Effect of Vectors on Human Endothelial Cell Signal Transduction

【摘要】  Objective- Endothelium is an important target for gene therapy. We have investigated the effect of viral and nonviral vectors on the phenotype and function of endothelial cells (ECs) and developed methods to block any activation caused by these vectors.

Methods and Results- Transduction of ECs with viral vectors, including adenovirus, lentiviruses, and Moloney murine leukemia virus, can induce a pro-inflammatory phenotype. This activation was reduced when nonviral vectors were used. We demonstrate that after transduction there is upregulation of dsRNA-triggered antiviral and PI3K/Akt signaling pathway. Blockade of the NF B, PI3-K, or PKR signaling pathways all operated to inhibit partially virally induced activation, and inhibition of both PKR and PI3-K pathways totally blocked EC activation. Furthermore, inhibition of IFN- /ß in addition to PI3-K was effective at preventing EC activation.

Conclusions- Viral vectors, although efficient at transducing ECs, result in their activation. Blockade of the signaling pathways involved in viral activation may be used to prevent such activation.

Viral vectors are efficient at transducing ECs. However, they activate the cells, causing an increased expression of adhesion molecules and cytokines. This is caused by stimulation of intracellular signaling pathways. Blocking these pathways, or cytokines that provide autocrine stimulation, can prevent activation after transduction.

【关键词】  endothelial cells gene therapy nonviral vectors signal transduction viral vectors

Introduction

The endothelium is an attractive target for gene therapy because of its accessibility and importance in the pathophysiology of a wide range of conditions. Gene delivery to endothelial cells (ECs) can be accomplished using viral or nonviral vectors. In general, viral methods 1-3 are effective at delivering genes to ECs; however, viruses can stimulate ECs. 4-7 Viral vectors can activate the dsRNA-triggered 2'-5' OA synthetase/RNase L, PKR, 8-10 and PI3 kinase (PI3-K) 8,9,11 pathways, resulting in alterations in surface phenotype and secretion of cytokines, including IFN /ß. 8 In contrast, nonviral vectors are less efficient at gene delivery to ECs, 12-16 but are less likely to activate these cells. 16

In this report we investigate the response of ECs to gene delivery mediated by viral and nonviral vectors. We show that viral vectors can activate ECs, leading to an increased expression of adhesion molecules and cytokines, but that this can be blocked by inhibition of a variety of signaling pathways and cytokines. This opens up novel therapeutic interventions aimed at modulating the effect of viral vectors on ECs.

Materials and Methods

EC Isolation and Culture

Human saphenous veins ECs (HSVECs) and foreskin microvascular endothelial cells were prepared and cultured as described. 17,18 For all experiments, the ECs were pooled from at least 5 patients as previously described. 17

Transfection and Transduction

The E1a-, partial E1b-, partial E3-adenovirus serotype 5 (Ad) vector, Ad-EGFP (Clontech, Palo Alto, Calif), Ad-0, RAd35, the HIV-EGFP/-0 and EIAV-EGFP/-0 (kind gifts from Oxford Biomedica, Oxford, UK), and the MMLV (pBullet-EGFP) were produced as described. 8 For transduction, 10 4 cells were incubated with virus in 100 µL optiMEM I for 2 to 3 hours, at which time the volume was increased to 0.5 mL by addition of culture medium. Immunolipoplexes were used to transfect cells as described. 1,16,19

Flow Cytometry

The phenotype of ECs was assessed by flow cytometry 2 days after transfection with nonviral vectors or 3 days after transduction with viral vectors, because it was at these times that maximal gene expression was seen. Cell staining was performed using mouse antibodies (Ab) conjugated with allophycocyanin (activated protein C ) or primary Ab, followed by goat anti-mouse-APC. 17,20 When indicated, ECs were stimulated with 80 ng/mL tumor necrosis factor (TNF)-, 80 ng/mL interleukin (IL)-1ß, and 80 ng/mL IFN-, followed by flow cytometric analysis 12 hours after stimulation. 16,20

Cytokines

IFN-, IL-12 p70, IL-1ß, IL-6, IL-8 and TNF- were measured as described. 8 Supernatants were obtained 4 days after transduction.

