Adenovirus-Mediated Erythropoietin Production by Airway Epithelia Is Enhanced by Apical Localization of the Coxsackie–Adenovirus Receptor in Vivo

Adenovirus-Mediated Erythropoietin Production by Airway Epithelia Is Enhanced by Apical Localization of the Coxsackie–Adenovirus Receptor in Vivo

Please download to get full document.

View again

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.

Graphics & Design

Publish on:

Views: 0 | Pages: 7

Extension: PDF | Download: 0

  Adenovirus-Mediated Erythropoietin Production byAirway Epithelia Is Enhanced by Apical Localization of theCoxsackie–Adenovirus Receptor   In Vivo  Benjamin Davis,* Jenny Nguyen,* David Stoltz, Dayna Depping,Katherine J.D. Excoffon, and Joseph Zabner  y  Department of Internal Medicine, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, 440 EMRB, Iowa City, IA 52242, USA *  These authors contributed equally to this work. y To whom correspondence and reprint requests should be addressed. Fax: (319) 335-7623. E-mail:  Available online 28 July 2004 In well-differentiated human airway epithelia, the coxsackie B and adenovirus types 2 and 5receptor (CAR) resides on the basolateral membrane. Replacing the transmembrane andcytoplasmic tail of CAR with a glycosyl-phosphatidylinositol anchor (GPI-CAR) allows apicallocalization of GPI-CAR, where it can bind adenovirus and enhance gene transfer   in vitro  . To testthis hypothesis further and to investigate requirements and barriers we developed an  in vivo  model that quantitatively assesses gene transfer of erythropoietin (EPO) to mouse airwayepithelia. Our data suggest that erythropoietin is secreted basolaterally, allowing possible accessto the bloodstream. The data also suggest that basolateral adenovirus-mediated airway epitheliaEPO secretion persists for long periods and could be used to study persistence  in vivo  .Additionally, the increase in hematocrit in response to the increased serum EPO could be used for therapeutic purposes. Finally, we tested the ability of apically localized CAR to enhance theinfection of AdEPO in mouse airway epithelia  in vivo  . The data suggest that apical receptors inairway epithelia may be sufficient to improve adenovirus infection of airway epithelia  in vivo  .Key Words: gene transfer, airway epithelia, adenovirus, EPO, erythropoietin, CAR I NTRODUCTION The first steps in adenovirus infection require primarilytwo proteins in the viral capsid: fiber and penton base[11,13,14]. The adenovirus fiber protein forms a trimer that binds to the cell via a high-affinity receptor, thecoxsackie virus B and adenovirus types 2 and 5 receptor(CAR) [4,30]. Recent structural and genetic studies sup- port a model in which the lateral cleft between twoneighboring knob domains on fiber interact with theextracellular amino-terminal IgV domain of CAR[5,9,26]. Interestingly, adenovirus-meditated gene trans- fer to lymphocyte and CHO cell lines does not require thewild-type transmembrane or cytoplasmic domains of CAR, suggesting that the interaction between fiber–knoband CAR mediates attachment to the cell surface[20,36,37]. In addition to the fiber–CAR interaction,the penton base interacts with  a V h 3  and  a V h 5  integrins,facilitating receptor-mediated endocytosis of adenovirus[14,21,40]. Thus, CAR is required for binding and infec- tion, and  a V h  integrins may function as co-receptors.Humanairwayepitheliaareatarget forgenetransferinthe genetic disease cystic fibrosis [27,38]. Earlier work showed that adenovirus infection and adenovirus-medi-ated gene transfer to differentiated airway epithelia isinefficient due to lack of CAR and integrins on the apicalmembrane [2,12,15,17,23–25,35,41,42]. Thus, lack of  fiber–knob binding to the apical membrane may be therate-limiting step for adenovirus-mediated gene transferto airway epithelia. Despite its absence on the apicalmembrane, CAR is present on the basolateral membrane[25,34]. Consequently, adenovirus efficiently infects air- way epithelia from the basolateral surface in a fiber-dependent manner [34].These results raised the question of whether CARlocalized to the apical membrane would be sufficientfor adenovirus-mediated gene transfer from the apicalsurface. Answering this question is important for under-standing the molecular mechanisms of adenovirus entryinto human airway epithelia. Walters  et al . recentlyshowed that apically localized CAR