Retrovirus-Polymer Complexes: Study of the Factors Affecting the Dose Response of Transduction Natalia Landa ´ zuri, Delfi Krishna, Monique Gupta, and Joseph M. Le Doux* The Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory University, Atlanta, Georgia 30332-0535 We have previously shown that complexes of Polybrene (PB), chondroitin sulfate C (CSC), and retrovirus transduce cells more efficiently than uncomplexed virus because the complexes are large and sediment, reaching the cells more rapidly than by diffusion. Transduction reaches a peak at equal weight concentrations of CSC and PB and declines when the dose of PB is higher or lower than CSC. We hypothesized that the nonlinear dose response of transduction was a complex function of the molecular characteristics of the polymers, cell viability, and the number of viruses incorporated into the complexes. To test this hypothesis, we formed complexes using an amphotropic retrovirus and several pairs of oppositely charged polymers and used them to transduce murine fibroblasts. We examined the effect of the type and concentration of polymers used on cell viability, the size and charge of the complexes, the number of viruses incorporated into the complexes, and virus binding and transduction. Transduction was enhanced (2.5- to 5.5-fold) regardless of which polymers were used and was maximized when the number of positive charge groups was in slight excess (15-28%) of the number of negative charge groups. Higher doses of cationic polymer were cytotoxic, whereas complexes formed with lower doses were smaller, contained fewer viruses, and sedimented more slowly. These results show that the dose response of transduction by virus-polymer complexes is nonlinear because excess cationic polymer is cytotoxic, whereas excess anionic polymer reduces the number of active viruses that are delivered to the cells. Introduction Recombinant retroviruses are frequently used for experimental and clinical gene transfer because they permanently modify the genome of target cells (1, 2), transfer genes to a wide variety of human cell types (3-5), and are relatively straightforward to construct and produce (6, 7). Nevertheless, recombinant retroviruses have had limited success in clinical gene therapy protocols, in part because it has proven difficult to predict and control the level of gene transfer (8, 9), and in many cases the transduction efficiency is too low to achieve the desired therapeutic effect (10-12). Recent work suggests that there are at least two significant causes for low and variable levels of retrovirus-mediated gene transfer. First, retroviruses are large particles that diffuse slowly but decay rapidly (with a half-life of about 7 h). As a result, most retroviruses are no longer infectious by the time they reach the cell surface (13, 14). Second, retrovirus stocks contain variable and often significant levels of inhibitors that block transduction (15-18). These findings suggest that the efficiency, predictability, and reproducibility of retrovirus-mediated gene transfer can be significantly increased by the development of improved methods for rapidly purifying viruses and delivering them to cells. We have previously found that gene transfer is enhanced by the combined addition of oppositely charged polymers such as Polybrene (PB) and chondroitin sulfate C (CSC) to retrovirus stocks (19). Gene transfer is increased primarily because the viruses become part of a polymer complex that can be rapidly purified from inhibitors of transduction and that rapidly sediment onto the surface of the tissue culture dish, significantly increas- ing the number of active viruses that reach the cells (20, 21). Interestingly, the dose-response curve of gene transfer by the virus-polymer complexes is a nonlinear function of the ratio of the concentrations of the two polymers. For example, we have previously observed that when viruses are mixed with a constant concentration of PB (80 µg/mL) and an increasing concentration of CSC (0-200 µg/mL), gene transfer increases sharply, reaches a peak when the concentration (80 µg/mL) of CSC equals that of PB, and then declines as the concentration of CSC begins to exceed that of PB (19). The factors that control the dose response of gene transfer as a function of polymer concentration are not known. One important factor may be the physicochemical properties of the polymers. For example, Davis et al. showed that the extent to which poly-L-lysine enhances retrovirus transduction is a strong function of its molecular weight (22). Another important factor may be the effect of the polymers on the viability of the cells since high doses of cationic polymers are often cytotoxic (19, 23). A third factor may be the extent to which the viruses are captured by the complexes, since an excess of anionic polymer might interfere with the ability of the viruses to bind to the complexes, which would in turn reduce the rate at which the viruses are delivered to the cells. To distinguish among these possibilities, we tested the hypothesis that the dose response of transduction as a function of polymer concentration is a complex function of the physi- cochemical characteristics of the polymers that are used, the * To whom correspondence should be addressed. Tel: 404-385-0632. Fax: 404-894-4243. E-mail: joe.ledoux@bme.gatech.edu. 480 Biotechnol. Prog. 2007, 23, 480-487 10.1021/bp060336y CCC: $37.00 © 2007 American Chemical Society and American Institute of Chemical Engineers Published on Web 03/21/2007