Reverse-Transcription Polymerase Chain Reaction Assay

After transduction or transfection, the ECs were isolated with magnetic beads coated with mAb against CD105 (Caltag, Silverstone, UK). Reverse-transcription polymerase chain reaction assays were performed using the paired primers for 2'-5' OA synthetase-1, IRF-1, IFN- 2, IFN-ß1, and ß-actin. 8,16

SDS-PAGE and Western Blotting

Cell lysates were analyzed by Western blot as described. 17,20 Antibodies used are listed in supplementary data.

Leukocyte Adhesion and Transmigration Assays

Adhesion and transmigration were performed as previously described. 17

Blockade of Signaling Pathways

The following inhibitors were used to block specific pathways: NF B pathway with PSI (10 µmol/L), MG132 (10 µmol/L), or SN50 (20 µmol/L) (all from Calbiochem), the PKR pathway with 2-amino-purine (Sigma) (10 mmol/L), MAPK p38 with SB23906 (10 µmol/L) (Tocris Cookson Ltd, Bristol, UK), JNK with SP600125 (10 µmol/L) (Calbiochem), STAT3 with STAT3 inhibitor peptide (3 mmol/L) (Calbiochem), PI3 K/Akt pathway with LY294002 (50 µmol/L) (Sigma), or Wortmannin (100 nM) (Sigma). Trichodion (Alexis Platform, Nottingham, UK) was used to inhibit the NF B, AP-1 and STAT-1 pathways (80 µmol/L), and Dexamethasone (Sigma) (5 x 10 -7 M) was used to inhibit several pathways. In all cases, ECs were cultured in the inhibitor for 24 hours before addition of viral particles. To block cytokines, ECs were cultured in the presence of mAbs to IL-1, IL-1ß, IFN-, IFN-ß and/or TNF- (20 µg/mL) (R&D systems).

Statistical Analysis

The Student t test was used for simple comparison between 2 means. For multiple comparisons, data were analyzed by ANOVA.

Results

Comparison of Efficiency of Viral and Nonviral Vectors at Transducing ECs

We first tested the ability of viral vectors to transduce primary HSVECs. We optimized transduction efficiency of all these viral vectors (Figure I, available online at http://atvb.ahajournals.org). As previously reported, 2,7,22 adenoviral and lentiviral vectors are efficient at transducing HSVECs. However, although MMLV vectors can transduce HUVECs, 23 we found that MMLV resulted in poor transduction of HSVECs. Little toxicity (<20%) was seen with all viral vectors. Whereas nonviral vectors can be used to delivery genes to ECs, 12,13 their efficiency is, in general, low. The best transfection efficiency was seen with immunoliposomes made with anti-CD71 Ab (&40%) (Figure I).

Effect of Transfection/Transduction on EC Phenotype

As previously reported, 7 we found that exposure of HSVECs to Ad vectors (with a MOI of 500) (either Ad-EGFP ( Figure 1; Figure II, available online at http://atvb.ahajournals.org) or control (Ad0) 8 (data not shown)) resulted in considerable upregulation of adhesion molecules and MHC class II molecules. After lentiviral transduction, we saw a similar upregulation of adhesion molecules CD54, CD62, and CD106 on HSVECs. Exposure to retroviral vectors also resulted in upregulation of adhesion molecule expression (though this was less in the case of CD106). However, transfection of HSVECs with nonviral vectors caused only slight upregulation of CD62 and CD106, equivalent to that seen in response to DNA alone ( Figure 1; Figure II). Incubation of HSVECs in medium/buffers used to prepare viruses did not affect the phenotype of the cells (data not shown).

Figure 1. Viral vectors change the phenotype of ECs. After transduction or transfection of the HSVEC or microvascular ECs with nothing (U), Ad, HIV, EIAV, MMLV, DNA alone, liposomes (L), dendrimers (D), or immunolipoplexes (ImmL), cells were analyzed by flow cytometry for the expression of EGFP and CD54. Relative mean fluorescence intensity of CD54 in EGFP-positive cells was determined. In parallel, supernatants were collected on day 5 and the levels of TNF- and IFN- were determined. Results are expressed as the mean of triplicates±SD. * P <0.05.

To determine whether similar upregulation was seen in ECs from other sources, we repeated these experiments with microvascular ECs isolated from human foreskin. Similar upregulation of CD54 ( Figure 1 ), CD62, and CD106 (data not shown) was seen after viral transduction.