enhances adenovirusinfection in human airway epithelia  in vitro . To address  M OLECULAR  T HERAPY   Vol. 10, No. 3, September 2004 500 Copyright  B   The American Society of Gene Therapy1525-0016/$30.00 A RTICLE doi:10.1016/j.ymthe.2004.05.032  this question  in vivo,  we studied adenovirus-mediatedgene transfer in differentiated murine airway epitheliaexpressing recombinant CAR lacking the cytoplasmic andtransmembrane domains but modified with a glycosyl-phosphatidylinositol anchor signal sequence (GPI-CAR)to target the apical membrane [20,31,37]. To learn whether apically localized CAR facilitates gene transferto the airways  in vivo  and to investigate the mechanismsinvolved, we studied mice inf ected first intranasally withthixo-formulated AdGPI-CAR [29] and then infected withan adenovirus expressing erythropoietin (AdEPO). EPOwas used as a reporter gene because it allows the use of anoninvasive assay so that infected animals can be studiedover time [3]. R ESULTS Recombinant EPO Is Secreted via the Basolateral Sidein Human Airway Epithelia To determine if adenovirus-encoding RhEPO would me-diate expression by human airway epithelia and toinvestigate if EPO is secreted in a polar fashion, weinfected human airway epithelia with the AdEPO virusat a multiplicity of infection (m.o.i.) of 100 in 60  A l of athixotropic solution or by pretreating the epithelia withthe calcium chelator 8 mM EGTA for 3 h. We collectedapical washes, cell lysates, and basolateral media after 48h and measured EPO by ELISA. Our data show that, asexpected, human airway epithelia do not make RhEPOand that both pretreatment with EGTA and the use of athixotropic solution resulted in high levels of Ad-medi-ated EPO expression. Although EGTA treatment resultedin higher RhEPO levels (up to twofold) than the thixo-tropic solution (Fig. 1A), we used the thixotropic solu- tion throughout these studies because it does not requiredisruption of the tight junction. Finally, we found alinear dose response in basolateral levels of RhEPO as wechanged the m.o.i. of AdRhEPO (Figs. 1B and 1C). These data suggest that adenovirus-mediated expression of RhEPO in airway epithelia would result in basolateralsecretion of RhEPO, giving it access to the bloodstreamwhere it could be measured as a reporter protein andcould mediate production of red blood cells. Adenovirus-Mediated Expression of EPO by HumanAirway Epithelia Persists over Time To test for the persistence of adenovirus-mediated EPOexpression by human airway epithelia  in vitro,  we addedAdRhEPO at an m.o.i. of 100 in a thixotropic solution tothe apical side for 3 h and assayed the basolateral medi-um for EPO by ELISA every 3 days. The basal medium waschanged every time it was sampled, thus the measure-ments reflect EPO made over 3 days prior to mediumcollection. Fig. 2 shows that we detected high levels of EPO in the basolateral medium by day 3 after infectionand that the levels peaked on day 14. Expression of EPOremained high for up to 25 days. Since we used first-generation Ad vectors, the decline in EPO expressionshown might be the result of Ad-associated toxicity.These data suggest that basolateral Ad-mediated airwayepithelia EPO secretion persists for long periods andcould be used to study persistence  in vivo . Adenovirus-Mediated Expression of EPO in the LungsResults in Increased Hematocrit in Mice To test adenovirus-mediated expression of EPO in airwaysand liver  in vivo,  we infected mice intranasally withAdRhEPO, AdmEPO, and AdGFP in a thixotropic solution FIG. 1.  Adenovirus-mediated expression of rhesus monkey EPO by humanairway epithelia. (A) The polarity of RhEPO secretion of primary cultures of human airway. Epithelia were infected apically with AdRhEPO (m.o.i. of 100)using either a thixotropic solution (Thixo) or EGTA. Forty-eight hourspostinfection an apical wash, the basal medium, and the cell lysate wereassayed for the presence of RhEPO. (B) The effects of dose of AdEPO in athixotropic solution applied to the apical surface without disruption of tightjunctions. Very low levels of RhEPO were detected in the medium only withthe highest doses of AdRhEPO. (C) The effects of dose of AdRhEPO applied tothe basolateral surface of airway epithelia (note the difference in scale).  