Alpha-defensin 5 differentially modulates adenovirus vaccine vectors from different serotypes in vivo

Adenoviral vectors have shown significant promise as vaccine delivery vectors due to their ability to elicit both innate and adaptive immune responses. α-defensins are effector molecules of the innate immune response and have been shown to modulate natural infection with adenoviruses, but the majority of α-defensin-adenovirus interactions studied to date have only been analyzed in vitro. In this study, we evaluated the role of α-defensin 5 (HD5) in modulating adenovirus vaccine immunogenicity using various serotype adenovirus vectors in mice. We screened a panel of human adenoviruses including Ad5 (species C), Ad26 (species D), Ad35 (species B), Ad48 (species D) and a chimeric Ad5HVR48 for HD5 sensitivity. HD5 inhibited transgene expression from Ad5 and Ad35 but augmented transgene expression from Ad26, Ad48, and Ad5HVR48. HD5 similarly suppressed antigen-specific IgG and CD8+ T cell responses elicited by Ad5 vectors in mice, but augmented IgG and CD8+ T cell responses and innate cytokine responses elicited by Ad26 vectors in mice. Moreover, HD5 suppressed the protective efficacy of Ad5 vectors but enhanced the protective efficacy of Ad26 vectors expressing SIINFEKL against a surrogate Listeria-OVA challenge in mice. These data demonstrate that HD5 differentially modulates adenovirus vaccine delivery vectors in a species-specific manner in vivo.


Introduction
Human α-defensins are small, cationic peptides consisting of 18-45 amino acids folded into a β-sheet structure that is stabilized by three disulfide bonds. To date, six α-defensins have been identified in humans. Human neutrophil peptides (HNP) 1, 2, 3 and 4 are primarily secreted from neutrophils and participate in systemic innate immunity [1], whereas human defensins (HD) 5 and 6 are typically secreted from intestinal Paneth cells and contribute to gastrointestinal tract innate immunity [2]. α-defensins have also shown the ability to neutralize bacteria, fungi, and viral targets [3,4].
α-defensins exhibit potent antiviral activity against adenoviruses [5]. In vitro experiments have shown that α-defensins inhibit expression from adenoviruses from species A, B1, B2, C, and E, whereas α-defensins may augment expression from adenoviruses from species D and F [5,6]. However, it has remained unclear how α-defensins modulate adenoviruses in vivo. Mechanistically, α-defensins are believed to inhibit adenoviruses by preventing uncoating, thereby restricting release of the endosomolytic protein (protein VI) during endosomal escape [6]. Consequently, adenoviral virions accumulate in endosomal compartments and fail to reach the nucleus. Although the precise mechanism of how α-defensins augment gene expression by species D and F adenoviruses is not fully understood, it has been proposed that increased numbers of Ad-defensin complexes may accumulate at cell surfaces [7] and may be internalized through defensin-specific receptors [8].
Several studies have suggested that enteric α-defensins enhance infectivity of mouse adenovirus types 1 and 2 in mice [8,9]. Furthermore, α-defensins have been shown to exert adjuvant-like properties when co-administered with antigens [10][11][12][13][14][15]. However, the potential impact of human α-defensins on human adenovirus vaccine vectors has not previously been studied in vivo and is important for defining the potential therapeutic use of α-defensins. We hypothesized that α-defensins, and in particular α-defensin 5 (HD5), may differentially modulate the immunogenicity of adenovirus vaccine vectors from different serotypes.
In this study, we assessed whether HD5 would modulate transgene expression from adenovirus serotypes 5, 26, 35, 48 (Ad5, Ad26, Ad35, Ad48) and a chimeric Ad5 vector containing the surface hexon hypervariable regions of Ad48 (Ad5HVR48), and whether HD5 would modulate immunogenicity by Ad5 and Ad26 vaccine vectors in mice. Co-administration of HD5 with Ad5 suppressed transgene expression, immunogenicity, and protective efficacy. In contrast, co-administration of HD5 with Ad26 enhanced transgene expression, increased Ad26elicited innate cytokines, antibody, and CD8 + T cells responses, and improved protective efficacy against a recombinant Listeria challenge.

HD5 alters adenovirus transduction sensitivity in A549 cells
Previous in vitro studies have shown that HD5 inhibits most adenovirus types, except adenoviruses from species D and F [5]. Based on these data, we postulated that HD5 might exhibit different effects with different adenovirus vectors in vivo. We studied two species D adenoviruses, Ad26 and Ad48 [16], expressing enhanced Green Fluorescent Protein (eGFP) for an initial in vitro study. For comparison, Ad5.eGFP (species C) and Ad35.eGFP (species B) were also included. We selected A549 cells for initial tests with HD5 as A549 cells are often used to study Ad infectivity [3,5,6,17]. Control infection was normalized to 100% of eGFP-positive cells 24 h post-infection in the absence of peptide. Ad5.eGFP was �99% inhibited at a concentration of �16 μM HD5 (p < 0.0001), and Ad35.eGFP showed 100% inhibition at a HD5 concentration of �33 μM (p < 0.0001) compared to control infection ( Fig 1A). In contrast, transduction of Ad26.eGFP and Ad48.eGFP was enhanced at concentrations of �8 μM HD5 (Ad26, p < 0.05; Ad48, p < 0.0001) (Fig 1A). Peak expression was at 72% (p < 0.0001) above control infection for Ad26.eGFP and 400% (p < 0.0001) above control infection for Ad48.eGFP with 33 μM HD5. In contrast, a negative control mutant HD5 peptide ( Fig 1B) (mHD5; 50 μM) did not modulate adenovirus infection compared to wild type (WT) infection (no HD5) for all viruses as expected (Fig 1A). Taken together, these data show that transgene expression from Ad26 and Ad48 vectors was enhanced by HD5 pre-treatment [6].  Prior studies have suggested that HD5 can bind to both Ad fiber and hexon proteins [5,18]. To evaluate potential interactions between HD5 and species D adenoviruses, we assessed the ability of HD5 to modulate the infectivity of a chimeric Ad5HVR48 virus, which is 99% Ad5 and contains the Ad5 fiber but has the surface hexon hypervariable regions (HVR) of Ad48 [19][20][21]. Chimeric Ad5/Ad26 vectors proved unstable and could not be constructed. Control infection in A549 cells was normalized to 100% of luciferase-positive cells 24 h post-infection in the absence of peptide. As shown in the previous experiment, expression from Ad5.luc was abrogated with 33 μM HD5 (p < 0.05), but expression from both Ad48.luc and Ad5HVR48. luc were markedly enhanced in the presence of HD5 (p < 0.01) (Fig 2). Since Ad5HVR48 shares the phenotype of Ad48, these data suggest that HD5 can interact with the surface hexon HVRs of Ad48, although we cannot exclude the possibility that HD5 may also interact with other Ad capsid proteins.

HD5 impacts adenovirus expression in vivo
We next immunized BALB/C mice with adenovirus vectors from different serotypes expressing luciferase to evaluate transgene expression in vivo. Ad5.luc (10 9 vp) and Ad26.luc (10 9 vp) were administered by the intramuscular (i.m.) route, and luciferase expression was monitored following intraperitoneal (i.p.) injection of luciferin substrate. Expression was assessed at 6 h, 1, 3, and 7 days post immunization. Animals that received Ad5.luc and 2.5 μM HD5 showed a 21-fold reduction in expression on day 3 (p < 0.001), whereas concentrations of �25 μM HD5 showed no detectable expression (Fig 3A and 3C). In contrast, a 3-fold to 8-fold enhanced luciferase expression was observed in animals receiving Ad26.luc and �25 μM HD5 (100 μM, p < 0.001 at day 7) (Fig 3B and 3D). As expected, the negative control mutant mHD5 showed no significant impact on any of the experimental groups assayed (Fig 3). These in vivo observations mirror the in vitro data (Fig 1A), which showed HD5 pre-treatment inhibits Ad5 and augments Ad26 transgene expression.

HD5 differentially impacts Ad-elicited antibody responses in vivo
We next evaluated whether HD5 administration would modulate the immunogenicity of Ad5 and Ad26 vaccine vectors in mice. Previous research has shown that defensins are capable of directly binding microbial antigens and forming complexes that may be more readily taken up by antigen-presenting cells, which effectively enhances the production of antigen-specific antibodies [24]. We therefore investigated whether HD5 impacted Env-specific IgG titers elicited by Ad5 and Ad26 vectors expressing SIV envelope protein (Env). Mice immunized with Ad5. SIVEnv (10 9 vp) showed a 1.7 log reduction in Env-specific IgG titers by day 14 (100 μM HD5; p < 0.0001) and a 1.6 log reduction by day 28 (100 μM HD5; p < 0.0001) with 100 μM HD5 compared with controls ( Fig 5A). In contrast, for Ad26.SIVEnv we observed a 9.3-fold enhancement of Env-specific IgG levels at day 14 (10 7 vp, p < 0.05; Fig 5B). These results indicate that HD5 suppressed humoral immune responses induced by Ad5.SIVEnv but augmented humoral immune responses induced by Ad26.SIVEnv.
We next immunized large groups of animals (n = 35/group) with a low dose (10 6 ) of Ad26. SIVEnv to evaluate if HD5 can boost humoral immune responses to a low vaccine dose. Fig 5C shows that 4 of 35 animals immunized with Ad26.SIVEnv and 100 μM HD5 had detectable Env-specific IgG titers on day 14, while no responses were detected in animals immunized with Ad26.SIVEnv alone. By day 28, 12 of 35 animals immunized with Ad26 and 100 μM HD5 had detectable Env-specific IgG titers, while only 1 of 34 animals immunized with Ad26. SIVEnv alone had detectable Env-specific IgG titers (Fig 5C; p = 0.0017, two-tailed Fisher's exact test). These data suggest that HD5 can boost humoral immune responses at low Ad26 vector doses.
We sought to characterize the protective efficacy of these antigen-specific CD8 + T cells by challenging vaccinated mice with Lm-OVA (10 5 CFU) 11 days post vaccination. Plaque assays from spleen on day 2 following Lm-OVA challenge showed significant differences in bacterial load counts compared to controls. Ad5 vaccinated mice with HD5 co-administration led to a significance among groups. The data are representative of experiments performed two times. Mean ± SEM and limit of detection (LOD) are shown. ���� p < 0.0001, ��� p < 0.001, �� p < 0.01, � p < 0.05, Wilcox statistical test. 5.7-log increase in bacterial loads (p < 0.01; Fig 6B), and the lack of protection in the Ad5 + HD5 group correlated to the reduction of SIINFEKL + CD8 + T cells. In contrast, Ad26 vaccinated mice that received HD5 co-administration exhibited a 6-log reduction in bacterial loads (p < 0.01; Fig 6B) compared to controls. We suspect that a threshold of SIINFEKL + CD8 + T cells may be necessary for protection in this model, consistent with prior studies [25][26][27]. These data demonstrated that HD5 differentially suppressed and augmented antigen-specific protective CD8 + T cells elicited by Ad5 and Ad26 vaccine vectors, respectively.

Discussion
The ability of α-defensins to modulate innate and adaptive immune profiles induced by Ad vaccine vectors in vivo remains poorly characterized. Here we demonstrate that HD5 differentially modulates the immunogenicity and protective efficacy of Ad5 and Ad26 vectors in mice in a species-specific manner. HD5 augmented innate, cellular and humoral immune responses elicited by Ad26 and improved protective efficacy against a recombinant Listeria challenge. In contrast, Ad5-elicited immune responses and protective efficacy were dampened by HD5. Taken together, these findings show that HD5 can have different effects on different subtype Ad vectors in vivo.  Prior studies have shown that Ad differential susceptibility to α-defensins can be divided in two categories based on whether α-defensins suppress or enhance Ad vector expression [5,6]. These studies showed in vitro that Ad sensitivity to α-defensins is largely dependent on species specificity, and that adenovirus serotypes 19, 23, 25, 37, 51 (species D) and 41 (species F) are uniquely enhanced by α-defensins, whereas other Ad species show inhibition. Our in vitro data confirm and extend these observations. Furthermore, we identified two additional species D serotypes, Ad26 and Ad48 that displayed enhanced expression in the presence of HD5. We also showed that surface hexon hypervariable regions contribute to the HD5-augmented phenotype exhibited by Ad5HVR48. However, relatively few prior studies have explored the ability of α-defensins to modulate the immunogenicity of adenovirus vectors in vivo [8,28]. Our data showed enhanced levels of innate cytokines and chemokines, enhanced antigen-specific serum IgG responses, and an enhanced protective capacity of antigen-specific CD8+ T cells in a Listeria challenge model in vivo with Ad26 in combination with HD5 as compared to Ad26 alone. In contrast, Ad5-elicited immunogenicity was suppressed by co-administration of HD5. Our data are consistent with previous studies that have shown that α-defensins can modulate innate immunity [4,29] and thereby enhance specific serum IgG antibody responses [30,31] as well as enhance CD8 + T cell responses in mice [31,32].
Ad vectors can be influenced in many ways by host antiviral responses. It is believed that HD5 exhibits antiviral activity that prevents non-species D or F Ad vectors from escaping the endosome during intracellular trafficking [6]. Although the mechanism of enhancement of species D and F Ad vectors is not fully understood, it has been proposed that non-neutralizing HD5 binding epitopes on the capsid surface mask the electronegative surface charge and facilitate entry into cells, similar to the use of polycation formulations, which enhance adenovirus uptake and subsequent transgene expression in vitro [33][34][35][36][37][38]. Recently, Vragniau et al.
[17] discovered a new mechanism used by a clade B, replication-competent Ad3 vector to overcome the antiviral effect of HD5 in vitro. Specifically, Ad3 produces subviral penton-dodecahedral particles that act as decoys for HD5, thus preventing the inactivation of progeny virus. These results may be due to the fact that a replication-competent virus was used in this particular case, whereas most other Ad-HD5 studies involve replication-defective viruses. Nonetheless, our study and others show that α-defensins have a significant impact modulating expression by Ad vectors and that more basic biology research is needed to fully elucidate these interactions.
Collectively, our data suggest that Ad vector-elicited immune responses are modulated by HD5 in a species-specific manner. The data presented here fills a knowledge gap by showing that HD5 has a profound impact on Ad vector-elicited innate, cellular, and adaptive immune responses in vivo. Such knowledge might facilitate the idea of using HD5 therapeutically to inhibit Ad types sensitive to HD5 like Ad5, but can also be explored for adjuvant discovery with Ad types similar to Ad26. To explore these effects to their full potential, further studies are warranted to define the specific mechanism of action of α-defensins on Ad-elicited adaptive immune responses, as well as the roles of the various cytokines and chemokines in modulating transgene-specific immune responses.

Viruses
Replication-incompetent, E1/E3-deleted Ad5, Ad26, Ad35, Ad48 and Ad5HVR48 vectors expressing enhanced green fluorescent protein (eGFP), SIV mac239 (SIV Env), luciferase (luc), or the ovalbumin immunodominant epitope (SIINFEKL) were as previously described [16]. Briefly, the plasmid/cosmid system for these vectors consisted of a pAdApt adaptor plasmid consisting of the left ITR; the packaging signal; an expression cassette involving a cytomegalovirus promoter, a multiple cloning site for insertion of a trans gene, and the simian virus 40 polyadenylation transcription termination signal; and a 2.0-2.5-kb fragment downstream of the E1 region that enables homologous recombination with pWE cosmids in E1-complementing cells such as PER.C6. A pWE cosmid containing the majority of each respective Ad genome spanning from the pIX sequence to the right inverted terminal repeat (ITR) with a deletion of the E3 region and a modified E4 open reading frame 6 (E4orf6) sequence was used for adenovirus construction. Viruses were grown in PER.C6 cells and maintained until virus cytopathic effect was observed. The vectors were plaque-purified, analyzed for transgene expression, amplified in 24 triple-layer T175 flasks, purified by double CsCl gradient ultracentrifugation, and dialyzed into phosphate-buffered saline (PBS) containing 5% sucrose. Purified rAd vectors were stored at -80˚C. Virus particle (vp) titers were determined by spectrophotometry [17]. Specific infectivity was assessed by PFU assays.

Mice
Six-to ten-week-old BALB/C and C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Animals were immunized intramuscularly (i.m.) in the quadriceps with virus, virus pretreated on ice for 1 h with HD5/mHD5, or PBS control in a volume of 100 μl divided equally between both legs unless noted otherwise. Experiments were conducted with n = 5 mice per group unless noted otherwise. All animal experiments were performed with approval from the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee.

Transduction assays in vitro
Transduction assays were conducted to determine the effect of HD5 or mHD5 on eGFP expression by Ad5, Ad26, Ad35, and Ad48 in vitro. A549 cells were seeded at a concentration of 5.0X10 4 in 24-well tissue culture plates the night before. The next day viruses were incubated with increasing concentrations of HD5 or mHD5 in serum-free DMEM and incubated on ice for 1 h. The mixture was then added to A549 cells that were pre-washed twice with 1X PBS to remove serum from the media. After a 2 h incubation at 37˚C, the media and unbound virus was aspirated and the cells were cultured in complete A549 media for 24 h at 37˚C. The next day the cells were washed twice with 1X DPBS, trypsinized, rinsed with AUTOMACS buffer (Myltenyi Biotec), and centrifuged at 1700 RPM (Sorvall Legend RT) for 5 min. Buffer was aspirated and cells were resuspended in 2% formaldehyde. Fixed cells were acquired on a LSR II flow cytometer (BD Biosciences) and analyzed for eGFP expression using FlowJo analysis software (Tree Star). Data are expressed as a percent of control infection in the absence of HD5. Virus concentrations were chosen that typically yielded 50% positive cells in control samples.
To analyze luciferase expression by Ad5, Ad48, and Ad5HVR48, A549 cells (1.0X10 4 cells/ well) cultivated in 96-well plates were infected with the indicated viruses at 2500 viral particles vp/cell to obtain a medium level of luminescence based on the detection limits of the instrument. After 2 hours, the infection medium was replaced with complete A549 medium. Culture supernatant was discarded and replaced with DPBS buffer 1 day post-infection. Steady Glo mix (100 μl) (Promega) was added into each well and incubated for 15 minutes at room temperature. Steady Glo/DPBS mix (100 ul) was transferred into new Black/White Isolaplate-96 plate and luminescence was detected by PerkinElmer 1420 multilabel reader. All experiments were performed in triplicate.

In vivo animal imaging
BALB/C mice were used due to their white coat color to prevent less non-specific background signal compared to other mouse strains. Animals were anesthetized with 2% isoflurane and oxygen for hair removal and imaged as previously described [39][40][41]. Briefly, hair was removed from the left hind leg on the ventral posterior area of the mouse. Mice were immunized in the left hind leg with Ad5.luc (10 9 vp) or Ad26.luc (10 9 vp) with and without increasing concentrations of HD5 or 100 μM mHD5. Animals were injected intraperitoneally (i.p.) with 150 μl of 30 mg/ml D-luciferin substrate (Caliper Life Sciences) according to the manufacturer's protocol. Luminescence was quantitated at 6 h, 1 day, 3 days, and 7 days post-immunization in an IVIS Lumina II charge-coupled device imaging system and Living Image software (Caliper Life Sciences). Images were integrated for 3 min, f/stop was 1.2, and binning was large.

Luminex
C57BL/6 mice were immunized i.m. with empty Ad5 (10 10 vp) and Ad26 (10 10 vp) vectors that were pretreated with 100 μM HD5 or PBS for 1 h on ice. Seven hours post immunization, animals were bled, and sera were isolated from whole blood. Sera were treated with 0.05% Tween-20 (Sigma) in 1X DPBS (Life Technologies) for 15 minutes at room temperature and then evaluated using the Millipore Milliplex Map Mouse Cytokine/Chemokine Magnetic Bead Panel (Millipore, Billerica, MA) according to the manufacturer's instructions. Samples were subsequently fixed with 4% formaldehyde in 1X DPBS for 1 h at room temperature, washed, and resuspended in Drive fluid (Luminex Corporation). Data were acquired on a MAGPIX running xPONENT version 4.2 software (Luminex Corporation) and analyzed using a 5-parameter logistic model with a 80%-120% standard acceptance range. Cytokine and chemokine levels were analyzed using GraphPad Prism v6.03 (GraphPad Software, CA, USA) and averaged for each experimental condition. Values below the limit of quantification were set to the limit of quantification of the analyte for analysis purposes. Statistical significance was assessed with a Wilcox statistical test using R language.
Two additional groups of animals (n = 35/group) were immunized with 10 6 vp of Ad26 with and without 100 μM HD5. We analyzed day 24 and 28 post-immunization time points by Endpoint ELISA (see above) and assessed the data for significance using a two-sided Fisher Exact Test by making comparisons between virus + HD5 and virus only groups.

Listeria challenge
C57BL/6 mice were immunized i.m. with Ad5.SIINFEKL (10 8 vp) or Ad26.SIINFEKL (10 8 vp) vectors expressing an immunodominant epitope from the ovalbumin protein, OVA 257-264 (SIINFEKL), which has been previously shown to protect against recombinant Listeria monocytogenes expressing the ovalbumin protein (LmOVA) [41]. Vectors were pretreated with 100 μM HD5 or PBS for 1 h on ice. Animals were bled 11 days post immunization and sera were stained for tetramer-specific CD8 + T cells using the MHC class I tetramer H-2K b loaded with the immunodominant OVA 257-264 (SIINFEKL) epitope from ovalbumin protein [44]. Animals were subsequently challenged with 1.0 X 10 5 colony forming units (CFU) of LmOVA by i.v injection. Each experimental condition was assayed for colony forming units by plating on BHI-agar plates as previously described [45].