Viral Vectors Induce Production of Th1 Cytokines and Inflammatory Cytokines

We saw an upregulation of Th1 cytokines, IL-12 and IFN-, as well as enhanced secretion of TNF-, IL-8, IL-1ß, and IL-6 ( Figure 1; Figure III, available online at http://atvb.ahajournals.org) in HSVECs and microvascular ECs after Ad, lentiviral, or MMLV transduction. Minimal effects on cytokine secretion were detected after nonviral transfection.

Activation Pathways Induced by Viral Transduction

ECs produce type I IFNs in response to double-stranded RNA (dsRNA) by a pathway that is partially dependent on the cytosolic dsRNA-binding enzyme protein kinase R. 9,10 Activation of this pathway has not been previously reported in ECs after viral vector transduction, although it was seen after the intracellular introduction of dsRNA into ECs. 24 In addition, the NF B pathway may be activated through the 2'-5' OA synthetase/RNase L pathway. Therefore, we determined whether these pathways are activated in ECs after transduction. All 4 viral vectors triggered the production of type 1 IFN /ß at mRNA ( Figure 2 A) and IFN-, ß, and at the protein level ( Figure 1; Figure III), together with upregulation of 2'-5' OA synthetase and interferon (IFN) regulatory factor 1 (IRF-1) at mRNA levels (2'-5' OA synthetase/RNase L pathway) ( Figure 2 A). We also saw phosphorylation of PKR and its downstream substrate eIF-2, indicative of PKR pathway activation, and upregulation of PI3 kinase and its downstream substrate PKB/Akt ( Figure 2 B). Similar data were observed with microvascular endothelium (data not shown).

Figure 2. Transduction activates EC signaling pathways. After viral transduction with Ad, HIV, EIAV, MMLV or transfection with lipoplexes, dendriplexes, immunolipoplexes, or plasmid DNA alone, HSVECs were analyzed 2 days later (A) by reverse-transcription polymerase chain reaction to measure mRNA levels for PKRK, 2'-5'OAS, IRF-1, IFN-, IFN-ß and ß-actin, and (B) by Western blotting for expression of phosphorylated and total proteins for PKR, eIF-2, I B, PKB/Akt, and PI3 kinase, as well as ß-actin. The data shown are representative of 3 experiments. C, ECs were analyzed after transduction at lower MOI (Ad-EGFP MOI 50, EIAV-EGFP MOI 75, and HIV-EGFP MOI 500). Alternatively, immunolipoplex-transfected cells were drug-selected to ensure 100% transfection efficiency (nonviral selected). To exclude the effect of differential production of EGFP, immunolipoplex-transfected HSVECs were sorted to select cells (nonviral sorted) with an equivalent MFI to virally transduced cells (Ad-EGFP). Cells transduced with Ad encoding ß-galactosidase (Ad-ß-gal) or transfected with a plasmid encoding ß-galactosidase (nonviral ß-gal) were also analyzed. Cells were transduced with Ad-0, EIAV-0, and HIV-0 (viruses not encoding a transgene). All ECs were Western-blotted for phosphorylated PKR, I B, PKB/Akt, and ß-actin.

These data suggest that all 3 pathways may operate to activate (among others) the NF B pathway (as shown by increases in phosphorylated I B) ( Figure 2 B). This is one of the central pathways for EC activation. Viral preparation medium did not activate any of the signaling pathways (data not shown).

To address whether the differences in EC activation between viral and nonviral vectors is caused by the much higher EGFP expression after viral transduction, we transduced HSVEC at a lower multiplicity of infection (MOI) (Ad-EGFP, MOI 50; EIAV-EGFP, MOI 75; HIV-EGFP, MOI 500), resulting in an equivalent number of cells expressing EGFP to that seen after immunolipoplex transfection (40% to 50%). All viral vectors activated the I B, PKR, and PI3K/Akt pathways, but this is not seen after transfection with immunolipoplexes ( Figure 2 C). We then transfected HSVECs with immunolipoplexes, followed by drug selection with G418 (200 µg/mL), resulting in 100% of cells expressing EGFP, equivalent to cells transduced with Ad-EGFP and EAIV-EGFP at MOI 500. No obvious activation of I B, PKR, and PI3K/Akt pathways was seen ( Figure 2 C). Finally, to confirm that the effects were not caused by higher levels of EGFP expression after transduction, we flow-sorted a population of HSVECs that, after immunolipoplex transfection, had a high expression of EGFP equivalent to that seen after viral transduction. These cells showed no evidence of activation ( Figure 2 C).