M OLECULAR   T HERAPY   Vol. 10, No. 3, September 2004 501 Copyright  B   The American Society of Gene Therapy doi:10.1016/j.ymthe.2004.05.032  A RTICLE  and via tail vein injection with AdRhEPO and AdmEPO.Intranasal administration of adenovirus in a thixotropicsolution results in airway delivery and expression of EPOmRNA predominantly in the lungs, whereas intravenousadministration targets the liver (Fig. 3). We collectedblood via eye bleeds and centrifuged it to measure thehematocrit. Fig. 4 shows that adenovirus-mediated mEPOand RhEPO resulted in increased hematocrit when ad-ministered either intravenously or intranasally. Althoughintranasal instillation of adenovirus in a thixotropicsolution resulted in a smaller increase in hematocritcompared to intravenous delivery, this difference wassmall.Consistent with our  in vitro  experiments, the increasein hematocrit in animals treated with AdmEPO lastedfor as long as we measured it (50 days). However,AdRhEPO induced an increase in the hematocrit thatreturned to baseline after 20 days and continued to dropbelow baseline levels after 50 days. These data suggestthat adenovirus-mediated expression of homologousmEPO in the lungs can result in a prolonged increasein hematocrit. This increase in hematocrit could be usedfor therapeutic purposes or as a simple end point forquantitative assessment of gene transfer to the airwayepithelia. Expression of Rhesus Monkey EPO in Mice Results inNeutralizing Ab to EPO Adenovirus expression of RhEPO in murine lungs andliver resulted in a transient increase in hematocrit andeventually the appearance of anemia in the mice. Wehypothesized that an immune-mediated response to theRhEPO that neutralized its activity and broke toleranceagainst the endogenous mEPO could be responsible. Thepossibilities include a cell-mediated immune responsethat destroys the cells making EPO or a humoral neutral-izing antibody response that inactivates EPO function. Todistinguish between these possibilities, we first measuredthe serum EPO concentrations via ELISA. Fig. 5 showsthat serum EPO concentrations in mice infected withAdRhEPO remained elevated for over 25 days despitethe fact that these mice were already anemic. These dataexclude the possibility of a T-cell-mediated destruction of cells making EPO. We then tested for the presence of antibodies directed against a commercially available hu-man EPO (hEPO) by Western blot on sera from uninfect-ed mice and at 25 days after AdRhEPO gene transfer. Fig.6 shows that mice exposed to AdRhEPO via either theintranasal or the intravenous route, but not mice exposedto AdGFP, developed antibodies that recognize hEPO.Antibodies were not detected in pre-infection sera (datanot shown). These data suggest that the adenovirus- FIG. 3.  Intranasal administration of AdRhEPO resulted in increased mRNAexclusively in the lung. Mice received 10 8 IU of AdRhEPO intravenously or byintranasal administration in a thixotropic solution. Lung and liver tissuesamples were removed 8 days after infection for RT-PCR. Mouse  h -actin wasdetected for control. RT-PCR products stained with ethidium bromide areshown. FIG. 4.  Adenovirus-mediated gene transfer of mEPO or RhEPO resulted inincreased mouse hematocrit. Mice received 10 8 IU of AdRhEPO intravenously(closed circles) or by intranasal administration in a thixotropic solution (opencircles). A second set of mice received 10 8 IU of AdmEPO intravenously (closedsquares) or by intranasal administration in a thixotropic solution (opensquares). Control mice received 10 8 IU of AdGFP by intranasal administrationin a thixotropic solution (open triangles). Hematocrit was measured every 6days for 7 weeks. Data are means  F  SEM ( n  = 4). FIG. 2.  Persistence of adenovirus-mediated expression of rhesus monkey EPOby human airway epithelia. Epithelia were infected apically with AdRhEPO(m.o.i. of 100) using a thixotropic solution. RhEPO concentrations weredetermined in the basolateral medium every 3 days for 4 weeks.  M OLECULAR  T HERAPY   Vol. 10, No. 3, September 2004 502 Copyright  B   The American Society of Gene Therapy A RTICLE doi:10.1016/j.ymthe.2004.05.032  mediated expression of RhEPO in mice activates an anti-body response that recognizes RhEPO and cross reactswith hEPO and mEPO. Taken together, the data suggestthat these antibodies neutralize the activity of bothRhEPO and endogenous mEPO. Apical Localization of CAR EnhancesAdenovirus-Mediated Gene Transfer from the ApicalSurface  in Vivo To determine if apical localization of CAR in murineairway epithelia  in vivo  is sufficient for adenovirus infec-tion, we investigated adenovirus-mediated gene transferfrom the apical surface of mouse airways that had beenpreviously induced to express GPI-CAR (Fig. 7). We trea- ted mice with 10 8 IU per mouse of AdGPI-CAR in athixotropic solution by nasal instillation. AdGFP wasinstilled as a negative control. Three days later, we treatedthe mice with 10 8 IU of AdmEPO. At this dose, we andothers have shown only minimal amount of gene transferinto naı¨ve mice. AdGFP showed minimal change inhematocrit. However, GPI-CAR expression substantiallyincreased the hematocrit after a low dose of AdEPO.These results indicate that expression of GPI-CAR inairway epithelia is sufficient to increase the efficiency of adenovirus infection of airway epithelia from the apicalsurface. D ISCUSSION In determining the quality of the AdEPO reporter systemin the murine model, our data suggested that erythropoi-etin is secreted basolaterally, allowing possible access tothe bloodstream. The data also support that basolateralAd-mediated airway epithelia EPO secretion persists forlong periods and could be used to study persistence  invivo . Additionally, the increase in hematocrit in response FIG. 5.  Adenovirus-mediated gene transfer of RhEPO resulted in increasedmouse serum EPO concentration. Mice received 10 8 IU of AdRhEPOintravenously (closed circles) or by intranasal administration in a thixotropicsolution (open circles). Control mice received 10 8 IU of AdGFP by intranasaladministration in a thixotropic solution (open triangles). EPO concentration inserum by ELISA was measured. Data are means  F  SEM ( n  = 4). FIG. 6.  Expression of rhesus monkey EPO in mice results in anti-EPO antibodies. We tested for the presence of antibodies directed against EPO by Western blot onsera from mice before and 25 days after AdRhEPO gene transfer. Recombinant hEPO is recognized by an anti-EPO Ab. Mice exposed to AdRhEPO via either theintranasal or the intravenous route, but not mice exposed to AdGFP, developed antibodies that recognize hEPO. Antibodies were not detected in preinfectionsera (data not shown). FIG. 7.  Expression of GPI-CAR by mouse airway epithelia  in vivo  results inincreased adenovirus-mediated gene transfer of mEPO. Mice were treatedwith 10 8 IU per mouse of AdGPI-CAR in a thixotropic solution by nasalinstillation. AdGFP was instilled as a negative control. Three days later, micewere treated with 10 8 IU of AdmEPO. Hematocrit was measured 7 days post- AdmEPO infection. In mice expressing GPI-CAR in the airways AdmEPOsubstantially increased the hematocrit. These results indicate that expressionof GPI-CAR in airway epithelia is sufficient for adenovirus-mediated genetransfer from the apical surface. Data are means  F  SEM ( n  = 6).  M OLECULAR   T HERAPY   Vol. 10, No. 3, September 2004 503 Copyright  B   The American Society of Gene Therapy doi:10.1016/j.ymthe.2004.05.032  A RTICLE  to the increased serum EPO could be used for therapeuticpurposes or as a simple end point for quantitative assess-ment of gene transfer to the airway epithelia [3]. Impor- tantly, these data also suggest that expression of foreignEPO activates an antibody response that neutralizes theactivity of both the exogenous EPO and endogenous EPO.Thus, species-specific EPO is important in this EPO re-porter system.The data show that placing a specific binding sitefor adenovirus (GPI-CAR) at the apical surface of airwayepithelia via a GPI anchor facilitates adenovirus-medi-ated gene transfer  in vivo . CAR is located in theadherens junctions of epithelia where it functions asa cell adhesion molecule [35]. We have shown that it may play a role in allowing wild-type adenovirus es-cape to the lumen of the airways [33]. Whereas it is still unclear how the primary infection occurs, our datasuggest that absence of apical CAR receptor is themajor hindrance for infection. The apical surface of airway epithelia expresses low levels of   a v  h 3  and  a v  h 3 integrins [12]. Thus, these data further suggest thatcoreceptors may not be required for adenovirus airwayepithelial infection. These data are also consistent withthose obtained from  in vivo  studies in the liver [16].GPI-anchored proteins localize to lipid rafts [32], which may be involved in endocytic sorting [22]. In addition, GPI proteins can be found in lipid raft domains inclose proximity