![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Journal of Virology, January 2008, p. 148-160, Vol. 82, No. 1
0022-538X/08/$08.00+0 doi:10.1128/JVI.01980-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.
Department of Microbiology and Immunology, The University of Melbourne, Victoria 3010, Australia
Received 10 September 2007/ Accepted 10 October 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2β1 and β2 integrins was up-regulated, whereas that of Vβ3, Vβ5, and 5β1 integrins, if present, was down-regulated. This differential regulation of integrins was reflected at the transcriptional level. It was unrelated to the use of integrins as rotavirus receptors, as both integrin-using and integrin-independent viruses induced integrin regulation. Using pharmacological agents that inhibit kinase activity, integrin regulation was shown to be dependent on phosphatidylinositol 3-kinase (PI3K) but independent of the activities of the mitogen-activated protein kinases p38 and ERK1/2, and cyclooxygenase-2. Replication-dependent activation of the PI3K/Akt pathway was observed following infection of intestinal and nonintestinal cell lines. Rotavirus activation of PI3K was important for regulation of 2β1 expression. Blockade of integrin regulation by PI3K inhibition led to decreased adherence of infected intestinal cells to collagen and a concomitant decrease in virus titer. These findings indicate that rotavirus-induced PI3K activation causes regulation of integrin expression in intestinal cells, leading to prolonged adherence of infected cells to collagen and increased virus production. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Integrins provide a means for cells to respond to their environment through intracellular signaling networks controlling crucial cellular processes such as adhesion, proliferation, migration, differentiation, and survival (1, 20, 38, 55, 69). These β heterodimeric glycoproteins have been identified as cellular receptors or entry cofactors for many viruses, including rotaviruses. The 2β1, 4β1, and 4β7 integrins are involved in rotavirus-cell binding and entry, and the Vβ3 and Xβ2 integrins are proposed to facilitate cell entry (17, 28, 29, 32, 34, 46, 47, 73). Of these integrins, 2β1, Vβ3, and Xβ2 facilitate rotavirus attachment and entry into the Caco-2 intestinal epithelial cell line (30).
Two viruses that use β1 and β3 integrins as entry receptors have been shown to modulate cellular integrin expression. The 2β1, Vβ3, and 6β1 integrins are important human cytomegalovirus receptors (23). Human cytomegalovirus infection has been shown to down-regulate 2β1 and 5β1 but up-regulate 6β1 on endothelial cells, down-regulate 1β1 on fibroblasts, and up-regulate β1 integrins on monocytes and prostate tumor cells to facilitate motility (8, 60, 63, 68). Nonpathogenic and pathogenic hantaviruses use β1 and β3 integrins, respectively, for endothelial cell entry. Irrespective of this receptor specificity, hantaviruses induced increased endothelial cell expression of β1 and β3 integrins (27).
Several viruses that use non-integrin receptors can modulate integrin expression to produce an altered cellular adhesion that may impact on virus pathogenesis. Ebola virus surface glycoprotein causes Vβ3 integrin down-regulation, leading to cell detachment and death (62, 65, 67, 72). In contrast, Epstein-Barr virus (EBV) up-regulates V integrins in transformed B lymphocytes, increasing cell proliferation and invasion (37). The EBV-associated up-regulation of 6 integrins might promote nasopharyngeal carcinoma metastasis (54).
In the only study of integrin protein expression modulation by rotavirus reported to date, 5β1 expression on Caco-2 cells was unaltered by SA11 rotavirus at 8 h postinoculation (p.i.) (21). Microarray analysis showed that levels of 2, 6, and β1 integrin subunit mRNA increased 3.1-, 2.7-, and 2.8-fold, respectively, in rhesus monkey rotavirus (RRV)-infected Caco-2 cell cultures at 16 h p.i., compared with levels in mock-infected cells (19). Levels of these integrin mRNAs were unaltered at 6 h or 12 h p.i. The expression of β3 was not up-regulated on the basis of the criteria used, and 5 mRNA levels were not analyzed (19).
Rotavirus infection increases integrin mRNA levels by an unknown mechanism. However, several signaling pathways have been implicated in controlling integrin expression. Leukotriene D4 stimulation of cyclooxygenase-2 (prostaglandin-endoperoxide synthase 2 [COX-2])-dependent prostaglandin production can regulate adhesion and migration of Caco-2 cells by up-regulating 2β1 expression (24, 50, 58). As levels of COX-2 mRNA and prostaglandins are increased after rotavirus infection (19, 71), COX-2 signaling could be involved in rotavirus-induced alterations in integrin mRNA expression. Expression of 2β1 integrin is positively regulated by the AP-1 transcription factor under the control of the mitogen-activated protein kinase (MAPK) ERK1/2 (22, 75). During rotavirus infection, AP-1 activation is influenced by the MAPKs c-Jun NH2-terminal kinase (JNK) and p38, so integrin expression might be regulated by MAPK activation (12, 35, 39). Furthermore, the phosphatidylinositol 3-kinase (PI3K)-dependent increases in β1 integrin expression on intestinal cells following treatment with bacterial endotoxin suggest that PI3K signaling might influence β1 integrin regulation by rotavirus (56).
