), we performed a transwell migration assay with Mg/EGTA-treated mock cells to induce adhesion in the absence of viral infection (Figure 3E). Treatment with Mg/EGTA strongly reduced spontaneous and CCL19-directed chemotaxis comparable to the reduction observed for HCMV-positive mDCs (Figure 1B), while additional HCMV infection did not further reduce CCL19 transwell migration.
Therefore, we conclude that HCMV activates β2-integrins on infected mDCs to facilitate increased adhesion to fibronectin as well as ICAM-1, which in turn leads to reduced CCL19-mediated transwell migration.
hcMV infection of mDcs leads to Decreased expression levels of cYTiP, the negative regulator of β2-integrin activation Next, we aimed to unravel the molecular mechanism of the observed β2-integrin activation on HCMV-positive mDCs. Cytohesin-1 is known to bind to the intracellular CD18 tail of β2-integrins and thereby causes a conformational change leading to their activation (28,29,34). Furthermore, it was demonstrated that the cellular protein CYTIP negatively regulates LFA-1 activity and consequently ICAM-1 adhesion (30,45). In particular, using siRNA-mediated CYTIP ablation, it was previously shown that the loss of CYTIP suffices to induce mDC adhesion and reduce migration capability (45). Therefore, we next investigated cytohesin-1 and CYTIP expression in HCMV-infected mDCs. In this regard, mDCs were mock-or HCMV-infected and subjected to FACS, based on their GFP/YFP expression. As a control, HSV-1-infected mDCs were analyzed; however, due to the high infection efficiency of nearly 100%, cell sorting was dispensable in this case. Confirming previous results, CYTIP protein levels were lost in HSV-1-infected mDCs 24 hpi (Figure 4A shows one representative experiment and Figure 4C summarizes the quantification of at least three independent experiments). In contrast, protein expression of cytohesin-1, which is the positive regulator of LFA-1 activity, is only affected to a minor extent. Interestingly, also in HCMV-positive mDCs CYTIP protein expression levels are reduced to approximately 50% (24 hpi), while cytohesin-1 protein expression is not significantly influenced (Figures 4B,C).
Thus, these results provide evidence that HCMV induces downregulation of CYTIP in order to increase LFA-1 activation on the surface of infected mDCs leading to induced adhesion and reduced migration.

hcMV induces cYTiP Downmodulation early after infection
Since it is known that HCMV replicates slower than HSV-1, we analyzed CYTIP downregulation at different time points post-infection. Therefore, mDCs were mock-or HCMV-infected for the indicated time periods and subsequently subjected to Subsequently, cells were stained with an antibody specific for an epitope specifically present on activated CD11/CD18 (β2) integrins and subjected to flow cytometric analyses. Mock (white bar) was set to "1," and relative activation is shown for the indicated conditions. The experiment was performed at least three times with cells of different healthy donors. (D) Mature DCs were infected and treated with (striped bars) or without Mg/EGTA (filled bars) as described in (c) and subjected to fibronectin adhesion assay. Percentages of adherent cells are depicted for each condition (n = 3). (e) Mock-or HCMV-infected mDCs (24 hpi) were treated with Mg/EGTA (striped bars) and subjected to transwell migration assays. For induction of chemotaxis, CCL19 (100 ng/ml) was added to the lower wells, while spontaneous migration without chemokine was analyzed as control (w/o). Migration was evaluated after 2 h. The percentages of migrated mDCs were calculated in the lower well by determining cell number and infection rates of migrated cells in relation to the respective starting condition. Untreated mock cells are depicted as control (white bars). The experiment was performed with cells of three different healthy donors. Only significant changes (**** = p < 0.0001; *** = p < 0.001; ** = p < 0.01; * = p < 0.05) are marked by asterisks.  . (a,B) Protein lysates were prepared and subjected to Western blot analyses with antibodies specific for CYTIP, cytohesin-1, ICP0, IE2, or GAPDH. One representative experiment out of at least three is shown. (c) Quantification of CYTIP and cytohesin-1 protein expression levels in HSV-1-infected or HCMV-positive mDCs, summarizing at least three independent experiments. Expression of CYTIP and cytohesin-1 was normalized to the GAPDH control. Mock was set to 100%. Significant changes (**** = p < 0.0001; *** = p < 0.001) are marked by asterisks, non-significant changes (p > 0.05) as "ns." intracellular flow cytometric analyses for detection of CYTIP and cytohesin-1 levels ( Figure 5). One representative experiment for HCMV-positive (green histograms) and HCMV-negative (filled black histograms) mDCs is shown in Figures 5A,C, respectively. Summarized data out of at least three independent experiments are depicted as median values relative to mock in Figure 5B (HCMV-positive) and Figure 5D (HCMV-negative). CYTIP reduction is already observed 12 hpi in HCMV-positive mDCs, becoming statistically significant 24 hpi and even more prominent 36 hpi with a reduction of approximately 70% in comparison to mock-infected cells (Figure 5B). In contrast, cytohesin-1 is not downregulated to a significant