to caveolae or inside the caveolae[1,28]. Hence, the activity of GPI-CAR might be aug-mented by its location in areas of the apical membranewith active internalization (endocytosis, potocytosis),thereby masking the need for integrins. In either event,these results bring into question the need for corecep-tors to facilitate internalization, at least when thereceptor is abundant or localized in an area of activemembrane turnover. Therefore, targeting strategies toairway epithelia may not be hindered by the lack of coreceptors. Whereas in some epithelia the glycocalyxmay impede adenovirus infection, our data indicatethat in human airway epithelia  in vitro , or in murineairway epithelia  in vivo , this glycocalyx is not ratelimiting.In conclusion, expressing CAR on the apical surface byGPI modification rescues adenovirus binding and genetransfer from the apical surface of airway epithelia  in vivo .These data suggest that targeting binding sites on theapical surface will enhance gene transfer. This conclusionis consistent with other approaches to enhance genetransfer. For example, increasing nonspecific apical bind-ing of adenovirus by incorporation in CaP i  coprecipitatesor complexes including cationic lipids facilitated genetransfer [7,8]. Moreover, targeting adenovirus to another GPI-linked protein, the urokinase plasminogen activatorreceptor, or to P2Y receptors enhanced gene transfer[6,19], suggesting that numerous different approaches of increasing binding may be sufficient to improve genetransfer. We predict that targeting vectors to high-affinityreceptors capable of internalization will result in efficientgene transfer to the airway epithelia. E XPERIMENTAL  P ROCEDURES Human Airway Epithelia Airway epithelial cells were obtained from trachea andbronchi of lungs removed for organ donation. Cells wereisolated by enzyme digestion as previously described[18,43]. Freshly isolated cells were seeded at a density of  5    10 5 cells/cm 2 onto collagen-coated, 0.6-cm 2 -areaMillicell polycarbonate filters (Millipore Corp., Bedford,MA, USA). The cells were maintained at 37 j C in a humid-ified atmosphere of 5% CO 2  and air. Twenty-four hoursafter plating, the mucosal medium was removed and thecells were grown at the air–liquid interface [18,43]. The culture medium consisted of a 1:1 mix of DMEM:Ham’sF12,5%UltroserG(BiosepraSA,Cedex,France),100U/mlpenicillin, 100  A g/ml streptomycin, 1% nonessential ami-no acids, and 0.12 U/ml insulin. Airway epithelia reachedconfluence and developed a transepithelial electrical re-sistance, indicating the development of tight junctionsand an intact barrier. Epithelia were allowed to differen-tiate by culturing for at least 14 days after seeding and thepresence of a ciliated surface was tested by scanningelectron microscopy [43]. Since airway epithelia do not expressCARontheapicalsurface,weinitiallyinfectedtheepithelia with adenovirus by either pretreating epitheliawith 8 mM EGTA delivered in H 2 O to transiently disruptthe tight junctions [35] or using a thixotropic solutionthat enhances viral contact time [29]. Two days after infection with CAR-expressing adenoviruses, epithelialintegritywas measured with an ohmmeter (EVOM; WorldPrecision Instruments, Inc., Sarasota, FL, USA). The trans-epithelial resistance values for all infected epithelia were>300  V    cm 2 . Epithelia were then evaluated or studied asdescribed below. We exposed the epithelia to recombi-nant adenovirus-expressing EPO constructs at an m.o.i. of 0.01to100for3h. ToexpressGPI-CARor EPOweutilizedan adenoviral vector with a CMV promoter. Cloning Mouse and Rhesus EPO Kidney tissues of mice and Rhesus monkeys were flashfrozen in liquid nitrogen. RNA was isolated using RNA-STAT as previously described. Reverse transcriptase PCR(Invitrogen Superscript One-Step RT-PCR, Carlsbad, CA,USA) was performed to isolate the EPO cDNA. The EPOmRNA was ligated into Invitrogen’s pcDNA3.1/CT-GFP-TOPO plasmid via the TOPO technology. This plasmidwas sequenced by the University of Iowa DNA Se-quencing Core and tested for expression in COS cells.Adenovirus vectors expressing either mEPO (mouse) orRhEPO (rhesus) were constructed as previously de-scribed [39].  M OLECULAR  T HERAPY   Vol. 10, No. 3, September 2004 504 Copyright  B   The American Society of Gene Therapy A RTICLE doi:10.1016/j.ymthe.2004.05.032
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks