key: cord-1040680-g0hucasa
authors: Arcasoy, SM; Latoche, JD; Gondor, M; Pitt, BR; Pilewski, JM
title: Polycations increase the efficiency of adenovirus-mediated gene transfer to epithelial and endothelial cells in vitro
date: 1997-12-18
journal: Gene Ther
DOI: 10.1038/sj.gt.3300349
sha: e2d78f91ecc771cdd49be85eb3e2f93970824748
doc_id: 1040680
cord_uid: g0hucasa

Recombinant adenoviruses are being developed for gene therapy for cystic fibrosis and other lung diseases, and for prevention and treatment of vascular thrombosis. A major limitation to the clinical utility of adenoviruses is the low efficiency of gene transfer achieved in vivo. In addition, little is known about the initial interactions between adenoviruses and the target cell. To address the hypothesis that the negative charge presented by membrane glycoproteins reduces the efficiency of adenovirus-mediated gene transfer, primary cultures of human airway, Madin–Darby canine kidney cells, an immortalized cystic fibrosis airway epithelial cell line, and primary cultures of sheep pulmonary artery endothelium were infected with recombinant adeno- virus containing the E. coli lacZ reporter gene (Ad2βgal2) in the presence of various polyions. For each cell type, adsorption of Ad2βgal2 in the presence of the polycations polybrene, protamine, DEAE-dextran, and poly-L-lysine significantly increased the percentage of cells that express lacZ. The polyanion heparin did not significantly alter gene transfer efficiency, but completely abrogated the effects of polycations. These data provide evidence that negatively charged moieties on the cell surface reduce the efficiency of adenovirus-mediated gene transfer, and that alteration of the charge interaction between adenoviruses and the cell surface may improve the potential clinical application of these vectors.

Apical membrane sialoglycoconjugates, either by steric Gene therapy for treatment of several hereditary and hindrance or charge interactions related to the abundant acquired diseases has gained increasing attention in the negatively charged sialic acid residues, may interfere past decade. 1-4 Despite remarkable progress in the develwith adenovirus binding to the target cell. The imporopment of both viral and nonviral vectors and the tance of sialic acid and cell surface charge in virus-cell initiation of clinical gene therapy trials, the low efficiency interactions has been demonstrated for several enveloped of gene transfer is one of the major obstacles to the cliniand nonenveloped viruses, such as retrovirus, 17, 18 murine cal application of many of these vectors. Recombinant sarcoma and leukemia viruses, [19] [20] [21] avian sarcoma virus, 22 adenoviruses have been evaluated extensively because of respiratory syncitial virus, 23 reovirus, 24 and rotatheir relatively high efficiency and ability to drive viruses. 25, 26 For example, Kahn et al 18 studied the effects expression of a foreign gene in nondividing cells.

of various concentrations of polybrene and DEAE-dex-Although efficient adenovirus-mediated gene transfer to tran on transduction of bovine aortic endothelial cells cultured epithelial 5 and endothelial 6, 7 cells, and to various using retroviral vectors, and found a significantly in vivo airway 8 and vascular models 9, 10 has been reported, enhanced transduction efficiency with the use of these differentiated epithelial and endothelial cells are relapolycations. tively resistant to adenovirus infection, requiring high Despite the relatively well understood phases of multiplicities of infection (MOI) and a relatively long adenovirus infection, 27 little is known about the potential physical contact time between adenovirus and target cell implications of charge interactions for cell binding and for significant gene transfer. [11] [12] [13] [14] [15] [16] entry of adenoviruses. Our recent work has revealed that The mechanisms for this relative inefficiency are poorly removal of negatively charged sialic acid residues from understood, but potentially include secreted mucins, cultured polarized epithelial cells significantly improves inflammatory mediators, cell surface sialoglycoconjuthe efficiency of adenovirus-mediated gene transfer to gates, tight junctions, polarized expression of viral bindthese cells. 28 In the current study, we hypothesized that ing or internalization receptors, and differential rates of the addition of positively charged polymers during adsorption of recombinant adenovirus would also increase the efficiency of gene transfer. HBE cells: As shown in Figures 1 and 2 , Ad2␤gal2 infection of primary HBE cells at an MOI of 25 resulted in lacZ expression in 48 ± 4% of the cells, whereas adsorption in marked. The percentage of lacZ-expressing cells increased more than two-fold, from 19 ± 1% to 46 ± 2% (P the presence of polybrene at concentrations of 1, 2, and 4 g/ml increased the percentage of cells expressing the Ͻ 0.0001 by ANOVA) when adsorption of Ad2␤gal2 was performed in the presence of 4 g/ml of polybrene (data transgene to 85 ± 1%, 94 ± 1%, and 97 ± 1% of HBE cells, respectively (P Ͻ 0.0001 by ANOVA). Similarly, viral not shown). As has been shown for the effects of polybrene on the adsorption in the presence of 1 g/ml of protamine resulted in transgene expression in 75 ± 6% of the cells (P efficiency of retrovirus-mediated gene transfer, 18 there is a limited dose response relationship for the polycation = 0.0002 by ANOVA). At an MOI of 10, the increase in gene transfer efficiency with polybrene was more effect. At polybrene concentrations of 4 g/ml and above (see Figure 2 ), and protamine concentrations of 5 g/ml and above (data not shown), the gene transfer efficiency reached a plateau insofar as further increases in polycation concentration resulted in minimal changes in the percentage of cells expressing the transgene. Therefore subsequent experiments were performed with polybrene and protamine concentrations of 4 and 5 g/ml, respectively.

MDCK cells: Ad2␤gal2 infection of MDCK type I cells at MOI of 10, 25 and 100 resulted in transgene expression in 1 ± 0.1%, 5 ± 0.4% and 12 ± 3.0% of cells, respectively. At each MOI the percentage of transgene-expressing cells increased significantly when viral adsorption was performed in the presence of a polycation. As shown in Figure  3 , the effect of polycations was most pronounced at lower MOIs, with an approximately 30-fold increase in the per- 

Although the relative increase in ␤-galactosidase at the same MOI.

expression following Ad2␤gal2 infection at an MOI of 100 was of lesser relative magnitude than at an MOI of 10, the difference was very significant insofar as nearly adsorption in the presence of polybrene increased 100% of the cells expressed lacZ when polycations were transgene expression more than 20-fold at an MOI of 10 present during viral adsorption (see Figure 3 ). The effect (P Ͻ 0.0001) and 16-fold at an MOI of 100 (P Ͻ 0.0001). at an MOI of 25 was similar (data not shown). When Similar results were obtained with protamine, as ␤polybrene was added to the media for 1 h after adenogalactosidase expression increased more than 40-fold at viral adsorption, there was no significant increase in lacZ an MOI of 10, and 17-fold at an MOI of 100 (Ͻ0.5% versus expression (12 ± 1% versus 16 ± 1%, P Ͼ 0.05), suggesting 22 ± 0.3%, and 4 ± 0.2% versus 66 ± 2%, respectively; P Ͻ that polybrene is affecting viral binding and/or internal-0.0001 for both MOI). ization. Figure 5 , at an MOI of 5, the percentage at an MOI of 10 and 100 resulted in transgene expression of cells that expressed lacZ was 44 ± 0.4%; when in Ͻ0.5% and 4 ± 0.2% of the cells, respectively. Viral

Ad2␤gal2 adsorption was performed in the presence of polycations, the percentage of positive cells increased arin. As shown in Figure 6 , at an MOI of 10, transgene expression increased four-fold (from 12 ± 1% to 48 ± 3% approximately two-fold to 89 ± 4% with polybrene (P Ͻ 0.0001), and 81 ± 2% with protamine (P = 0.0002). At an of the cells) with polybrene incubation, but was not different from the control when adsorption was performed MOI of 100, transgene expression was observed in over 90% of cells in the presence or absence of polycations with heparin alone. Moreover, Ad2␤gal2 adsorption in the presence of both polybrene and heparin resulted in (data not shown).

lacZ expression in 13.2 ± 1% of cells, which is not significantly different from adenovirus alone. These data dem-The effects of polycations are observed with adsorption of virus at either 4°C or 37°C, consistent with a onstrate that polyanions such as heparin have no effect on gene transfer efficiency, but that polyanions inhibit the polycation-mediated increase in adenoviral binding To begin to determine whether the observed increase in polybrene-mediated increase in gene transfer efficiency.

In addition, as shown in Figure 6 , similar effects on gene gene transfer efficiency is due to enhanced viral attachment or internalization, viral adsorption to IB3 cells was transfer were observed when adenoviral adsorption was performed after a 2-min wash of cells with polybrene or performed at either 37°C or 4°C. The magnitude of increase in the percentage of cells that express lacZ was both polybrene and heparin immediately before adenoviral incubation. Polybrene wash increased gene transfer similar under both conditions, as shown in Figure 5 . When adsorption was performed at 37°C, the percentage efficiency (from 12 ± 1% to 34 ± 12%, P = 0.0018), indicating that the polycation need not be present during viral of positive cells increased approximately three-fold with polybrene and two-fold with protamine (P = 0.0001 for adsorption. Combined polybrene and heparin wash did not result in a significant change in the percentage of both conditions). When Ad2␤gal2 adsorption was performed at 4°C in the presence of polybrene or protamine, lacZ-expressing cells (12 ± 1% versus 13 ± 1%, P = 0.864). the percentage of positive cells was slightly lower, but the magnitude of the polycation effect was nearly identical.

The effects of polycations are observed with polymers ranging in size from 3000 to over 500 000 Da To determine whether the size of the polycation was criti-The effects of polycations are abrogated by co-incubation with heparin cal to its effects on gene transfer efficiency, Ad2␤gal2 at an MOI of 25 was adsorbed in the presence of the follow-Initial experiments with heparin alone were done to determine the effect of polyanions on recombinant ing polycations that vary in molecular weight: polybrene (3000 Da), protamine (4000 Da), DEAE-dextran (500 000 adenovirus infection. In each of the cell lines, viral adsorption in the presence of heparin at up to 50 U/ml Da), and poly-l-lysine (70 000-150 000 Da). As shown in Figure 7 , each of the polycations significantly increased did not result in statistically significant differences in the percentage of cells expressing lacZ (see Figures 2, 3 and the efficiency of gene transfer (15 ± 1% with adenovirus alone versus 44 ± 4% with polybrene, 51 ± 1% with poly-l-6). To determine whether co-incubation of heparin and polycations during viral adsorption would alter trans-lysine, 33 ± 4% with protamine, and 42 ± 1% with DEAEdextran, P Ͻ 0.0001 for each polycation by ANOVA). As gene expression, Ad2␤gal2 was adsorbed to HBE cells at 37°C for 1 h in the presence of both polybrene and hep-demonstrated in Figure 7 , the percentage of lacZ-expressing cells was unaltered by the addition of free sialic acid (P = 0.74) or glycophorin A (P = 0.91) during adsorption of Ad2␤gal2, however, both free sialic acid or glycophorin A abrogated the effects of polycations (data not shown), suggesting that the effect is due to charge interactions rather than interaction with specific oligosaccharide moieties, and is independent of chain length. We speculate that the effect of polycations could occur via a number of mechanisms, including neutralization of of viruses. [17] [18] [19] [20] [21] [22] [23] [24] For example, influenza virus and bovine coronaviruses require sialic acid residues for cell bind-cell surface negative charge, formation of a bridge between the negatively charged cell surface and adeno-ing. 29, 30 In contrast to these enveloped viruses, the molecular mechanisms of cell entry employed by adeno-virus particle, or an increase in the permeability of cell monolayers. In the first instance, one need not implicate viruses are only partially understood, 26,31 and the importance of cell surface charge in viral attachment and additional or alternative receptors. Binding and internalization would occur via fiber and penton base, albeit entry has not been evaluated to date. Our recent work has revealed that enzymatic removal of negatively more efficiently after negation of the charge barrier. In contrast, a bridging effect of the polycations could occur charged sialic acid residues from epithelial cell surface glycoconjugates significantly increases adenovirus-through enhancement of fiber or penton base binding, or through bridging of an alternative viral protein to a cell mediated gene transfer, 28 suggesting that the negative charge of sialic acid residues on cell surface glycocon-receptor. Lastly, we cannot rule out an effect of polycations on epithelial permeability, since Peterson and col-jugates may impair the binding of adenoviruses to epithelial cells.

leagues 37 have demonstrated that polycations such as protamine decrease the transepithelial electrical resist-In the current study, we extend these observations and show that the presence of positively charged polymers of ance and increase the permeability of MDCK monolayers to mannitol. However, two observations argue against different sizes during or immediately before adenovirus adsorption markedly increases the efficiency of adeno-increased epithelial permeability as the primary mechanism for a polycation-mediated increase in gene trans-virus-mediated gene transfer to both epithelial and endothelial cells. This effect occurs when viral adsorption is fer. First, in published studies to date, the reduction in transepithelial electrical resistance becomes evident after performed at either 37°C or 4°C, suggesting that polycations enhance the attachment of adenovirus to the cell.

10 min of exposure at 37°C. 37, 38 However, when protamine at concentrations up to 100 g/ml was added at Moreover, negatively charged moieties such as heparin do not significantly alter gene transfer efficiency. How-4°C there is no increase in permeability. 39 In our studies, an increase in gene transfer efficiency of similar magni-ever, heparin completely abolishes the effect of increased gene transfer caused by polycations. These data demon-tude is observed after exposure of cell monolayers to polycations for 2 min at 4°C before adenoviral adsorp-strate for the first time that charge interactions influence the efficiency of adenovirus-mediated gene transfer to tion. The time course and temperature independence of the polycation effect on gene transfer efficiency therefore both epithelial and endothelial cells, and suggest that modifying these interactions may be a useful strategy to speaks against increased permeability as a mechanism. Second and more importantly, the effects of polycations improve the efficiency of recombinant adenovirusmediated gene transfer to these cell types.

are also observed in subconfluent cell cultures in which a tight monolayer and cell polarization are not present. Recent work suggests that adenoviruses enter cells in two stages. The initial binding of the knob of fiber protein This argues against altered epithelial permeability as a mechanism for increased gene transfer. to an unknown receptor is followed by internalization that is mediated in part by ␣v␤3 and ␣v␤5 integrin recep-In summary, adenovirus-mediated gene transfer to epithelial and endothelial cells is modulated by the presence tors. 31 The fiber protein is a trimer of the 581 (serotype 5) or 582 (serotype 2) amino acid fiber polypeptides, 32 and of negatively charged cell surface molecules that appear to alter viral binding to cells. The addition of cationic contains at least one glucosamine residue. 33 Some evidence suggests that the interactions of fiber with the cell polymers before or during viral adsorption significantly increases the efficiency of gene transfer. This finding has involve amino acids rather than the carbohydrate moieties present on both the putative receptors and fiber. 34 important implications for the use of recombinant adenoviruses for gene therapy. Administration of a polycation Although the amino acid sequence of the fiber protein in adenovirus serotype 2 has been found to contain six with recombinant adenovirus may permit a decrease in the required quantity of virus, which may in turn impact negatively charged amino acids in the C-terminus, the exact conformation and net charge of this protein, and the on the host inflammatory response. 1,3 In addition, by determining and altering the binding sequence and/or precise binding site within the fiber knob remain unclear.

Epithelial and endothelial cell membranes contain gly-the charge of the fiber protein, adenovirus-mediated gene transfer may potentially be made more efficient, thereby coconjugates that are sialylated and/or sulfated, and thereby confer a negative charge to the cell surface. In improving the prospects for successful clinical application. airway epithelia, glycoproteins, such as the transmembrane mucin MUC1, and glycolipids, such as the gangliosides GM1 and GM2, are sialylated and expressed Materials and methods abundantly in the apical membrane. 35 Structurally similar glycolipids and glycoproteins, such as the endothelial mucins GlyCAM-1 and CD34, are expressed by endo-Production of recombinant adenoviruses Serotype 2, E1-deleted recombinant adenovirus contain-thelial cells. 36 Exposed on the lumenal surface, these anionic glycoconjugates are ideally situated to present a ing the E. coli lacZ reporter gene driven by the cytomegalovirus promoter (Ad2␤gal2) was constructed, purified, charge barrier to lumenally delivered gene transfer vectors. Thus, epithelial and endothelial cells have simi-and assayed for the number of plaque forming units (p.f.u.) per ml as previously described. 40 ,41 Viral stocks larities in cell structure that may constitute barriers to were stored in 5% sucrose and kept frozen at −80°C medium. In addition, adenovirus adsorption was performed before incubation with medium containing poly-until use.

brene to evaluate whether the effect of polybrene is specific to the viral adsorption phase. Isolation and primary culture of human bronchial epithelial (HBE) cells

In each experiment, wells that received Ad2␤gal2 alone were used to determine the baseline efficiency of gene HBE cells were isolated from native lungs of transplant recipients, as previously described. 42 In brief, airways transfer by recombinant adenovirus. Negative control experiments were done by incubating wells with media were dissected from surrounding adventitium, and placed in ice-cold HEPES-buffered minimum essential alone, and with media containing polyions without virus.

In each experiment, duplicate wells were used for each medium containing penicillin, streptomycin and amphotericin B. After incubation for 12-16 h at 4°C in 0.1% Pro-condition, and each experiment was performed at least twice for confirmation of results. As there was some vari-tease XIV (Sigma, St Louis, MO, USA), airway epithelial cells were obtained by gently scraping the epithelium ation from experiment to experiment due to minor differences in cell number, representative experiments are with the blunt end of a forceps. Supernatant from the washed tissue was spun, and the cell pellet was plated on shown. type IV human placental collagen (Sigma) coated tissue culture flasks in bronchial epithelial growth medium Detection of lacZ expression (BEGM; Clonetics, San Diego, CA, USA). Exempt

Forty-eight hours after viral adsorption, expression of the approval for the use of human lung tissue was obtained ␤-galactosidase transgene was determined by staining from the University of Pittsburgh Investigational the cells with 1 mg/ml of 5-bromo-4-chloro-3-indoyl-␤-Review Board.

galactopyranoside (X-gal; Boehringer Mannheim, Indianapolis, IN, USA) solution for 4 to 6 h. 42 The percentage Cell lines of transgene expressing (blue) cells in three representa-Type I Madin-Darby canine kidney (MDCK) cells were tive fields (Ͼ1000 cells per field) was determined by obtained from ATCC and grown in Dulbecco's minimum counting under inverted phase microscopy. essential medium (DMEM)/Ham's F12 supplemented with 3% FBS. Sheep pulmonary artery endothelial cells Statistical analyses (SPAEC) were obtained from collagenase-digested pul-

The mean percentage of X-gal positive cells in duplicate monary arteries, enriched by fluorescence activated cell wells under various conditions was analyzed for statistisorting of di-I-acetylated low density lipoprotein uptake 43 cal difference using ANOVA with Statview software and used from passage 12-15. Immortalized bronchial (Abacus Concepts, Berkeley, CA, USA). When statistical epithelial cells (IB3) were kindly provided by Dr Pamela differences were found, individual comparisons were Zeitlin (Johns Hopkins University, Baltimore, MD, USA) made using Fisher's PLSD post-hoc analysis. and grown in Ham's F12 supplemented with 10% FBS. The growth medium was changed every 3 days. MDCK

Effect of polycations on the early nant human ␣1-antitrypsin cDNA to human endothelial cells. stages of reovirus infection

Comparison of human

Protection of human endothelial cells from ian, and bovine rotaviruses for requirement of sialic acid in oxidant injury by adenovirus-mediated transfer of the human hemagglutination and cell adsorption

In vivo transfer of the human cystic fibrosis 26 Yolken RH et al. Sialic acid glycoproteins inhibit in vitro and in transmembrane conductance regulator gene to the airway epivivo replication of rotaviruses

In vivo adenovirus-mediated gene transfer mantling of adenovirus 2 during entry into cells

In vivo gene 28 Arcasoy SM et al. MUC1 and other sialoglycoconjugates inhibit transfer and expression in normal, uninjured blood vessels adenovirus-mediated gene transfer to epithelial cells. Am J using replication deficient recombinant adenovirus vectors

Target cell membrane sialic acid

Adenovirus-mediated transfer of the CFTR modulates both binding and fusion activity of influenza virus. gene to lung of nonhuman primates: biological efficacy study

Bovine coronavirus uses N-acetyl-9

Inefficient gene transfer by adenovirus vector acetylneuraminic acid as a receptor determinant to initiate the to cystic fibrosis airway epithelia of mice and humans. Nature infection of cultured cells

Adenovirus-mediated gene transfer to submu-␣v␤3 and ␣v␤5 promote adenovirus internalization but not cosal glands in vivo using xenografts

The sequence of adenovirus fiber: simi

A double-blind vehicle-controlled study of larities and differences between serotypes 2 and 5

adenoviral vector mediated gene transfer in the nasal epithelium

of patients with cystic fibrosis

The polypeptides of adenovirus. VI. 831. Early and late glycoproteins

Repeat administration of an adenovirus vector 34 Hennache B, Boulanger P. Biochemical study of KB cell receptor encoding cystic fibrosis transmembrane conductance regulator for adenovirus

In vivo adenovirus-mediated gene transfer via the pulmonary artery of rats

Mucins in the mainstream

Protamine sulfate as an effective 630-631. alternative to polybrene in retroviral-mediated gene transfer

meability of cultured epithelial monolayers

Optimization of retroviral-68

Correlation between polyion 39 Hammes M, Singh A. Effect of polycations on permeability of effect on cell susceptibility to in vitro infection with murine Cglomerular epithelial cell monolayers to albumin. J Lab Clin Med type viruses and polyion effect on some membrane related func-1994

Development and analysis of recombinant adeno

Effect of polycations on viruses for gene therapy of cystic fibrosis

/3T3 cells to murine leukemia and sarcoma

Characterization of an adenovirus gene

Enhancing effect of DEAE-dextran on the focustransfer vector containing an E4 deletion

forming titre of a murine sarcoma virus (Harvey strain)

Enhancement and inhibition of avian expression on bronchial epithelium after recombinant adenosarcoma viruses by polycations and polyanions

Collagen is a survival factor against LPS-induced

Interaction of respiratory syncytial virus with polyapoptosis in sheep pulmonary artery endothelial cells

Gerene Baldauff, and other members of the Lung Trans-Protamine sulfate was obtained from Eli Lilly plant team at the University of Pittsburgh for assisting (Indianapolis, IN, USA). Polybrene, poly-l-lysine, DEAEwith the procurement of human lung tissue, and to Dr dextran, sialic acid, heparin, and glycophorin A were Emily Yee for providing primary endothelial cell culobtained from Sigma.tures. We also thank David Turner and Dr Simon Watkins of the University of Pittsburgh Structural Biology

Imaging Center for their technical assistance. This work The number of cells in one well of a 24-well plate was was supported in part by the American Lung Association determined and Ad2␤gal2 dilutions were prepared in the (JMP), Cystic Fibrosis Foundation (Q933 to JMP and G320 growth medium appropriate for the cell type to obtain to BRP), and NIH HL 32154 (BRP). MOIs of 10, 25 or 100. Wells were washed with HBSS once, and viral adsorption was performed for 1 h at 4 or

37°C in a minimal volume to maximize contact of viral particles with cells. Plates were rocked every 15 min to 1 Crystal RG. Transfer of genes to humans: early lessons and assure even distribution of viral solution. immediately before adenoviral adsorption in growth