extent, even 36 h post-HCMV infection. HCMV-negative mDCs showed no reduction in CYTIP and cytohesin-1 protein levels until 36 hpi ( Figure 5D). We, therefore, conclude that HCMV specifically downregulates CYTIP protein expression, while cytohesin-1 levels are not impaired. We further investigated the mechanism by which HCMV mediates CYTIP downmodulation. Since our previous findings in HSV-1-infected mDCs showed a proteasome-dependent pathway of CYTIP degradation (45), we examined whether a similar mechanism is responsible for CYTIP downmodulation in the context of an HCMV infection. Thus, mDCs were mock-or HCMV-infected for 24 h, treated with or without the proteasome inhibitor MG-132 and analyzed for their CYTIP as well as cytohesin-1 expression via intracellular flow cytometry (Figures 6A,B). One representative experiment is depicted in Figure 6A, which shows CYTIP (left panel) and cytohesin-1 (right panel) protein levels in HCMV-positive (green histograms) and HCMV-negative (black histograms) mDCs 24 hpi. Figure 6B summarizes the flow cytometric data from at least three independent experiments.
Remarkably, CYTIP protein expression levels were restored in MG-132-treated HCMV-positive mDCs (green striped pattern) to approximately 80%. Regarding cytohesin-1, no significant differences in protein expression levels could be observed. CYTIP and cytohesin-1 protein levels are not affected in HCMV-negative mDCs compared to mock ( Figure 6B, right panel). These observations are confirmed by immunofluorescence microscopy analyzing CYTIP and cytohesin-1 expression in mock-and HCMV-infected mDCs 24 hpi with or without MG-132 treatment ( Figure 6C).
Taken together, our present study demonstrates that HCMV induces a proteasomal degradation of CYTIP in HCMV-positive mDCs early after infection. This leads to induced β2-integrin activation, and therefore increased mDC adhesion, which in turn inhibits CCL19-directed mDC transwell migration. Thus, not only the α-herpesvirus HSV-1 but also the β-herpesvirus HCMV uses this strategy to hamper the induction of antiviral immune responses. However, since HCMV does not downregulate the expression of CCR7, the reduction of CCL19-mediated migration is most likely due to induced adhesion.

DiscUssiOn
Dendritic cells play a crucial role in the orchestration of immune responses by transferring the immunological status of peripheral tissues to areas with high lymphocyte density, i.e., to SLOs. There (naïve) T cells and/or B cells can be activated by presentation of antigens on DCs, in order to induce a potent immune response. Considering this, migration of DCs is a prerequisite for the successful initiation of adaptive immunity, important for elimination of e.g., viral infections. Particularly herpesviruses, which are able to establish lifelong persistent infections, have developed several strategies to counteract elimination by the immune system. In this study, we provide evidence that HCMV induces the activation of β2-integrins via the proteasomal degradation of the negative regulator CYTIP. This degradation of CYTIP is functionally relevant as we found induced adhesion and impaired transwell migration of HCMV-infected mDCs toward the chemokine CCL19, which is known to be predominantly expressed in lymph nodes. Therefore, immune evasion mechanisms targeting DC migration may allow HCMV to effectively evade cell-mediated immunity in vivo, supporting the production of new progeny virus and the establishment of latency.
Our finding that HCMV inhibits CCL19-mediated migration ( Figure 1B) is the first report in which HCMV was described to reduce mDC migration toward this specific chemokine exclusively via the induction of adhesion, i.e., by increasing the activation status of β2-integrins (Figure 3). Thus, HCMV is able to not only inhibit upregulation of the respective chemokine receptor CCR7, when infecting iDCs prior to maturation (61), but also hamper migration toward CCL19 beyond CCR7-targeting, via the induction of adhesion upon HCMV infection of mDCs (Figures 3D,E). This very likely caused the reduced CCL19-mediated migration using fibronectin-coated membranes, as expression of the CCL19-receptor CCR7 was not affected by HCMV until 24 hpi (Figure 2B). Adhesion is mediated by integrins, thus we investigated the activation status of β2-integrins in uninfected or HCMV-infected mDCs, since our group previously showed that fibronectin adhesion of mDCs is predominantly mediated by the β2-integrin LFA-1 (45). Interestingly, HCMV-positive mDCs exhibit significantly increased activation of β2-integrins, similar to that of Mg/EGTA-treated mDCs, in comparison to mock cells (Figure 3C), which in turn resulted in a strong induction of fibronectin adhesion ( Figure 3D). Additionally, adhesion of HCMV-infected mDCs on the LFA-1 specific ligand ICAM-1 was also strongly induced 24 hpi (Figure 3A). This hints toward a mechanism by which HCMV upregulates β2-integrin activation, i.e., LFA-1, and thus mDC adhesion in order to reduce migration. HSV-1-infected mDCs also showed impaired CCL19-mediated chemotaxis, however, to a far greater extent than for HCMV. This HSV-1-mediated inhibition of CCL19-specific-migration was not only due to stronger mDC adhesion upon HSV-1 infection, but also based on a decrease in CCR7 surface expression (44,45).
