Gary Vellekamp

Empty Capsids in Column-Purified Recombinant Adenovirus Preparations

Empty capsids from adenovirus, that is, virus particles lacking DNA, are well documented in the published literature. They can be separated from complete virus by CsCl density gradient centrifugation. Here we characterize the presence of empty capsids in recombinant adenovirus preparations purified by column chromatography. The initial purified recombinant adenovirus containing the p53 tumor suppressor gene was produced from 293 cells grown on microcarriers and purified by passage through DEAE-Fractogel and gel-filtration chromatography. Further sequential purification of the column-purified virus by CsCl and glycerol density gradient centrifugations yielded isolated complete virus and empty capsids. The empty capsids were essentially noninfectious and free of DNA. Analysis of empty capsids by SDS-PAGE or RP-HPLC showed the presence of only three major components: hexon, IIIa, and a 31K band. This last protein was identified as the precursor to protein VIII (pVIII) by mass spectrometric analysis. No pVIII was detected from the purified complete virus. Analysis by electron microscopy of the empty capsids showed particles with small defects. The amount of pVIII was used to determine the level of empty capsid contamination. First, the purified empty capsids were used to quantify the relation of pVIII to empty capsid particle concentration (as estimated by either light scattering or hexon content). They were then used as a standard to establish the empty capsid concentration of various recombinant adenovirus preparations. Preliminary research showed changes in empty capsid concentration with variations in the infection conditions. While virus purification on anion-exchange or gel-filtration chromatography has little effect on empty capsid contamination, other chromatographic steps can substantially reduce the final concentration of empty capsids in column-purified adenovirus preparations.

Purification of Adenovirus

The advancement of adenoviral technology for delivery of therapeutic agents in human clinical trials has brought about the need for new methods for producing adenoviruses at large scale. This chapter discusses the physical properties of adenoviral particles, and the characteristics of the milieu from which they are purified. These considerations are essential to the design of large-scale recovery and purification methods. Purification processes using chromatography matrices are effective for both small- and large-scale purification as well as analysis. An understanding of the physical nature of the adenoviral particle is necessary to understand the choice of purification procedures and assay methods needed for following the process.

Complement component C1q and anti-hexon antibody mediate adenovirus infection of a CAR-negative cell line.

Realization of the potential clinical utility of recombinant adenovirus for gene therapy or vaccine development depends on a better understanding of the role of naturally occurring or therapy-induced anti-adenovirus antibodies. This study addresses the impact of anti-adenovirus neutralizing antibodies and the complement protein C1q on adenovirus infection of coxsackie and adenovirus receptor (CAR)-positive, and especially CAR-negative cells. Initially, transduction efficiency of adenovirus vectors was assessed in the presence or absence of human sera derived from healthy individuals that were seropositive for anti-adenovirus neutralizing antibodies. Infection was monitored by transgene expression in vitro using a replication-deficient adenovirus encoding green fluorescent protein (Ad-GFP). HeLa cells (CAR-positive) were readily infected by Ad-GFP and increasing concentrations of pooled sera increasingly inhibited infection. In contrast, rhabdomyosarcoma (RD) cells, a CAR-negative cell, were poorly infected by Ad-GFP. However, in the presence of human serum, robust GFP expression was observed. This expression was completely abrogated if the human serum was heat-inactivated. Addition of purified human C1q protein to the heat-inactivated serum restored GFP expression. Similar results were seen when human C1q protein was added to purified anti-hexon antibodies, but not to anti-fiber or anti-penton base antibodies, thus implicating anti-hexon antibodies as the infective antibody component of the human sera. These studies suggest that complement protein C1q and anti-hexon antibodies together can mediate efficient adenovirus infection in CAR-negative cell types.