Drug combination treatment as a strategy for inhibiting human adenovirus replication Authors: Mackenzie J Dodge, Katelyn M MacNeil, Tanner M Tessier, Andris M Evans, Kaitlyn R Palmateer, Jason B Weinberg, Joe S Mymryk Published: 2026 ABSTRACT AbstractHuman adenoviruses (HAdVs) are ubiquitous human pathogens that infect the respiratory, ocular, and gastrointestinal tissues. Despite the severity of these infections in immunocompromised patients, there are no clinically approved antiviral medications to treat HAdV infections. Over the past decade, many compounds have been found to interfere with different parts of the HAdV replication cycle. One critical barrier to developing a successful HAdV therapy arises if the drug concentration required for antiviral efficacy is clinically unachievable or too toxic for patient use. This problem can be diminished by using a combination of drugs that function synergistically, potentially allowing the use of lower drug concentrations that are clinically achievable and/or exhibit acceptable toxicity profiles. In this exploratory study, we examined the antiviral activity of pairwise combinations of six drugs that have been previously shown to disrupt HAdV replication: ivermectin, digitoxin, deguelin, niclosamide, rosiglitazone, and remdesivir. Combinations of these drugs showed a stronger reduction in HAdV progeny production, protein expression, and genome replication efficiency compared to their individual effects. These experiments serve to illustrate the feasibility and benefits of drug combinations that synergize against HAdV replication.Keywords: Human adenovirus, Combination therapy, Synergy, Antiviral, Ivermectin, Deguelin, Niclosamide Introduction Human adenoviruses (HAdVs) are ubiquitous human pathogens that infect the respiratory, ocular, and gastrointestinal tissues. There are currently over 100 types of HAdVs, with the most common infections being caused by the HAdV-C species (Hong et al., 2025; Lion, 2019; Lynch and Kajon, 2016). As a whole, HAdVs have a high global prevalence, as most individuals experience an infection with at least one type during early childhood (MacNeil et al., 2023; Lynch and Kajon, 2021). Notably, HAdV-E4, -B7, and -B14 have been associated with outbreaks and severe infections (Lion, 2019; Lynch and Kajon, 2016; Kajon et al., 2019; Cai et al., 2024; Saint-Pierre Contreras et al., 2023; Gray and Chorazy, 2009). HAdV infections in immunocompetent individuals are typically mild and self-resolving. However, severe HAdV associated disease can occur in immunocompromised individuals, particularly those receiving hematopoietic stem cell transplants (HSCT) and solid organ transplants (SCT), with severe disease in those at-risk patients being associated with higher mortality rates (Lynch and Kajon, 2021; Lion, 2014; Ison, 2006, 2022; Dailey et al., 2019; Fisher et al., 2024). Despite the severity of HAdV infections in immunocompromised patients, there are no clinically approved antiviral medications for HAdV infections (Ison, 2006). Instead, standard treatment focuses on supportive care and, when feasible, reduction of immunosuppressive treatment. In some cases, antiviral therapy is considered for patients with severe disease (Ison, 2006; Florescu et al., 2019). Drugs like cidofovir (CDV), ribavirin, and ganciclovir have been used to treat patients with severe HAdV infections, though they have varying degrees of efficacy and substantial toxicities (Ison, 2006; Florescu et al., 2019). CDV is most often used to treat patients with severe HAdV infections. However, CDV has variable efficacy and low bioavailability, and myelosuppression and nephrotoxicity frequently complicate its clinical utility (Dodge et al., 2021; Carretero-Ledesma et al., 2025; De, 2011; Toth et al., 2018; Wold et al., 2019; Cesaro, 2023). Brincidofovir, a CDV analog, showed promising results for treating HAdV infections in HSCT recipients, but unfortunately failed in phase III trials due to adverse gastrointestinal toxicities (Detweiler et al., 2018; Grimley et al., 2017). Traditional approaches to treating viral infections focus on the development of direct-acting antivirals (DAAs) that target