In contrast to the reduced CCL19-mediated migration in HCMV-positive mDCs, we could not observe changes in CXCL12-mediated chemotaxis compared to mock, HCMVnegative or HCMV-UV-infected mDCs ( Figure 1B). Strikingly, the surface expression of the respective chemokine receptor CXCR4 was found to be significantly induced on HCMV-positive mDCs ( Figure 2D). Previous reports demonstrated that infection of different cell types with HCMV has diverse effects on CXCR4. Lecointe et al. showed that HCMV infection of microglial cells, but not of macrophages or astrocytes, leads to the downregulation of CXCR4 (62). Moreover, CXCL12-mediated migration of HCMV-infected cytotrophoblasts was found to be impaired, although CXCR4 surface expression was induced (63). Hence, it is tempting to speculate that HCMV may differentially regulate the CXCR4 axis in a cell-type specific manner. In addition, high expression of CXCL12, recognized by CXCR4, by osteoblasts in the bone marrow could explain why HCMV does not influence the migration toward this chemokine. In particular, HCMV is known to establish latency in CD34 + hematopoietic progenitor cells (39)(40)(41) and CD14 + monocytes (42,43), and therefore, this directed migration of HCMV-infected mDCs to the bone marrow could stimulate dissemination and translocation to an ecological niche rich in potential target cells and thereby increase the reservoir of latently infected myeloid progenitor cells. Considering that HCMV uses multiple strategies to modulate DC migration by, e.g., regulating host chemokine receptor expression as well as expressing a plethora of chemokine and chemokine receptor homologs itself, it emphasizes the crucial role to attack the migratory ability of DCs to hamper early immune responses of the host.
The inside-out-signaling for β2-integrin activation is known to be based on intracellular binding of talin or cytohesin-1 to the CD18 subdomain and to membrane-bound PIP3 (27,33,64). Cytohesin-1 is a specific binding partner for the family of β2integrins and highly expressed in mDCs (64). Consistent with our previous report on HSV-1-infected mDCs (45), cytohesin-1 protein expression was not significantly affected by HCMV infection (Figures 4 and 5). In contrast of directly regulating cytohesin-1 expression, the Kaposi's sarcoma-associated herpesvirus cytohesin-1 homolog kaposin A was found to modulate the subcellular localization of cytohesin-1. Kaposin A directly interacts with cytohesin-1 to induce its membrane localization, finally leading to increased ICAM-1 adhesion (65). However, HCMV does not encode a kaposin A homolog, and thus, alternative mechanisms must be induced. Apart from a regulation of cytohesin-1 itself, we tested whether HCMV influences the expression of CYTIP, which is the direct interaction partner of cytohesin-1 and mediates the abrogation of cytohesin-1 induced LFA-1 activation (30). As recently shown by our group, the silencing of CYTIP indeed leads to induced mDC adhesion and reduced migration, hence CYTIP plays a crucial role in the regulation of adhesion and migration (45). Remarkably, we detected a strong proteasome-dependent degradation of CYTIP in HCMVpositive mDCs (Figure 6). Hence, we provide evidence that at least the prototypic members of αand β-herpesviridae, HSV-1 and HCMV, respectively, share a similar mechanism to induce CYTIP degradation in order to induce adhesion and inhibit mDC migration.
The results presented here reveal that HCMV targets different aspects of DC biology, in particular adhesion and migration. By manipulating chemokine receptor expression and the activation status of integrins, HCMV impairs the migration of infected mDCs to block induction of antiviral immune responses. Further research concerning the molecular mechanism of CYTIP degradation in HCMV-infected mDCs and, thus, a better understanding of this viral immune evasive mechanism might lead to the development of new anti-herpesviral therapeutics.

eThics sTaTeMenT
For the generation of monocyte-derived DCs from leukapheresis products of healthy donors, a positive vote from the local ethics committee has been obtained (reference number 4556). This study was carried out in accordance with the recommendations of the ethics committee of the "Friedrich-Alexander-Universität Erlangen-Nürnberg" with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki.