To develop adenoviral cell- or tissue-specific gene delivery, understanding of the infection mechanisms of adenoviruses is crucial. Several adenoviral attachment proteins such as CD46, CAR and sialic acid have been identified and studied. However, most receptor studies were performed on non-human cells. Combining our reporter gene-tagged adenovirus library with an in vitro human gene knockout model, we performed a systematic analysis of receptor usage comparing different adenoviruses side-by-side. The CRISPR/Cas9 system was used to knockout CD46 and CAR in the human lung epithelial carcinoma cell line A549. Knockout cells were infected with 22 luciferase-expressing adenoviruses derived from adenovirus species B, C, D and E. HAdV-B16, -B21 and -B50 from species B1 as well as HAdV-B34 and -B35 were found to be CD46-dependent. HAdV-C5 and HAdV-E4 from species E were found to be CAR-dependent. Regarding cell entry of HAdV-B3 and -B14 and all species D viruses, both CAR and CD46 play a role, and here, other receptors or attachment structures may also be important since transductions were reduced but not completely inhibited. The established human knockout cell model enables the identification of the most applicable adenovirus types for gene therapy and to further understand adenovirus infection biology.

Carcinomas are characterized by a widespread upregulation of intercellular junctions that create a barrier to immune response and drug therapy. Desmoglein 2 (DSG2) represents such a junction protein and serves as one adenovirus receptor. Importantly, the interaction between human adenovirus type 3 (Ad3) and DSG2 leads to the shedding of the binding domain followed by a decrease in the junction protein expression and transient tight junction opening. Junction opener 4 (JO-4), a small recombinant protein derived from the Ad3 fiber knob, was previously developed with a higher affinity to DSG2. JO-4 protein has been proven to enhance the effects of antibody therapy and chemotherapy and is now considered for clinical trials. However, the effect of the JO4 mutation in the context of a virus remains insufficiently studied. Therefore, we introduced the JO4 mutation to various adenoviral vectors to explore their infection properties. In the current experimental settings and investigated cell lines, the JO4-containing vectors showed no enhanced transduction compared with their parental vectors in DSG2-high cell lines. Moreover, in DSG2-low cell lines, the JO4 vectors presented a rather weakened effect. Interestingly, DSG2-negative cell line MIA PaCa-2 even showed resistance to JO4 vector infection, possibly due to the negative effect of JO4 mutation on the usage of another Ad3 receptor: CD46. Together, our observations suggest that the JO4 vectors may have an advantage to prevent CD46-mediated sequestration, thereby achieving DSG2-specific transduction.