Background
Viruses bind to specific cellular receptors to infect their hosts. The specific receptors a virus uses are important factors in determining host range, cellular tropism, and pathogenesis. HIV-1 is one of the best characterized viruses in terms of viral entry. HIV-1 first binds to CD4, its primary receptor [
1,
2]. Although CD4 binding was initially thought to be sufficient for infection, it was later found that a second interaction between HIV and chemokine co-receptors CCR5 or CXCR4, is also required [
3‐
5]. Binding to CD4 occurs first, triggering conformational changes in the HIV protein gp120, revealing the previously hidden binding site for its co-receptors, which then trigger membrane fusion [
6,
7]. The discovery of HIV's requirement for co-receptors in addition to CD4 represented a significant shift in our understanding of viral entry. The idea that a single virus bound to a single entry receptor was replaced with the idea that viral entry is the result of distinct sequential events requiring multiple surface proteins.
In keeping with this multistep entry model, adenoviruses have been proposed to use a primary receptor to mediate binding and co-receptors to mediate internalization [
8]. Adenoviruses are non-enveloped double stranded DNA viruses associated with respiratory disease, ocular disease, and gastroenteritis [
9]. Adenoviruses have three major capsid proteins: hexon, which forms the bulk of the capsid and is present in 240 copies, penton base, which is present in five copies at each of the twelve vertices, and fiber, a homotrimeric protein that protrudes from each vertice, extending outward from the penton base. More than 50 human serotypes of adenovirus have been identified to date [
10,
11]. The best studied of these are the species C adenoviruses, including Adenovirus Serotype 2 (Ad2) and Adenovirus Serotype 5 (Ad5). The primary receptor for species C adenoviruses is thought to be Coxsackie and Adenovirus Receptor (CAR), which binds to the globular knob domain of fiber [
12]. This high affinity interaction docks the virus to the cell, thus allowing secondary interactions to occur. Following fiber binding to CAR, the penton base engages αvβ3 and αvβ5 integrins to initiate endocytosis and viral entry [
8]. Adenoviruses bind to integrins via an RGD motif present in the penton base. The penton base-integrin interaction is proposed to be exclusively involved in virus internalization and not to contribute to virus binding [
8].
Several studies have reported alternate mechanisms for adenovirus entry. Huang et al demonstrated that adenovirus binds to hematopoietic cells via a penton base interaction with Integrin αMβ2, an integrin not expressed on epithelial cells, but still requires αv integrins for virus internalization [
13]. Additionally, Ad5 has also been proposed to use heparan sulphate glycosaminoglycans as receptors [
14,
15] and to use lactoferrin as a bridge between viral particles and the cell surface [
16,
17]. In both of these systems, adenovirus fiber is the viral protein required for binding. Further complication is observed in vivo. Infection of liver cells, which has been well characterized, is CAR-independent and instead depends on adenovirus hexon binding the blood coagulation factor F(X) which leads to infection [
18‐
24]. Additionally, in both mice and non-human primates, adenoviruses with mutant fibers ablated for CAR binding show a similar biodistribution compared to wild type viruses [
18,
21,
25,
26]. Similarly, a lack of correlation between CAR expression and adenovirus infection has been observed in cell lines, though the mechanism by which infection of cells with low CAR is achieved is undefined [
12,
27,
28].
We also observe a lack of correlation between CAR expression and infection in cell lines. Indeed, we report here that an Ad5 vector can efficiently transduce cancer cells which express little to no CAR. Further, Ad5 binds to these cells via integrin αvβ5, a surface protein previously thought to be used exclusively for internalization and thus classified as a secondary co-receptor. These observations lead us to further question the two-step model for adenovirus infection, in which adenovirus must first bind to CAR, the primary receptor, in order to bind to integrins and trigger viral entry.
Discussion
In this study we report that cells which do not express CAR can be efficiently transduced by Ad5. Adenovirus entry is not dependent on fiber binding to cells but instead is blocked by an RGD peptide that interferes with the RGD domain on the adenovirus penton base binding cellular integrins. Further, we find that binding to low CAR cells is inhibited specifically by a blocking antibody to integrin αvβ5, demonstrating that integrin αvβ5 is required for Ad5 attachment to these cells. The binding mediated by integrin αvβ5 is extremely high affinity, in the picomolar range. Our data further challenges the prevailing model of adenovirus infection, in which binding to a primary receptor, CAR, is required in order for subsequent interactions between adenovirus and integrins to initiate viral entry.
Ad5 binding to low CAR cells is independent of fiber and instead depends on an interaction with integrin αvβ5. Therefore, Ad5 does not require an independent binding receptor to dock it to the cell before it can interact with internalization receptors. Other viruses are also reported to use both primary binding and internalization receptors. HIV-1 first binds to CD4 followed by binding to the chemokine receptors CCR5 or CXCR4, which trigger membrane fusion [
49]. Binding to CD4 induces conformational changes in the HIV protein gp120, revealing the previously hidden binding site for its coreceptors [
6]. Variants with mutations in gp120 allowing for direct interaction with coreceptors have been isolated
in vitro; however, these variants are sensitive to neutralizing antibodies and therefore selected against
in vivo[
50]. Therefore, the role of CD4 binding in HIV-1 infection may be particularly critical in evading the immune system of the host.
Unlike HIV-1 binding to CD4, Ad5 binding to CAR does not induce conformational changes in viral proteins, thus facilitating subsequent entry steps [
51]. Rather, CAR is thought to facilitate a high affinity interaction to fiber, thus docking the virus to the cell surface and allowing the subsequent interaction between the penton base and integrins to initiate internalization. Indeed, only the extracellular domain of CAR is required for CAR-mediated adenovirus entry [
52]. However, previous studies have observed CAR-independent infection and the work in this paper demonstrates that Ad5 can bind directly to integrin αvβ5, previously identified as an internalization receptor [
13,
14,
17,
22,
45]. Therefore, our results and the results of others suggest this initial binding step is not required for entry. Nevertheless, CAR binding is conserved in a number of adenovirus serotypes and the fiber-CAR interaction is one that is well characterized and is itself of high affinity [
10,
53]. Therefore, CAR binding clearly plays an important role in the adenovirus infection cycle. One possibility is that, even when CAR is not needed to enter cells, CAR functions as an exit receptor [
54]. Walters et al report that when Ad5 lyses a cell, excess fiber is released and through binding to CAR, disrupts neighboring cell-cell junctions, allowing for release of the virus back to the apical surface where it may continue infecting cells [
54]. Additionally, recent work has demonstrated that adenovirus binding to CAR may induce downstream signaling events that increase integrin activation, thus promoting infection[
55]. Therefore, binding to CAR may facilitate entry in ways beyond simply docking the virus to the cell surface. Indeed, at least two serotypes of adenovirus, Ad9 and Ad37, have fibers which bind CAR but do not use CAR as an attachment receptor, supporting the idea that CAR binding is important for steps other than attachment [
10,
56]. Further, it is likely advantageous to the virus to be able to use multiple entry routes, enabled by its ability to engage multiple different receptors.
Both HIV, as evidenced by CD4-independent variants isolated
in vitro, and Ad5, as evidenced by our results and the results of others can infect cells without binding to their so-called primary receptors. Binding to these receptors, instead of being strictly required for infection, may contribute to other necessary parts of the virus infection cycle, such as evading the host immune system or facilitating virus escape. Many other viruses with less characterized receptors seem to also use multiple receptors, some classified as binding receptors [
57]. For example, rotaviruses are thought to first bind to a sialic acid (SA)-containing molecule, which anchors the virus to the cell, and then bind to coreceptors to initiate viral entry [
57]. Mutant variants of rotaviruses that are SA-independent and interact directly with coreceptors have been isolated
in vitro, suggesting that similarly to HIV and Ad5, binding to the primary receptor is not strictly required for infection [
58,
59]. Therefore, the interaction between rotaviruses and SA-containing molecules may facilitate an as yet unidentified aspect of rotavirus infection. As more functional roles of virus receptors in infection are elucidated, the use of binding receptors in other aspects of the viral life cycle may emerge as a general principle of viral pathogenesis.
In addition to being used as a model system for viral entry, much effort has been put into developing adenoviruses, especially species C adenoviruses including Ad5, as vectors for gene therapy. In fact, adenoviral vectors have been used in more than one quarter of gene therapy trials worldwide [
60]. Cancer is one of the most common targets of adenovirus-mediated gene therapy. As mentioned previously, CAR expression is often lost as cancers progress and this loss has been viewed as a major hurdle to using adenovirus-based therapies in cancer [
31,
33‐
36]. However, integrin αvβ5 has been reported to be overexpressed in cancers [
61,
62]. Therefore, our conclusion that Ad5 can use integrin αvβ5 to bind to and infect cells lacking CAR suggests that cancer cells having lost CAR expression may still be good targets for adenovirus-based therapies. Recent work has shown that erythrocytes sequester adenovirus by binding CAR, thus limiting systemic infection; therefore, using CAR-ablated vectors, a strategy many groups are attempting, may improve delivery for gene therapy for reasons beyond changing receptor interactions [
63,
64]. We also observed what may be an as yet unidentified obstacle to these therapies, however. T47D cells, which express CAR (Figure
2a) and integrin αvβ5 (data not shown) are still resistant to Ad5 infection (Figure
1). Wang et al showed the cellular protein CEACAM6 blocks adenovirus trafficking to the nucleus in human pancreatic cancer cell lines[
65]. Future studies to determine if this protein blocks infection in T47D cells, or if resistance is due to a novel mechanism are needed.
Methods
Cell Lines and Viruses
SkMel2 and WM278 cells are human melanoma cell lines. MCF7, MDA-MB-453, BT549, T47D, MDA-MB-435, and MDA-MB-231 are human breast cancer cell lines. All cells were obtained from the ATCC. SkMel2 cells were cultured in MEM supplemented with sodium pyruvate, non-essential amino acids, and 10% FBS. WM278 were cultured in DME-H16 and supplemented with 10% FBS. All other cancer cell lines were cultured in DMEM supplemented with 10% FBS. Virus used was replication incompetent E1A deleted and expressed GFP (Ad5-GFP). Virus was propagated in HEK293/E4/pIX cells and harvested by CsCl gradient ultracentrifugation as previously described [
66,
67]. Virus titers were determined as previously described [
68].
Antibodies and Peptides
The MAb RmcB was used to detect CAR expression [
12]. The MAbs LM609, P1F6, and JB1A directed against integrins αvβ3, αvβ5, and the β1 subunit respectively were purchased from Chemicon. The secondary antibody Alexa 488 was purchased from Molecular Probes. The synthetic peptides GRGDSP and GRGESP were purchased from Sigma. All antibodies were IgG.
Recombinant Fiber
Full length Ad5 fiber was cloned from a Gateway entry vector into a his-tagged destination vector using the Gateway system per manufacturer's instructions (Invitrogen). Fiber was transformed into and grown up in BL21 Star (DE3) E.coli (Invitrogen). Overnight starter culture was diluted 1:100 in LB/amp and grown until bacteria reached log phase. 50 uM IPTG was added and bacteria were grown at room temperature overnight. Pellets were disrupted using Bugbuster (Novagen) per manufacturer's instructions. Fiber was purified via its his-tag by incubation with Probond resin (Invitrogen), several washes with 20 mM Imidizole, and elution using Poly-Prep Chromatography Columns (BioRad) with 0.2 M Imidizole. Purified recombinant fiber was then dialyzed into PBS for use in experiments. Approximately 1 mg/L of purified fiber was obtained.
Cell Infection Assay
8 × 105 cells were plated in 6-well plates and incubated overnight at 37°C. Cells were infected with Ad5-GFP at MOI 25 in DMEM with 2% FBS. After overnight incubation at 37°C, cells were trypsinized, washed with PBS, and GFP expression quantitated using flow cytometry. The BD FACSCalibur Flow Cytometer is the instrument used, 10,000 events were acquired for each experiment, gating for live cells, and FlowJo software was used to generate histograms and analyze data. For fiber blocking experiment, prior to addition of Ad5-GFP, different quantities of soluble fiber (1 ug/mL, 5 ug/mL, or 25 ug/mL) were added to cells, incubated at room temperature for 1 hr. Then Ad5-GFP was added to cells at MOI 25 and cells were incubated overnight at 37°C before flow cytometry analysis.
Surface expression levels
Cells were trypsinized, washed with PBS, and 1 × 106 cells were incubated with primary antibody for 30 minutes on ice. Cells were washed, incubated with secondary for 30 minutes on ice, and analyzed by flow cytometry. The BD FACSCalibur Flow Cytometer is the instrument used, 10,000 events were acquired for each experiment, gating for live cells, and FlowJo software was used to generate histograms and analyze data. Dilutions were as follows: RmcB (1-50), LM609 (1-100), P1F6 (1-100), JB1A (1-100), Alexa 488 (1-100). Secondary antibody (Alexa 488) alone was used as a control for each cell line.
Statistics
Microsoft Excel was used to do a 2 tailed, type 3 T Test to determine statistical significance.
Quantitative PCR
Total RNA was isolated from cells using RNeasy Mini Kit (Qiagen). PCR was performed by the Genome Analysis Core Facility, Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco. PCR was conducted in triplicate with 20 uL reaction volumes of 1× TaqMan buffer (1× Applied Biosystems PCR buffer, 20% glycerol, 2.5% gelatin, 60 nM Rox as a passive reference), 5.5 mM MgCl2, 0.5 mM each primer, 0.2 uM each deoxynucleotide triphosphate (dNTP), 200 nM probe, and 0.025 unit/uL AmpliTaq Gold (Applied Biosystems) with 5 ng cDNA. A large master mix of the above-mentioned components (minus the primers, probe, and cDNA) was made for each experiment and aliquoted into individual tubes, one for each cDNA sample. cDNA was then added to the aliquoted master mix. The master mix with cDNA was aliquoted into a 384-well plate. The primers and probes were mixed together and added to the master mix and cDNA in the 384-well plate. PCR was conducted on the ABI 7900HT (Applied Biosystems) using the following cycle parameters: 1 cycle of 95° for 10 minutes and 40 cycles of 95° for 15 seconds, 60° for 1 minute. Analysis was carried out using the SDS software (version 2.3) supplied with the ABI 7900HT to determine the Ct values of each reaction. Ct values were determined for three test and three reference reactions in each sample, averaged, and subtracted to obtain the ΔCt [ΔCt = Ct (test locus) – Ct (control locus)]. PCR efficiencies were measured for all custom assays and were greater than or equal to 90%. Therefore, relative fold difference was calculated for each primer/probe combination as 2-ΔCt × 100. PCR primer and TaqMan probe sequences were synthesized by Integrated DNA Technologies (Coralville, IA) [or purchased from Applied Biosystems]. The sequences were as follows
Human CAR
Amplicon:
GGCGCTCCTGCTGTGCTTCGTGCTCCTGTGCGGAGTAGTGGATTTCGCCAGAAGTTTGAGTATCACTACTCCTGAAGAGATGATTGAAAAAGCCAAAG
Forward: GGCGCTCCTGCTGTGC
Reverse: CTTTGGCTTTTTCAATCATCTCTTC
Probe: TGCGGAGTAGTGGATTTCGCCAGAAG
Human GapDH:
Amplicon: ATTCCACCCATGGCAAATTCCATGGCACCGTCAAGGCTGAGAACGGGAAGCTTGTCATCAATGGAAATCCCA
Forward: ATTCCACCCATGGCAAATTC
Reverse: TGGGATTTCCATTGATGACAAG
Probe: ATGGCACCGTCAAGGCTGAGAACG
Ad5 Binding Assay
Cells were plated in 96-well SigmaScreen poly-D-lysine coated plates (Sigma) and incubated overnight at 37°C. For peptide and antibody blocking experiments, cells were prechilled at 4°C for 30 minutes followed by addition of either peptide at indicated concentration or antibody (500 ug/mL) for 1 hr. Ad5-GFP (0.04 ug/uL, protein concentration determined by Bradford assay) diluted in DMEM with 50% FBS was added to cells and incubated for 6 hrs at 4°C. Cells were washed several times and fixed with ice cold solution of 95% EtOH/5% Acetic Acid. Cells were washed 1× TBST (0.05M Tris, 0.15 M NaCl, 0.5% Tween-20, pH 7.5) and incubated in Superblock (Pierce) for 1 hr at RT. Cells were washed 2× Superblock followed by incubation with a non-related control IgG antibody to block any non-specific interactions for 30 min, RT. Cells were washed 1× TBST and incubated with polyclonal rabbit anti-Ad5 antibody (Access Biomedical) for 30 min, RT, followed by washing 4xTBST. Cells were next incubated with Goat-anti-Rabbit-AP (Pierce) for 30 min, RT, followed by washing 4xTBST. Signal was then amplified and detected using an Elisa Amplification System per manufacturer's instructions (Invitrogen). For determining KD of Ad5 binding to cells, the above protocol was used except cells were incubated with Ad5 at varying concentrations for 18 hrs at 4°C prior to fixing
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CL carried out the experiments, participated in the design and coordination of the study, and drafted the manuscript. FM participated in the design and coordination of the study and contributed to the writing of the manuscript. All authors read and approved the final manuscript.