In this study, we have characterized the regulation of integrin expression following rotavirus infection and the subsequent effects on cell adhesion and virus replication. The 2β1, Vβ3, Vβ5, 5, and β2 integrins were chosen for analysis based on their predominance on rotavirus-permissive intestinal cells (33, 36, 42, 50, 74) and the role of a subset in rotavirus cell entry. It is demonstrated that rotavirus infection of intestinal cells differentially regulates rotavirus receptor and non-receptor integrin mRNA and protein expression by a PI3K-dependent pathway that is activated by infection. The differential regulation of integrins, observed in infected cells only, was independent of integrin receptor binding and dependent on viral replication. The modulation of integrin levels might be important for rotavirus replication, as inhibition of PI3K-dependent 2β1 expression resulted in decreased collagen adhesion by infected cells and reduced infectious virus yield.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2 integrin subunit (12F1-H6, immunoglobulin G2a [IgG2a]; and AK7, IgG1) were purchased from Pharmingen (San Diego, CA) and AbD Serotec (Kidlington, United Kingdom), respectively. Antibodies P4H9-G11 (β2 integrin subunit), LM609 (vβ3 integrin), W6/32 (major histocompatibility complex [MHC] class I), and isotype control MOPC21 (IgG1) were obtained as previously described (29). Antibody P1F6 (Vβ5 integrin, IgG1) was purchased from Chemicon (Temecula, CA). Additional control monoclonal antibodies UPC10 (IgG2a) and FLOPC21 (IgG3) were purchased from Sigma (St. Louis, MO) and matched with test antibody for isotype, diluent type, and protein concentration. Antibody A091 (5 integrin subunit [IgG1]) was obtained by participation of B.S.C. in the Sixth International Workshop on Human Leukocyte Differentiation Antigens (17). Rabbit antibodies to phospho-Akt (Ser473; monoclonal) and β-actin (polyclonal) were purchased from Cell Signaling Technology (Beverly, MA). Rabbit polyclonal antibodies to RRV were produced as described previously (18). Rotavirus antibody-negative rabbit serum was used as a negative control. Pharmacological agents that inhibit the kinase activity of COX-2 (PTPBS), the ERK1/2 kinase MEK (PD98059), p38 (SB203580), and PI3K (LY294002) were purchased from Calbiochem (La Jolla, CA) and dissolved in dimethyl sulfoxide (DMSO) to give stock solutions. Wortmannin, bovine pancreatic insulin, type I collagen from human placenta, and bovine serum albumin (BSA) were obtained from Sigma.
Cell lines and viruses.
The colorectal adenocarcinoma (HT-29) cell line was obtained from the ATCC and grown in Dulbecco's modified Eagles' medium (Thermo Electron, Melbourne, Australia), containing 0.075% (wt/vol) sodium bicarbonate, 20% (vol/vol) heat-inactivated fetal calf serum (FCS; JRH Biosciences, Lenexa, KS), 20 mM HEPES (Boehringer Mannheim, Mannheim, Germany), 26.6 µg/ml gentamicin (Pharmacia, Bentley, WA, Australia), 2 µg/ml amphotericin B (Fungizone: Apothecon, Bristol-Myers Squibb, Princeton, NJ), and 2 mM L-glutamine (Gibco, Grand Island, NY). Caco-2 and MA104 cells were propagated as described previously (47). K562 and 2-K562 cells were maintained and monitored for surface integrin expression as previously described (34).
The origins of RRV (P5B[3],G3), porcine rotavirus CRW-8 (P9[7],G3), and human rotaviruses Wa (P1A[8],G1) and 10/76 (P2A[6],G2) have been described previously (16, 51). Viruses were cultivated in MA104 cells with trypsin, purified by glycerol gradient ultracentrifugation, and titrated by immunofluorescent infectivity assay in MA104 cells as described previously (34, 40).
Psoralen inactivation of rotavirus infectivity. Rotavirus infectivity was inactivated as described previously (31). Inactivated virus showed undetectable infectivity at a dilution of 1:100 and was recognized to the same level at the same dilution as the original purified infectious virus by rabbit antiserum to RRV in an enzyme immunoassay that was described previously (18), showing that virus antigenicity was unaltered by inactivation.
Inhibitor toxicity assay. The potential toxicity of PTPBS, PD98059, SB203580, wortmannin, and LY294002 was tested by examining their effects on cellular metabolic activity using the CellTiter 96 AQueous One solution cell proliferation assay (Promega, Madison, WI). Confluent cell monolayers in 96-well trays were left untreated or treated with inhibitors in serum-free medium for 16 h at 37°C with 5% CO2. CellTiter 96 solution was incubated with cells for 1 h, and the absorbance at 490 nm was recorded with a microplate reader.
Flow cytometric analysis of viral and cellular protein expression. For analysis of viral antigen and total integrin or MHC class I expression, virus infectivity was activated with porcine trypsin (Sigma; 10 µg/ml) at 37°C for 20 min and added at the indicated multiplicity of infection (MOI), in a total volume of 200 µl of serum-free medium, to confluent cell monolayers in 24-well trays (Nunc, Wiesbaden, Germany). Following virus adsorption for 1 h, the inoculum was removed and replaced by serum-free medium. When utilized, inhibitors or DMSO was present at the given concentrations in both the viral inoculum and the medium added to cells following inoculum removal, unless otherwise indicated. Inhibitor presence during inoculation did not affect infectivity (data not shown). At the indicated times p.i., cell monolayers were washed twice with phosphate-buffered saline (PBS) and detached by incubation at 37°C for 5 min in PBS containing 0.1% (wt/vol) trypsin (Difco Laboratories, Detroit, MI) and 0.02% (wt/vol) EDTA (PBS-trypsin-EDTA). Cells were permeabilized by methanol fixation for 7 min at –20°C. For analysis of each integrin or MHC class I, duplicate aliquots of 3 x 104 cells were reacted for 45 min on ice with the specified test or control mouse antibody diluted to 20 µg/ml in PBS-FCS containing 0.1% (vol/vol) NaN3 (PBS-FCS-Az). Washed cells were reacted as described above with allophycocyanin-conjugated goat anti-mouse IgG (BD Pharmingen, San Diego, CA) diluted 1:50 in PBS-FCS. For virus antigen detection, washed cells were reacted with rabbit polyclonal antibodies to RRV or control polyclonal antibodies diluted 1:500, followed by fluorescein isothiocyanate-conjugated goat anti-rabbit IgG (Chemicon) at a 1:50 dilution. Antibodies to rotavirus and cellular proteins were combined for dual staining of virus antigen and each protein. Cellular fluorescence was analyzed on a FACSort flow cytometer with CellQuest software (Becton Dickinson, Cockeysville, MD). Infected or uninfected populations were gated on the basis of virus antigen expression.
The effect of insulin on integrin expression was determined by treatment of cell monolayers with insulin (10 µg/ml) for 4 h. Treated cells were incubated in fresh medium for a further 12 h before flow cytometric analysis as described above.
Cell surface protein expression was analyzed by flow cytometry as described previously (47). Briefly, cells were infected, detached, stained with mouse antibodies, fixed with 1% (vol/vol) ultrapure formaldehyde (Polysciences, Warrington, PA) in PBS, and analyzed by flow cytometry as described above.
Cellular RNA extraction and quantitative real-time PCR analysis.
Confluent cell monolayers in 24-well trays were infected with rotavirus at an MOI of 20 or mock infected with harvests of uninfected MA104 cells that had been trypsin treated and diluted as for virus-infected cells. In some experiments, LY294002 or DMSO was included at the indicated concentrations in the virus inoculum and medium added to cells after inoculum removal. At 16 h p.i., total RNA from 6 x 106 cells was collected using the RNeasy Midi kit (QIAGEN, Hilden, Germany) according to the manufacturer's protocol. Relative levels of 2, 5, V, β1, β3, and COX-2 mRNA in mock-, RRV-, and CRW-8-infected cells were determined by using real-time PCR. Specific forward and reverse primer sets were designed using Primer Express software (version 1.5; Applied Biosystems, Foster City, CA) and purchased from Geneworks (Hindmarsh, SA, Australia) as listed: 2 (forward, 5'-TTGAAGCCTATTCTGAGACTGC-3'; reverse, 5'-CTTGAGCAGCTGGTATTTGTCG-3'), 5 (forward, 5'-ATCTCAACAACTCGCAAAGCG-3'; reverse, 5'-TTACTGGGAATAGCACTGC-3'), V (forward, 5'-TCTTCTCTCGGGACTCCTGC-3'; reverse, 5'-CCGAAGTAACTTCCCTCG-3'), β1 (forward, 5'-TGCCATCATGCAAGTTGC-3'; reverse, 5'-ACCAAGTTTCCCATCTCCAGC-3'), β3 (forward, 5'-ATGCATCCCACTTGC-3'; reverse, 5'-ACACTGCCCGTCATTAG-3'), COX-2 (forward, 5'-TCAGCCATACAGCAAATCC-3'; reverse 5'-TCCTGTCCGGGTACAATCG-3') and 18S rRNA (forward, 5'-CGGCTACCACATCCAAGGAA-3'; reverse, 5'-GCTGGAATTACCGCGGCT-3'). The primer concentrations and dynamic range of RNA were optimized for each real-time PCR target. For each reaction, diluted target mRNA (5 µl) was added to 20 µl of a mix containing Multiscribe reverse transcriptase (Applied Biosystems), forward and reverse primers (300 or 900 µM), RNase-free water, RNase inhibitor, and 12.5 µl of 2x SYBR green PCR master mix (Applied Biosystems), which contained nucleotides, AmpliTaq Gold DNA polymerase, and optimized buffer components. Real-time PCRs were performed on an ABI Prism 7700 sequence detector (Applied Biosystems) with an amplification profile of 30 min at 48°C, 10 min at 95°C, and 45 cycles of 15 s at 95°C and 1 min at 60°C. As a no-template control, diluted target mRNA was replaced with RNase-free water. All samples and standards were tested in triplicate in at least three independent experiments. The data were analyzed with the sequence detector software V1.7 (Applied Biosystems). Transcript abundance was calculated using the comparative change in threshold cycle (
CT) method with 18S rRNA as the normalization standard. Levels of 18S rRNA were unaltered by RRV and CRW-8 infection of Caco-2 cells (data not shown). Differential expression based on real-time PCR measurements was defined as a change in transcriptional accumulation of 
1.5. The specificity of amplification of each assay was verified by electrophoretic separation of PCR products on agarose gels and dissociation curves generated by slow denaturation of the PCR end products that were analyzed using ABI Prism 7700 dissociation curve analysis software V1.0 (Applied Biosystems).
Detection of Akt activity. Confluent cell monolayers in 24-well trays were serum starved for 24 h and then mock infected, infected with purified rotavirus (MOI of 10), or treated with insulin (10 µg/ml). At given times, cells were lysed in 50 mM Tris-HCl buffer (pH 8.0), containing 1 mM Na3VO4, 2 mM phenylmethylsulfonyl fluoride (Sigma), aprotinin (10 µg/ml; Sigma), and leupeptin (5 µg/ml; Sigma) for 5 min at 4°C. Lysates were clarified by centrifugation and boiled with sample buffer. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrophoretically transferred to a membrane as described previously (35). After blocking in PBS containing 2% (wt/vol) BSA and 0.1% (vol/vol) Tween 20 (PBS-BSA) for 1 h at 37°C, the membrane was washed with PBS containing 0.1% (vol/vol) Tween 20 (PBS-T) and reacted with antibody to phospho-Akt diluted in PBS-BSA for 18 h at 4°C. After washing in PBS-T, membranes were washed in PBS-T and reacted with horseradish peroxidase-conjugated sheep anti-rabbit antibody (Chemicon), optimally diluted in PBS-T, for 1.5 h at room temperature. Bound antibody was detected by enhanced chemiluminescence (ECL) with ECL Hyperfilm (Amersham, Arlington Heights, IL) as specified by the manufacturer. Membranes were stripped as previously described (35) and reprobed as described above with antibody to β-actin diluted in PBS-BSA. Films were scanned, and densitometric analysis was carried out using Image Gauge software (V3.3, Fujifilm, Tokyo, Japan).
Cell adhesion assay.
Plates (96 wells) were coated (100 µl) with collagen or BSA (10 µg/ml) in PBS overnight at 4°C. Unbound protein was removed by washing twice with PBS. Further incubations were carried out at 37°C in 95% (vol/vol) air-5% (vol/vol) CO2. Wells were blocked with 1% (wt/vol) BSA in PBS for 1 h, followed by three washes with PBS. Confluent cell monolayers in 24-well trays were infected with RRV or CRW-8 at an MOI of 10 in medium containing 20 µM PI3K inhibitor (LY294002) or matched DMSO control. At 16 h p.i., cells were detached with PBS-trypsin-EDTA, washed, and resuspended in serum-free medium containing 2% (wt/vol) BSA to block nonspecific binding. For adhesion blockade, 1.5 x 104 cells in 300 µl were reacted with anti-2 or control antibody at 20 µg/ml and incubated with rolling for 30 min. Cells (5.0 x 103 per well) were allowed to adhere in triplicate to coated wells for 1 h. Unbound cells were removed by three gentle washes with PBS. Following medium replacement, cells were allowed to spread for a further 4 h. Adherent cells were counted under a light microscope. Bound cells (round) were distinguished from spreading cells (extended with protrusions) by morphology and enumerated separately.
Analysis of cell detachment during infection. Confluent cell monolayers, cultured in 24-well trays coated (200 µl) with collagen as described above or untreated, were infected at a MOI of 0.1 in the presence of PI3K inhibitor LY294002 at 20 µM or a DMSO-matched control. Additional LY294002 at 20 µM was added at 24 h p.i. to replace any lost activity. The number of cells detached at the given times after infection was determined by counting in a hemocytometer under a light-phase microscope.
Rotavirus growth assay. Rotavirus growth assays in adherent cell lines were performed as previously described (47). Briefly, cell monolayers growing in 24-well trays that had been either untreated or pretreated with collagen, as described above, were washed twice with medium. Trypsin-activated virus (MOI of 0.1) was adsorbed to cells for 1 h at 37°C. Monolayers were washed twice, medium containing 1 µg/ml porcine trypsin was added, and cultures were incubated for the indicated times. Virus replication was terminated by freezing at –70°C, and infectious virus was harvested by two rounds of freezing and thawing. Infectious virus was quantitated by infectivity titration in MA104 cells.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2β1, β2, Vβ3, and 5β1 integrins and MHC class I was determined using dual-color flow cytometric analysis of fixed cells stained for rotavirus antigen and each cellular protein. As 2 and 5 subunits pair with β1 only, their expression reflects 2β1 and 5β1 levels. The β2 subunit pairs with L, M, X, and D, so β2 expression indicates any β2 heterodimer present (38). The anti-Vβ3 antibody used is specific for this heterodimer (11). Initially, the proportions of cells infected with rotaviruses at a range of MOI were determined from density plots, such as those illustrated for RRV at an MOI of 10 in Fig. 1. As shown in Table 1, intestinal cell lines Caco-2 and HT-29 were slightly less susceptible to rotavirus infection than the kidney cell line MA104. At MOI of 3.5 and 10, 93% to 95% of MA104 cells were infected at 16 h p.i. by RRV and CRW-8. In contrast, maximum levels of infection of Caco-2 and HT-29 cells by these rotaviruses occurred at an MOI of 10 (77% to 83%). Human rotaviruses Wa and 10/76 infected fewer Caco-2 cells (69%) than RRV or CRW-8 (77% to 82%).
|
|
2 and β2 integrins and down-regulation of Vβ3 and 5 integrins. Similar alterations in 2 and β2 occurred in RRV-infected cells within RRV-inoculated HT-29 cultures. However, HT-29 cells did not express Vβ3 or 5 integrins, in agreement with previous reports (11, 59). No shift in the Vβ3 and 5 histograms of HT-29 cells was detected, so their expression was not induced de novo by infection. In contrast to the intestinal cell lines, integrin expression was unaltered in the infected cells within RRV-inoculated MA104 cultures. MHC class I expression was unaltered in RRV-infected Caco-2, HT-29, and MA104 cells.
These studies were extended to additional rotaviruses, cell lines, and MOI (Table 1). Compared with mock-infected cells, inoculation of Caco-2 cells with RRV, CRW-8, Wa, or 10/76 rotavirus (MOI of 10) produced increased expression by infected cells of 2 (2.1- to 2.4-fold) and β2 (1.6- to 1.8-fold) and decreased expression of Vβ3 (0.7-fold) and 5 (0.8-fold). The change (fold) in integrin expression by Caco-2 cells expressing RRV or CRW-8 antigen did not vary with MOI over the range of 0.1 to 10, although only 9 to 10% of cells were infected at an MOI of 0.1 (Table 1). RRV or CRW-8 infection of HT-29 cells resulted in increased expression of 2 (2.1-fold) and β2 (1.4- and 1.5-fold, respectively) at 16 h p.i. over mock-infected cells. These increases were proportionally the same as in Caco-2 cells, even though HT-29 cells expressed five- to ninefold less 2 and β2 than Caco-2 cells. Irrespective of MOI, total integrin levels in uninfected Caco-2 and HT-29 cells exposed to RRV, CRW-8, Wa, or 10/76 were similar to those in mock-infected cells (Fig. 1 and Table 1). Thus, the differential expression of integrins was not caused by a soluble factor secreted from infected cells. In contrast to integrins, total MHC class I expression was unaltered by rotavirus infection in Caco-2 and HT-29 cells. This highlights the specificity of integrin regulation by infected intestinal cells (Fig. 1 and Table 1). RRV or CRW-8 infection of MA104 cells did not affect expression of 2, β2, Vβ3, 5, or MHC class I (Fig. 1 and Table 1). Also, 2 levels in K562 or 2-K562 cells were unaltered by RRV infection (Table 1). Overall, rotavirus inoculation of intestinal cells led to differential regulation of 2β1, β2, Vβ3, and 5β1 expression in infected cells within inoculated cultures, provided the integrin was initially present. This regulation was independent of virus usage of 2β1, β2, and Vβ3 for cell attachment and entry, as CRW-8 and 10/76 do not use these integrins but regulated integrin expression similarly to the integrin-using rotaviruses RRV and Wa. These studies also show that Vβ3 is not essential for cell entry by integrin-using rotaviruses, as HT-29 cells lacked Vβ3 but were infected by RRV similarly to Caco-2 cells (Table 1).
The kinetics of integrin expression regulation by rotaviruses were examined in Caco-2 cells (Fig. 2). Cellular integrin levels in RRV-infected cells first altered between 8 and 12 h p.i. with maximum regulation at 16 h p.i. The unchanged 5 expression at 8 h p.i. observed in this study confirms the previous report that 5 is unaltered at this time (21). The kinetics of integrin regulation in Caco-2 cells infected with CRW-8 or RRV were indistinguishable (data not shown). Caco-2 cell exposure to inactivated RRV did not alter cellular integrin expression (Fig. 2). This indicates that rotavirus replication was required for differential regulation of integrin expression.
|
2, β2, Vβ3, and 5 expression in parallel with fixed, permeable cells (Fig. 3). As rotavirus antigen could not be analyzed in unfixed cells, infected cells were not distinguished from uninfected cells in these experiments. However, the majority (77 to 82%) of cells were infected at the MOI (10) chosen (Table 1). As shown in Fig. 3, RRV infection produced similar alterations in total levels of 2, β2, Vβ3, and 5 expression to earlier experiments (Fig. 1 and 2 and Table 1). Surface expression of these integrins on RRV-infected Caco-2 cells was modulated to a similar extent to the total cellular level (Fig. 3A and B). CRW-8 infection of Caco-2 cells at the same MOI similarly altered total and cell surface expression of 2, β2, Vβ3, and 5 (data not shown). The levels of cell surface and total MHC class I expression were unaltered following Caco-2 cell infection with RRV or CRW-8 (Fig. 3A) (data not shown).
|
Vβ5 integrin expression were examined using a monoclonal antibody specific for this heterodimer. Following RRV infection, Vβ5 levels in Caco-2 and HT-29 cells were reduced by 43% and 52%, respectively (Fig. 4A and C). Inhibitors of p38, COX-2, and MEK did not affect integrin protein expression in RRV- or CRW-8-infected Caco-2 cells compared with that in infected cells treated with DMSO alone (Fig. 4A and B). In contrast, PI3K inhibitors LY294002 and wortmannin extensively blocked the differential regulation of integrins caused by rotavirus in Caco-2 and HT-29 cells (Fig. 4). RRV or CRW-8 infection of Caco-2 cells in the presence of these inhibitors blocked 2 and β2 up-regulation by 71 to 90% and Vβ3, Vβ5, and 5 down-regulation by 44 to 99% compared with control cells (Fig. 4A and B). Similarly, 2 and β2 up-regulation following RRV or CRW-8 infection of HT-29 cells was reduced by 60 to 83% in the presence of PI3K inhibitors and Vβ5 down-regulation after RRV infection was reduced by 79% (Fig. 4C). For Caco-2 and HT-29 cells, LY294002 inclusion during the first 5 h after RRV infection was almost as effective in restoring 2 to control levels as LY294002 treatment for 16 h, whereas LY294002 treatment for 5 to 16 h after infection had little effect on 2 up-regulation (Fig. 4A and C). Thus, PI3K signaling early in the rotavirus replicative cycle was required for regulation of 2 expression. MHC class I expression following rotavirus infection was unchanged by treatment with PI3K, p38, COX-2, and MEK inhibitors (Fig. 4A, B, and C). This demonstrates the specificity of PI3K inhibition for blockade of rotavirus-induced integrin regulation. Overall, these data show that integrin expression modulation after rotavirus infection of intestinal cells required PI3K signaling in the first 5 h after infection and was independent of p38, COX-2, and MEK kinases.
|
2 mRNA levels, 1.9-fold (RRV) and 2.0-fold (CRW-8) in β1 mRNA levels, and 2.5-fold (RRV) and 2.6-fold (CRW-8) in COX-2 mRNA levels. In contrast, V mRNA levels were reduced to 0.6-fold (RRV) and 0.5-fold (CRW-8) and 5 mRNA levels to 0.4-fold (RRV) and 0.6-fold (CRW-8). Levels of β3 mRNA in infected cells were unaltered compared with those in mock-infected cells, and β5 was not analyzed. Due to primer-dimer formation, β2 mRNA levels could not be determined using two alternative primer sets. A useable dynamic range could not be obtained in the Taqman gene expression assay (Applied Biosystems), again precluding accurate measurement of β2 mRNA. RRV infection in HT-29 cells produced a 7.1-fold increase in 2 mRNA levels, similar to Caco-2 cells. Inactivated RRV did not affect 2 mRNA levels in either cell line. PI3K inhibitor LY294002 reduced the elevated 2 mRNA levels in Caco-2 and HT-29 cells by 80 to 81% (Fig. 5). These data indicate that PI3K-dependent changes in integrin protein expression following rotavirus infection were reflected at the transcriptional level.
|
|
|
2β1 integrin, a key ligand for adhesion of these intestinal cells to type 1 collagen (5, 33). The functional significance of the increased 2β1 expression was therefore examined by quantitation of collagen adhesion by mock- and rotavirus-infected Caco-2, HT-29, and MA104 cells. The role of 2β1 in this adhesion was analyzed by anti-2 antibody blockade (Fig. 8). This antibody reduced collagen adherence by 29% (Caco-2), 48% (HT-29), and 53% (MA104) for mock-infected cells and 38% (Caco-2), 57% (HT-29), and 39% (MA104) for RRV-infected cells. Adhesion of uninfected Caco-2 and HT-29 cells to type 1 collagen was similarly reduced by anti-2 antibodies in previous studies (33, 36, 50). Our experiments show this function is retained in rotavirus-infected cells. Although fewer HT-29 (26%) than Caco-2 (70%) cells adhered to collagen, HT-29 adhesion showed greater dependence on 2β1. As increased 2β1 expression following rotavirus infection was dependent on PI3K activity, the ability of the PI3K inhibitor LY294002 to affect cellular adhesion to collagen also was determined. This inhibitor had no effect on collagen adherence by mock-infected Caco-2, HT-29, or MA104 cells (Fig. 8). Infection reduced adherent Caco-2 cell numbers by 24% compared with mock-infected cells. However, significantly greater reductions of 57% (RRV) and 52% (CRW-8) in adherent Caco-2 cell numbers were observed following infection in the presence of LY294002 compared to DMSO alone (P < 0.0047). Consistent with their greater dependence on 2β1 for collagen adhesion, there were 42% (RRV) and 41% (CRW-8) fewer adherent virus-infected HT-29 cells than mock-infected cells. LY294002 reduced adherence of virus-infected HT-29 cells to collagen by 74% (RRV) and 67% (CRW-8), significantly greater reductions in cellular adherence than those induced by infection alone (P < 0.0038). Thus, PI3K inhibition in Caco-2 and HT-29 cells during infection decreased collagen adhesion. As PI3K inhibition blocked the majority of 2β1 expression up-regulation in rotavirus-infected cells and 2β1 mediated 38 to 57% of collagen adhesion, a substantial part of the reduced collagen adherence after LY294002 treatment was related to lowered 2β1 levels.
|
2β1 up-regulation through PI3K activation was not observed in MA104 cells, the lack of effect of PI3K inhibition in these cells is further evidence that 2β1 up-regulation was responsible for the increased binding of rotavirus-infected intestinal cell lines to collagen. Various proportions of adherent cells spread on collagen, demonstrating their viability. Rotavirus-infected Caco-2, HT-29, and MA104 cell cultures showed a marked decrease in the number and proportion of attached cells that were spreading compared with mock-infected cells. The proportion of spreading Caco-2 and HT-29 cells following RRV and CRW-8 infection in the absence of LY294002 (14 to 41%; mean ± standard deviation [SD] of 27% ± 10%) was similar to the uninfected proportion (17 to 23%; mean ± SD of 20% ± 3%; Table 1). Additionally, bound infected cell numbers fell by 49% to 52% (Caco-2) and 77% to 88% (HT-29) in the presence of LY294002, but spreading cell numbers were unaltered by LY294002 treatment. Thus, most cells spreading on collagen within rotavirus-infected HT-29 and Caco-2 cultures probably were uninfected. The almost complete absence of spreading MA104 cells after rotavirus exposure likely results from the high proportion of infected cells (93 to 95%; Table 1).
The effect of PI3K-dependent increases in 2β1 expression and adherence to collagen on cell survival and virus replication was investigated. Initially, the proportion of HT-29 and Caco-2 cells grown on collagen that remained adherent after rotavirus infection in the presence or absence of LY294002 was determined (Fig. 9A and B). At 40 h p.i. with RRV or CRW-8, adherent cell numbers fell by 22 to 28% (HT-29; P = 0.01) and 7% (Caco-2; P = 0.01) in the presence of LY294002, compared with the DMSO control (Fig. 9A and B). This indicated that a proportion of HT-29 and Caco-2 cell attachment to collagen during infection was dependent on PI3K activity, consistent with other findings (Fig. 8). Titers of RRV and CRW-8 at 40 h p.i. in these HT-29 cells grown on collagen were reduced by 37% and 27%, respectively, in the presence of LY294002 (Fig. 9C and D; P < 0.008). RRV titers at 40 h p.i. in these Caco-2 cells were reduced by 18% (Fig. 9C; P = 0.002). Both RRV yield and collagen adhesion were reduced to a greater extent in HT-29 than Caco-2 cells. Overall, through 2β1 integrin up-regulation, cellular PI3K activity played a significant role in maintaining cellular adherence (and therefore viability) and the yield of infectious rotavirus.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2β1 and reduced virus replication.
Up-regulation of 2β1 and β2 was evident by 12 h p.i., while Vβ3, Vβ5, and 5β1 were down-regulated at this time. This regulation was observed only in Caco-2 and HT-29 cells, suggesting an intestinal cell-specific mechanism. Changes in integrin expression were not due to a general cellular response to loss of cell viability during rotavirus infection, as Caco-2 cells had lost little viability and metabolic activity at 16 h p.i. when maximum integrin regulation was observed. In contrast, MA104 cells at 16 h p.i. showed significant loss of viability and metabolic activity with no regulation of integrin expression (P. Halasz and B. S. Coulson, unpublished data). The changes in integrin protein levels occurred both within the infected cell and at the cell surface, which is indicative of modulation of protein expression rather than trafficking pathways. Changes in integrin protein levels also were generally reflected at the mRNA level, identifying transcriptional regulation as a key mechanism of modulation of integrin levels during rotavirus infection. An exception was β3 mRNA, which was unaltered by infection, even though V mRNA expression was reduced. It is possible that IIb (CD41), the other β3 partner, might be up-regulated by rotavirus infection or V but not β3 may be targeted for regulation. The latter is supported by the down-regulation of Vβ5. The rotavirus-induced increases in 2 and β1 mRNA levels confirm the previous microarray finding of increases in mRNA expression of these integrin subunits in Caco-2 cells under similar conditions (19).
Many integrin-using and integrin-independent rotaviruses replicate to high titer in both Caco-2 and HT-29 cells (43, 66). It is noteworthy that Vβ3 is absent from HT-29 cells, indicating that Vβ3 is not necessary for RRV entry. Taken together, these data show that of the proposed integrin receptors for rotaviruses, 2β1 and Xβ2 are sufficient to mediate efficient attachment and entry into epithelial cells by integrin-using rotaviruses.
The extent of regulation of 2β1, β2, Vβ3, Vβ5, and 5β1 integrin expression following infection of Caco-2 cells by integrin-using rotaviruses RRV and Wa was indistinguishable from that of integrin-independent rotaviruses CRW-8 and 10/76. This demonstrates that the altered integrin expression induced following rotavirus infection was unrelated mechanistically to rotavirus usage of 2β1, Xβ2, and Vβ3 during cell attachment and entry. Rotavirus infection does not alter 2β1, β2, Vβ3, and 5β1 expression in surrounding uninfected cells, as uninfected cells within virus-infected cell cultures showed similar levels of these integrins to mock-infected cultures. Thus, virus spread to uninfected cells is not enhanced through integrin receptor up-regulation. These findings also indicate that the primary function of integrin regulation by rotaviruses is not to facilitate virus superinfection of the infected cells or a cellular defense mechanism to block virus release (19). However, in the case of the integrin-using rotaviruses, these functions might occur as secondary consequences of the integrin regulation. As altered Caco-2 cell expression of all integrins first occurred at the same time (8 h to 12 h) after infection and the levels of regulation of all integrins depended to similar extents on PI3K, it is likely that the mechanisms and signaling pathways involved in the regulation of each integrin are linked. Most trypsin-activated (infectious) RRV is internalized by MA104 cells within 3 to 5 min (41). The much later time after infection at which integrin expression was altered in our study suggests a requirement for rotavirus replication rather than signaling events induced following receptor engagement or entry, although these cannot be formally ruled out. This argument is further supported by the demonstration that Caco-2 cell exposure to inactivated RRV, which can enter cells but not replicate (61), does not induce integrin regulation.
