SHORT COMMUNICATION Adenovirus coxsackie adenovirus receptor-mediated binding to human erythrocytes does not preclude systemic transduction LA Rojas 1 , R Moreno 1 , H Calderón 2,3 and R Alemany 1 There is great skepticism in the capability of adenovirus vectors and oncolytic adenoviruses to reach specic organs or tumors upon systemic administration. Besides antibodies, the presence of CAR (coxsackie and adenovirus receptor) in human erythrocytes has been postulated to sequester CAR-binding adenoviruses, commonly used in gene therapy and oncolytic applications. The use of non-CAR-binding bers or serotypes has been postulated to solve this limitation. Given the lack of integrins in erythrocytes and therefore of internalization of the CAR-bound virus, we hypothesized that the interaction of adenovirus type 5 (Ad5) with CAR in human erythrocytes could be reversible. In this work, we have studied the effects of Ad5 interaction with human erythrocytes via CAR. Although erythrocyte binding was observed, it did not reduce viral transduction of tumor cells in vitro after long-term incubations. Transplantation of human erythrocytes into nude mice did not reduce Ad5 extravasation and transduction of liver and human xenograft tumors after systemic administration. These ndings indicate that despite human erythrocytes are able to bind to Ad5, this binding is reversible and does not prevent extravasation and organ transduction after systemic delivery. Thus, the poor bioavailability of systemically delivered CAR-binding adenoviruses in humans is likely due to other factors such as liver sequestration or neutralizing antibodies. Cancer Gene Therapy advance online publication, 21 October 2016; doi:10.1038/cgt.2016.50 INTRODUCTION Adenoviruses are broadly used as gene therapy vectors as well as oncolytic agents for the treatment of cancer. More than 50 human adenovirus serotypes have been described, 1 but particularly adenovirus type 5 (Ad5) has been the most widely used for these applications. In the case of disseminated cancer or when multiple organs need to be transduced, a systemic delivery of the virus is required by injection into the bloodstream. However, clinical experience has shown that local delivery has usually out- performed intravenous administration, which has rarely demon- strated therapeutic effects. 25 Ad5 encounters several important barriers when injected intravenously that may block transduction. Liver Kupffer cells and liver sinusoidal endothelial cells efciently sequester circulat- ing viral particles. 1 In addition, human blood represents a hostile environment for Ad5 because of multiple neutralizing interactions that reduce the bioavailability of the virus. Antibodies that occur naturally or develop upon infection directly block capsid infectivity or enhance its clearance by Fc receptors in phagocytes 6 and by complement receptor 1 in erythrocytes. 7 Independently of the presence of antibodies, human erythrocytes may also bind and inactivate Ad5 directly. Recent studies have demonstrated that unlike mouse erythrocytes, human erythrocytes express the primary adenovirus receptor, the Coxsackie and adenovirus receptor (CAR). 7,8 This CAR binding was shown to inhibit Ad5 infectivity, contrary to earlier reports of Ad5 titration in the presence of human blood. 9 However, the absence of integrins in erythrocytes precludes internalization of the erythrocyte-bound Ad5. 7 Therefore, we hypothesized that the interaction of Ad5 with CAR in human erythrocytes could be reversible, and upon long- term incubation periods, the virus could be available for infection. Our results support this hypothesis. MATERIALS AND METHODS Cell lines and viruses Human lung adenocarcinoma A549 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained with Dulbeccos modied Eagles medium supple- mented with 5% fetal bovine serum at 37 °C with 5% CO 2 . Cells were routinely tested for mycoplasma. AdGL is an Ad5-derived E1-deleted rst-generation vector expressing the EGFP-luciferase fusion protein cassette from pEGFPLuc (Clontech, Mountain View, CA, USA). AdGL was propagated in human embryonic kidney 293 cells and puried using cesium chloride double gradients according to standard techniques. Isolation of human erythrocytes Human blood samples were obtained by venipuncture into lithium-heparinized tubes (Greiner Bio-One, Monroe, NC, USA) with donor consent and approval by the IDIBELL's Ethics Committee. After centrifugation (2000 g for 10 min), plasma and leukocyte layer were removed, and erythrocytes were washed with phosphate-buffered saline (PBS) four times. Erythrocytes were resuspended in PBS at the indicated concentrations. 1 Virotherapy and Gene therapy Group, ProCure Program, Translational Research Laboratory, Instituto Catalan de Oncología-IDIBELL, Barcelona, Spain; 2 Psioxus Therapeutics, Oxford, UK and 3 Department of Oncology, University of Oxford, Oxford, UK. Correspondence: Dr R Alemany, Virotherapy and Gene Therapy Group, ProCure Program, Translational Research Laboratory, Instituto Catalan de Oncología-IDIBELL, Av Gran Via de lHospitalet 199-203, Via S/N Km2, 7, LHospitalet de Llobregat, Barcelona 08907, Spain. E-mail: ralemany@iconcologia.net Received 1 July 2016; revised 22 September 2016; accepted 23 September 2016 Cancer Gene Therapy (2016) 00, 1 4 © 2016 Nature America, Inc., part of Springer Nature. All rights reserved 0929-1903/16 www.nature.com/cgt