We performed transduction with Ad encoding for ß-galactosidase (RAd35) or Ad-0, HIV-0, and EIAV-0 that did not contain a transgene. In all cases, we saw similar activation of I B, PKR and PI3K/Akt pathways ( Figure 2 C). This indicates that it is the viral vector itself that activates ECs rather then the expression of the transgenes.

Altered Functions of ECs After Viral Transduction

To test the functional consequences of viral transduction, we transduced HSVECs with viral vectors, and then used them as in leukocyte adhesion and transmigration assays. When the ECs were transduced with Ad, lentivirus, and MMLV, the adherence and transmigration of leukocytes (in particular granulocytes and lymphocytes) was markedly increased to a comparable extent to that seen after cytokine activation (Figure IV, available online at http://atvb.ahajournals.org). There was no increase in adhesion or transmigration after transfection with the nonviral vectors (data not shown).

Virally Activated HSVECs Fail to Respond to Subsequent Activation by Proinflammatory Cytokines

After transfection or transduction, HSVECs were cultured for 5 days. They were then challenged with proinflammatory cytokines. Virally transduced HSVECs were less responsive in term of adhesion molecule expression and TNF- secretion but had higher (in the case of Ad-transduced HSVECs) production of IL-8 ( Figure 3 A; Figure V, available online at http://atvb.ahajournals.org). Similar observations were seen when the subsequent activation was with oxidized LDL (data not shown).

Figure 3. Virally transduced HSVECs are "resistant" to further activation. HSVECs were transduced with Ad, HIV, EIAV, or MMLV, or transfected with lipoplexes (L), dendriplexes (D), immunolipoplexes (ImmL), or plasmid DNA, or were untransfected/untransduced (U). Five days later, they were activated with IL-1ß, TNF-, and IFN-, and CD54 expression was determined 12 hours later (A). Results are expressed as the relative fluorescence intensity (normalized to isotype control staining). B, (upper panel) ECs were cultured in the presence of IL-1 and TNF- for 5 days before stimulation (or not) with IL-1ß and TNF-, in the presence or absence of anti-TNF-, IL-1, and ß neutralizing antibodies. The expression of adhesion molecules and HLA-DR was determined 12 hours later. B, (lower panel) HSVECs were transduced with viruses in the presence or absence of neutralizing antibodies to IL-1 and TNF-. Five days later, cells were stimulated with IL-1ß, TNF-, and IFN- for 12 hours, and expression of adhesion molecules and human major histocompatibility complex (HLA)-DR was determined. The results are shown as the mean of triplicate wells±SD. Data are representative of 2 experiments. * P <0.05.

To determine why HSVECs previously transduced with virus had a reduced response to pro-inflammatory agents, we cocultured the HSVECs in the presence of IL-1ß and TNF- for 5 days. HSVECs that were chronically stimulated with these cytokines failed to upregulate adhesion molecules and MHC class expression on subsequent activation ( Figure 3 B). If HSVECs were transduced with virus in the presence of neutralizing antibodies to IL-1 and TNF-, they were able to respond to activation stimuli in a similar manner to untransduced HSVECs ( Figure 3 B). These data indicate that IL-1 and TNF- are necessary and sufficient for the effect seen.

Inhibition of Signaling Pathways After Viral Transduction

After viral transduction, there is activation of PI3-K and PKR pathways, which stimulate the NF B pathway. We used 3 strategies to inhibit NF B activation after viral transduction: (1) prevention of NF B (p55 and p65) translocation (SN50); (2) prevention of I B breakdown (proteasome inhibitors PSI and MG132); and (3) induction of I B synthesis 25 and direct interference with NF B-dependent transactivation 26 (dexamethasone). Activation was determined by upregulation of adhesion molecules ( Figure 4 ), cytokine secretion, and effects on the signaling molecules (Figures VI and VII, available online at http://atvb.ahajournals.org). Inhibition of the NF B pathway partially prevented HSVEC activation after exposure to Ad. Trichodion, which inhibits AP-1 and STAT-1 in addition to NF B, 27 partially prevents HSVEC activation. These data suggest that additional inhibition of AP-1 and STAT-1 has little effect in preventing viral activation of HSVECs over and above that seen with inhibition of NF B alone.

Figure 4. Effect of inhibition of signaling pathways on EC activation. HSVECs were cultured with SN50, PSI, MG132 (NF B inhibitors), SB23906 (MAPK p38 inhibitor), SP600125 (JNK inhibitor), STAT3 inhibitor peptide (STAT-3 inhibitor), 2-AP (PKR inhibitor), Trichodion (an inhibitor of NF B, AP-1 and STAT-1), Dexamethasone (inhibitor of NF B and MAPK p38), LY294002, Wortmannin (inhibitors of PI3K), or neutralizing antibodies against IL-1, TNF-, IFN-, and IFN-ß for 24 to 48 hours before transduction with Ad; 48 hours after transduction, the expression of CD54, CD62, and CD106 was analyzed. For all panels, data from untreated HSVECs (with and without transduction) are shown.

Having shown that the downstream inhibition of virally activated signaling did not completely prevent EC activation, we investigated the upstream inhibition of PI3-K, PKR, MAPK p38, JNK, and STAT3 pathways. Inhibition of MAPK p38, JNK, and STAT3 pathways had a minimal effect on EC activation. However, inhibition of the PI3-K or PKR pathways resulted in a partial inhibition of activation.

Effect of Blocking Cytokine Production on HSVEC Activation

The PKR pathway induces IFN- and IFN-ß, whereas the PI3K pathway produces TNF- and IL-1 and IL-1ß. These cytokines are capable of autocrine signaling resulting in positive feedback loops that amplify activation of HSVECs. We therefore blocked these cytokine pathways, alone or in combination with inhibition of the signaling pathways. Addition of neutralizing antibodies to TNF- and IL-1 and IL-1ß only minimally inhibited the NF B and PKR pathways after viral transduction, and had little effect on adhesion molecule expression. Addition of neutralizing antibodies to IFN- or IFN-ß (alone or in combination) had little effect on HSVEC activation, whereas blockade of all 4 cytokines did partially inhibit HSVEC activation.

Effect of Inhibition of Both PKR and PI3K Pathways

Inhibition of either PKR or PI3K on their own results in partial inhibition of HSVEC activation after transduction. However, addition of both LY294002 and aminopurine totally inhibits upregulation of HSVEC activation after viral transduction ( Figure 4 ). We inhibited the cytokines stimulated by these pathways. Thus, inhibition of the PI3K pathway with LY294002 at the same time as blocking IFN- and IFN-ß results in total inhibition of HSVEC activation ( Figure 4 ). However, inhibition of PKR with aminopurine together with anti-TNF- and IL-1 antibodies only partially inhibited activation (data not shown). The use of all these inhibitors and all these neutralizing antibodies did not affect the transgene expression (data not shown).

Effect of Inhibitors on Leukocyte Adherence and Transmigration

NF B, JNK, MAPK p38, and PKR inhibitors, Dexamethasone and Trichodion, which partially inhibited HSVEC activation, also partially blocked the increase in adhesion and transmigration of leukocytes after Ad transduction. However, treatment of HSVECs with a combination of PKR and PI3K inhibitors, or with PI3K inhibition in conjunction with IFN- and IFN-ß blockade, resulted in no augmentation of leukocyte adhesion or transmigration after Ad transduction (Figure VIII, available online at http://atvb.ahajournals.org).

Discussion

In this report we have characterized the effect of viral transduction on EC phenotype and function. Infection of A594 cells, 28 vascular neointimal cells, or pulmonary tissues 29 with Ad vectors induces upregulation of CD54, with E1-E4+ Ad vectors, but not E1-E4- vectors, inducing long-term expression of CD54 and CD106. 7 We have observed not only an upregulation of adhesion molecule expression but also an increase in the secretion of Th1-type proinflammatory cytokines by ECs after transduction with Ad, lentivirus, and MMLV.