PI3K inhibitors blocked rotavirus-induced regulation of integrin mRNA and protein expression in intestinal cells to a large extent, clearly identifying a specific role for PI3K. Involvement of the COX-2, p38, and MEK pathways was ruled out by the lack of effect of their inhibitors on rotavirus-induced modulation of integrin expression. The up-regulation of 3β1 integrin on enterocytes by bacterial endotoxin also depends on PI3K activation (56). Interestingly, induction of PI3K activity in Caco-2 cells with insulin did not alter integrin expression. Thus, although the PI3K pathway is crucial for rotavirus-induced changes in integrin expression, other rotavirus-dependent factors are also required.
Papillomaviruses and human immunodeficiency virus type 1 reduce MHC class I levels on infected cells to escape detection and killing by cytotoxic T lymphocytes (2, 14). MHC class I expression was unaltered by rotavirus infection, showing that this immune evasion strategy is not used by rotaviruses. Although both rotaviruses and Ebola virus glycoprotein induce vβ3 down-regulation, only Ebola virus also led to MHC class I down-regulation (62, 65, 67, 72). This suggests that these viruses reduce Vβ3 expression by different mechanisms, consistent with the rotavirus requirement for PI3K activity and the Ebola virus glycoprotein dependence on reduced activation of the MAPK ERK2 (72).
Rotavirus activated Akt in a PI3K-dependent process at 1 to 4 h after infection. PI3K activity inhibition during this time almost completely abolished 2β1 up-regulation in intestinal cells, whereas PI3K inhibition after this time had little effect. These findings demonstrate that rotavirus-induced PI3K activation is crucial for regulation of integrin expression. As integrin levels were unchanged in rotavirus-infected MA104 cells despite Akt activation, factors present only in intestinal cells appear to be required in addition to PI3K activity for rotavirus-induced integrin regulation. The 4- to 8-h delay between rotavirus-induced Akt activation and integrin protein regulation might reflect a more complex cellular process, rather than a direct effect of integrin transcriptional activation through the PI3K pathway. For example, integrin transcriptional regulation might depend on prior PI3K-dependent expression of particular proteins.
Virus activation of PI3K can occur through integrin recognition. Adenovirus stimulates PI3K activation within 5 min of virus binding to V integrin receptors, which aids virion internalization (45). Papillomavirus-like particles induce PI3K activation by 10 min after 6β4 integrin engagement (25). Our work shows that rotavirus-induced PI3K activation is unlikely to occur via integrin receptor binding as the integrin-independent CRW-8 strain induced PI3K activation to a similar extent to RRV. Also, rotavirus-induced PI3K activation begins later than would be expected for receptor-induced activation. Although the timing could be consistent with activation during cell entry, this is unlikely. Inactivated rotavirus neither activated PI3K nor altered integrin expression, despite its reported ability to bind sialic acids and penetrate the cell membrane (61). Rotavirus replication transiently induces heat shock protein 70 (Hsp70) at 2 to 7 h p.i. in Caco-2 but not MA104 cells (10, 70). Although the Hsp70 response can involve PI3K activation (57), PI3K involvement in Hsp70 induction by rotavirus seems unlikely as we found that Akt was phosphorylated during infection of both MA104 and Caco-2 cells. PI3K activation is utilized to decrease infection-induced apoptosis by influenza A virus, hepatitis C virus, flaviviruses, rubella virus, and EBV (15). The antiapoptotic effect of rotavirus-induced PI3K activation is worthy of further investigation.
Inhibition of PI3K activity reduced the 2β1-dependent adhesion of rotavirus-infected intestinal cells to collagen, consistent with their lower 2β1 expression under these conditions. In contrast, PI3K inhibition did not affect the number of rotavirus-infected MA104 cells bound mainly through 2β1 to collagen, in line with their unaltered 2β1 expression. These findings support the argument that rotavirus-induced 2β1 up-regulation is responsible for the increased collagen binding by rotavirus-infected intestinal cells. Infected intestinal cells adhered to collagen for longer times in the absence of PI3K inhibition. The extent of this adhesion increase was related to the rotavirus infectious yield, strongly suggesting that these are linked. During multiple rounds of infection, this could provide a distinct replicative advantage. In the human intestine, the major ligand for 2β1 is collagen (13). Therefore, it is likely that any rotavirus-induced up-regulation of enterocyte 2β1 expression in vivo would similarly extend cell adhesion and survival before infected cells are ultimately lost (9). In Caco-2 cells, Vβ3 down-regulation increases resistance to apoptosis, so the rotavirus-induced Vβ3 down-regulation we observed in these cells might be an additional mechanism that could prolong their viability (44).
From the studies presented here, PI3K-dependent integrin regulation appears to be used by rotavirus to increase the attachment and survival of infected intestinal cells and facilitate virus replication. These findings provide novel insights into rotavirus replicative strategies, the cell signaling cascades important for rotavirus replication, and the effects of rotavirus on intestinal cell interactions with the extracellular environment.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by project grants (208900 and 350252) and research fellowships (299861 and 350253) awarded to B.S.C. from the National Health and Medical Research Council of Australia (NHMRC). S.J.T. holds an NHMRC RD Wright Fellowship (359207).
|
FOOTNOTES |
|---|
Published ahead of print on 17 October 2007.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