The NF B pathway is primarily responsible for activation of ECs. However, there are several ways in which NF B can be activated by viral transduction. In response to viral infection, ECs can be activated by dsRNA. This is partially dependent on the cytosolic dsRNA-binding protein kinase R (PKR). In dendritic cells (DCs), activation by dsRNA does not depend on the toll-like receptor (TLR) 3, a surface receptor for dsRNA, 30 but requires endosomal recognition through TLR7 and MyD88. 30,31 In this report, we show that viral vectors activate the 2'-5' OA synthetase/RNase L and PKR pathway in ECs. The consequences of viral transduction can vary, with ECs having a reduced ability to upregulate adhesion molecules and cytokine production when activated 5 days after transduction. This effect could be reproduced by addition of TNF- and IL-1ß for 5 days, and was reversed by addition of neutralizing antibodies to these cytokines, indicating that this resulted from secretion of proinflammatory cytokines by transduced ECs.

Ad vectors can produce inflammatory responses at high doses after intravenous administration in vivo through production of cytokines and chemokines. 32 The administration of transcription-defective Ad-GFP particles induced the early but not the late peak of chemokine and cytokine gene expression. These data indicate that Ad vector-induced inflammation is capsid-dependent in the early phase and transcription-dependent in the late phase. 32 Ad-mediated activation of DCs was recently attributed to the high levels of TNF- expression by these cells. 11 Ad-induced TNF- production was found to be necessary for DC maturation, in a manner dependent on signaling by PI3-K but not MyD88. Our observations are consistent with a similar pathway operating in human ECs, in as much as we saw the activation of PI3-K and high production of TNF- after Ad transduction.

Having established pathways that were activated after viral transduction, we investigated the effect of inhibiting them. Inhibition of NF B did not completely abolish virally induced adhesion molecule expression and pro-inflammatory cytokine production. However, inhibition of both PI3-K and PKR resulted in total inhibition of EC activation, suggesting that the PI3-K and PKR pathways not only activate NF B but also other downstream effector pathways. Use of other inhibitors suggested that blockade of the MAPK p38, JNK, STAT-1, and STAT3 pathways has little effect on the activation of ECs by viral transduction. Both the PKR and the PI3-K pathways are associated with cytokine production (IFN- /ß for PKR and TNF- /IL-1 /ß for PI3-K), serving as positive feedback loops that may help "lock" the ECs into an activated phenotype. Inhibition of these pathways on their own has only a partial effect on EC activation; however, inhibition of the PI3-K pathway in combination with blockade of type 1 IFNs totally abolished EC activation. These data indicate strategies that might prevent nonspecific EC activation by viral vectors.

In this report, we have chosen HSVEC as our primary model because saphenous veins are used widely as conduit in many bypass operations, and genetic approaches have been proposed to prevent graft restenosis. However, there are differences in EC biology, depending on their origin. Microvascular ECs are an important target for some applications, and we found similar activation of these cells after transduction.

In conclusion, we have shown that transduction with viral vectors has dramatic consequences for ECs in terms of their phenotype, the activation pathways induced, and their function. This can modulate the effect of gene therapy by inducing local inflammation, which may modulate the underlying disease process and also induce anti-vector or anti-transgene immune responses. Nonviral vectors, however, are relatively innocuous and fail to activate significantly ECs. However, they are less efficient at mediating gene expression! An understanding of the consequences of EC transduction and the pathways by which it is mediated is important in allowing the design of improved strategies involving the viral vector modification of ECs.

Acknowledgments

This research is supported by fellowships and grants from the MRC, the RCS Edinburgh (P.H.T.), Getrud-Kusen-Stiftung Hamburg (S.C.B.), the MRC, and the BBSRC. A.J.T.G. is a BBSRC Research Development Fellow. The authors thank Oxford Biomedica for providing lentiviral vectors and the theater nursing staff at Hammersmith and St. Mary?s Hospitals for collecting vessels.

【参考文献】
  Zoldhelyi P, McNatt J, Xu XM, Loose-Mitchell D, Meidell RS, Clubb FJ, Jr., Buja LM, Willerson JT, Wu KK. Prevention of arterial thrombosis by adenovirus-mediated transfer of cyclooxygenase gene. Circulation. 1996; 93: 10-17.

Muhlhauser J, Merrill MJ, Pili R, Maeda H, Bacic M, Bewig B, Passaniti A, Edwards NA, Crystal RG, Capogrossi MC. VEGF165 expressed by a replication-deficient recombinant adenovirus vector induces angiogenesis in vivo. Circ Res. 1995; 77: 1077-1086.

Dishart KL, Denby L, George SJ, Nicklin SA, Yendluri S, Tuerk MJ, Kelley MP, Donahue BA, Newby AC, Harding T, Baker AH. Third-generation lentivirus vectors efficiently transduce and phenotypically modify vascular cells: implications for gene therapy. J Mol Cell Cardiol. 2003; 35: 739-748.

Shahgasempour S, Woodroffe SB, Garnett HM. Alterations in the expression of ELAM-1, ICAM-1 and VCAM-1 after in vitro infection of endothelial cells with a clinical isolate of human cytomegalovirus. Microbiol Immunol. 1997; 41: 121-129.

Zietz C, Hotz B, Sturzl M, Rauch E, Penning R, Lohrs U. Aortic endothelium in HIV-1 infection: chronic injury, activation, and increased leukocyte adherence. Am J Pathol. 1996; 149: 1887-1898.

Harcourt BH, Rota PA, Hummel KB, Bellini WJ, Offermann MK. Induction of intercellular adhesion molecule 1 gene expression by measles virus in human umbilical vein endothelial cells. J Med Virol. 1999; 57: 9-16. <a href="/cgi/external_ref?access_num=10.1002/(SICI)1096-9071(199901)57:1

Rafii S, Dias S, Meeus S, Hattori K, Ramachandran R, Feuerback F, Worgall S, Hackett NR, Crystal RG. Infection of endothelium with E1(-)E4(+), but not E1(-)E4(-), adenovirus gene transfer vectors enhances leukocyte adhesion and migration by modulation of ICAM-1, VCAM-1, CD34, and chemokine expression. Circ Res. 2001; 88: 903-910.

Tan PH, Beutelspacher SC, Xue SA, Wang YH, Mitchell P, McAlister JC, Larkin DF, McClure MO, Stauss HJ, Ritter MA, Lombardi G, George AJ. Modulation of human dendritic-cell function following transduction with viral vectors: implications for gene therapy. Blood. 2005; 105: 3824-3832.

Williams BR Signal integration via PKR. Sci STKE. 2001; 2001: RE2.

Malmgaard L. Induction and regulation of IFNs during viral infections. J Interferon Cytokine Res. 2004; 24: 439-454.

Philpott NJ, Nociari M, Elkon KB, Falck-Pedersen E. Adenovirus-induced maturation of dendritic cells through a PI3 kinase-mediated TNF-alpha induction pathway. Proc Natl Acad Sci U S A. 2004; 101: 6200-6205.

Tan PH, Manunta M, Ardjomand N, Xue SA, Larkin DFP, Haskard DO, Taylor KM, George AJT. Antibody targeted gene transfer to endothelium. J Gene Med. 2003; 5: 311-323.

Hellgren I, Drvota V, Pieper R, Enoksson S, Blomberg P, Islam KB, Sylven C. Highly efficient cell-mediated gene transfer using non-viral vectors and FuGene6: in vitro and in vivo studies. Cell Mol Life Sci. 2000; 57: 1326-1333.

Manunta M, Tan PH, Sagoo P, Kashefi K, George AJ. Gene delivery by dendrimers operates via a cholesterol dependent pathway. Nucleic Acids Res. 2004; 32: 2730-2739.

Turunen MP, Hiltunen MO, Ruponen M, Virkamaki L, Szoka FC Jr, Urtti A, Yla-Herttuala S. Efficient adventitial gene delivery to rabbit carotid artery with cationic polymer-plasmid complexes. Gene Ther. 1999; 6: 6-11.

Tan PH, Beutelspacher S, Wang YH, Ritter MA, McClure MO, Lombardi G, George AJ. Immunolipoplexes: an efficient, non-viral alternative for transfection of human dendritic cells with potential for clinical vaccination. Mol Ther. 2005; 11: 790-800.

Tan PH, Chan C, Xue SA, Dong D, Ananthesayanan B, Manunta M, Kerouedan C, Cheshire C, Wolfe J, Haskard DO, Taylor KM, George AJT. Phenotypic and functional differences between human saphenous vein (HSVEC) and umbilical vein (HUVEC) endothelial cells. Atherosclerosis. 2004; 173: 171-183.

Mason JC, Yarwood H, Tarnok A, Sugars K, Harrison AA, Robinson PJ, Haskard DO. Human Thy-1 is cytokine-inducible on vascular endothelial cells and is a signaling molecule regulated by protein kinase C. J Immunol. 1996; 157: 874-883.

Tan PH, King WJ, Chen D, Awad HM, Mackett M, Lechler RI, Larkin DF, George AJ. Transferrin receptor-mediated gene transfer to the corneal endothelium. Transplantation. 2001; 71: 552-560.

Tan PH, Sagoo P, Chan C, Yates JB, Campbell J, Beutelspacher SC, Foxwell BM, Lombardi G, George AJ. Inhibition of NF-{kappa}B and Oxidative Pathways in Human Dendritic Cells by Antioxidative Vitamins Generates Regulatory T Cells. J Immunol. 2005; 174: 7633-7644.

Tan PH, Yates JB, Xue SA, Chan C, Jordan WJ, Harper JE, Watson MP, Dong R, Ritter MA, Lechler RI, Lombardi G, George AJ. Creation of tolerogenic human DC via intracellular CTLA4: a novel strategy with potential in clinical immunosuppression. Blood. prepublished online June 30, 2005; DOI 10.1182/blood-2005-05-1826.

Corbeil J, Evans LA, McQueen PW, Vasak E, Edward PD, Richman DD, Penny R, Cooper DA. Productive in vitro infection of human umbilical vein endothelial cells and three colon carcinoma cell lines with HIV-1. Immunol Cell Biol. 1995; 73: 140-145.

Xu XM, Ohashi K, Sanduja SK, Ruan KH, Wang LH, Wu KK. Enhanced prostacyclin synthesis in endothelial cells by retrovirus-mediated transfer of prostaglandin H synthase cDNA. J Clin Invest. 1993; 91: 1843-1849.

Harcourt JL, Hagan MK, Offermann MK. Modulation of double-stranded RNA-mediated gene induction by interferon in human umbilical vein endothelial cells. J Interferon Cytokine Res. 2000; 20: 1007-1013.

Auphan N, DiDonato JA, Rosette C, Helmberg A, Karin M. Immunosuppression by glucocorticoids: inhibition of NF-kappa B activity through induction of I kappa B synthesis. Science. 1995; 270: 286-290.

De Bosscher K, Schmitz ML, Vanden Berghe W, Plaisance S, Fiers W, Haegeman G. Glucocorticoid-mediated repression of nuclear factor-kappaB-dependent transcription involves direct interference with transactivation. Proc Natl Acad Sci U S A. 1997; 94: 13504-13509.

Erkel G. Trichodion, a new inhibitor of inflammatory signal transduction pathways from a Trichosporiella species. FEBS Lett. 2000; 477: 219-223.

Stark JM, Amin RS, Trapnell BC. Infection of A549 cells with a recombinant adenovirus vector induces ICAM-1 expression and increased CD-18-dependent adhesion of activated neutrophils. Hum Gene Ther. 1996; 7: 1669-1681.

Nicolis E, Tamanini A, Melotti P, Rolfini R, Berton G, Cassatella MA, Bout A, Pavirani A, Cabrini G. ICAM-1 induction in respiratory cells exposed to a replication-deficient recombinant adenovirus in vitro and in vivo. Gene Ther. 1998; 5: 131-136.

Diebold SS, Montoya M, Unger H, Alexopoulou L, Roy P, Haswell LE, Al-Shamkhani A, Flavell R, Borrow P, Reis e Sousa C. Viral infection switches non-plasmacytoid dendritic cells into high interferon producers. Nature. 2003; 424: 324-328.

Diebold SS, Kaisho T, Hemmi H, Akira S, Reis e Sousa C. Innate antiviral responses by means of TLR7-mediated recognition of single-stranded RNA. Science. 2004; 303: 1529-1531.

Liu Q, Zaiss AK, Colarusso P, Patel K, Haljan G, Wickham TJ, Muruve DA. The role of capsid-endothelial interactions in the innate immune response to adenovirus vectors. Hum Gene Ther. 2003; 14: 627-643.

作者单位：Peng H. Tan; Shao-An Xue; Maria Manunta; Sven C. Beutelspacher; Henrieta Fazekasova; A.K.M. Shamsul Alam; Myra O. McClure; Andrew J.T. GeorgeFrom the Department of Immunology (P.H.T., S.-A.X., S.C.B., M.M., H.F., A.K.M.S.A., A.J.T.G.), Division of Medicine, Imperial College London, Hammersmith Hospi