virally-encoded enzymes, such as polymerases and proteases. While these agents often provide highly effective treatments, they require substantial investments from pharmaceutical agencies to develop (Sahoo et al., 2021). Unfortunately, due to the small patient population, there is little incentive for the pharmaceutical development of HAdV-specific DAAs (Harvey et al., 2009; He et al., 2024). An alternative and particularly intriguing approach is the development of host-directed therapies (HDTs) (Dodge et al., 2021; Mast et al., 2020). This approach takes advantage of the fact that viruses are obligate intracellular parasites and therefore depend on their hosts for replication. By targeting host proteins and processes that are essential for the virus, even low concentrations of these drugs can preferentially target infected cells, while allowing healthy cells to survive (Dodge et al., 2021; Mast et al., 2020). This strategy is also more likely to identify broad-spectrum therapies that are effective against multiple different viral infections (He et al., 2024). Over the past decade, many compounds have been found to interfere with different parts of the HAdV replication cycle, most likely via HDT-based mechanisms (Dodge et al., 2021). Progress at identifying a safe and effective antiviral to treat HAdV infections has often been focused on identifying opportunities for repurposing existing drugs that are clinically approved to treat other diseases (Dodge et al., 2021). Existing data on pharmacokinetics, safety, and dosing of the candidate HDT drugs can allow for streamlined clinical approval for a different condition. Despite the work done evaluating these drugs and their antiviral effects, there has been no successful development of an effective treatment specific to HAdV infections. One critical barrier to developing a successful HDT therapy arises if the drug concentration required for antiviral efficacy is clinically unachievable or may be too toxic for patient use. This problem can be diminished by using a combination of drugs that function synergistically, potentially allowing the use of lower drug concentrations that are clinically achievable and/or exhibit acceptable toxicity profiles (Good et al., 2025; Ianevski et al., 2022). In this exploratory study, we examined the antiviral activity of pairwise combinations of six drugs that have been previously shown to disrupt HAdV replication: ivermectin, digitoxin, deguelin, niclosamide, rosiglitazone, and remdesivir. (Carretero-Ledesma et al., 2025; Grosso et al., 2017; King et al., 2020; Jans and Wagstaff, 2020; Romanowski et al., 2021; Balsera-Manzanero et al., 2023; Xu et al., 2020a, 2020b, 2021; Marrugal-Lorenzo et al., 2019; Wang et al., 2020). Combinations of these drugs showed a stronger reduction in HAdV progeny production, protein expression, and genome replication efficiency compared to their individual counterparts. Our results suggest that combinations of antiviral drugs should be considered in the ongoing pursuit of a safe and effective therapy for HAdV infections. Moreover, using repurposed drugs avoids the need for de novo development of candidate antivirals, allowing us to focus on already approved agents as a therapy for patients with HAdV infections. These experiments serve to illustrate the feasibility and benefits of drug combinations that synergize against HAdV replication. Materials & methods Cell lines Human A549 lung adenocarcinoma cells (ATCC, CRM-CCL-185), were grown at 37°C with 5% CO2 in Dulbecco’s modified Eagle’s medium (DMEM; Multicell Technologies) supplemented with 1% penicillin/streptomycin (Multicell) and 10% fetal bovine serum (FBS; Multicell). Viruses and infection of cells Infections were carried out using either wildtype HAdV-C5 (ATCC; VR-1516) or the AdNTE reporter adenovirus (King et al., 2024). In brief, the HAdV-C5-based AdNTE reporter was constructed to have the E1A gene linked to nanoluciferase through a 2A self-cleaving linker, providing an enzymatic assay to measure changes in E1A expression. The AdNTE reporter provides accurate measurements of changes to overall viral progeny production that correlate with plaque assay measurements (King et al., 2024). A549 cells were infected at ~50% confluence (MOI of 5) for 1 h at 37°C, after which the infection inoculum was replaced with fresh DMEM containing indicated drug treatments. Drug treatments were added at 1 h post-infection (hpi) using fresh DMEM containing 10% FBS and 1% penicillin/streptomycin. Indicated drug concentrations (or equivalent volume of DMSO solvent control) and harvesting timepoints are shown in-text. All experiments were done in biological triplicate (n = 3) with technical duplicates in each plate. Cell viability assay A549 cells were seeded in 96-well plates at 2500 cells per well and left to adhere overnight. Medium containing the indicated drug (digitoxin, deguelin, remdesivir, rosiglitazone, or niclosamide) was added at concentrations ranging from 1 nM to 1 mM and left for 72 h before viability was assessed. For each condition, three replicates were used. Viability was measured using PrestoBlue reagent (Thermo Fisher) according to the manufacturer’s instructions. Normalized relative fluorescent units (RFUs) of the drug-treated replicates were calculated as a percentage of the mean RFU of control (DMSO only) treatment replicates. CC50 values, defined as the concentration of drug in which cell viability was 50%, were calculated by nonlinear regression (Prism 8; GraphPad Software, Inc.). Western blotting and antibodies A549 cells were infected with wildtype HAdV-C5 for 24 or 48 h before lysis. At 1hpi, the inoculum was removed and replaced with fresh DMEM containing indicated concentrations of ivermectin, niclosamide, digitoxin, and/or deguelin. Western blotting was performed as previously described (King et al., 2020). A representative image is shown for each blot. Primary antibodies used include mouse α-E1A (clone M73; sc-25, Santa Cruz), mouse α-DBP (CSB-PA365892ZA01HIL, Cusabio), rabbit α-adenovirus type 5 (ab6982, Abcam), and mouse α-actin (8H10D10; 3700, Cell Signalling). Horseradish peroxidase-conjugated secondary antibody was detected using Luminata Forte substrate (Millipore). qPCR/genome replication efficiency Samples for genome replication efficiency were prepared using the DNeasy blood and tissue kit as per the manufacturers instructions (Qiagen). Relative DNA levels were measured by qPCR using SensiFAST SYBR (Froggabio/bioline meridian) with oligonucleotide sequences that specifically recognize the left end of the HAdV-C5 genome and GAPDH as an endogenous cellular control. Results represent the mean of three independent replicates (n = 3) and were normalized to cellular GAPDH, calculated using the ΔΔCT method, and set as relative to DMSO-treated samples. Pan E1A forward: AGAGGCCACTCTTGAGTGC; Ad5 reverse: CGTCACGTCTAAATCATAC. GAPDH forward: ACTGCTTAGCACCCCTGGCCAA; reverse: ATGGCATGGACTGTGGTCATGAGTC. AdNTE reporter assay Detailed description of the reporter virus construction and validation are described in (King et al., 2024). Briefly, A549 cells were infected with the AdNTE reporter virus (MOI of 5) for 1 h before replacing the infection inoculum with fresh DMEM containing the indicated drug combinations/concentrations. At each of the indicated timepoints post infection, A549 cells were harvested and lysed via three cycles of freeze/thaw. Lysate from these cells was then aliquoted and diluted 1 in 5 in fresh DMEM. 20 μl of the diluted lysate was used to infect wells in a new plate of A549 cells for 1 h before replacing the inoculum with fresh DMEM not supplemented with drug. At 6hpi, relative light units (RLU) were measured using the Nano-Glo luciferase assay system (Promega) as per the manufacturer’s instructions. Assays were conducted in biological triplicate (n = 3) with technical duplicates in each plate. EC50 values, defined as the concentration of drug in which HAdV progeny production was 50% of the DMSO control levels, were calculated by nonlinear regression (Prism 8; GraphPad Software, Inc.). SynergyFinder Data from AdNTE reporter assays was input into the SynergyFinder program to determine potential synergy between indicated drug combinations (Ianevski et al., 2017, 2020; Zheng et al., 2022). Briefly, SynergyFinder calculated four synergy scores (ZIP, HAS, Bliss, and Loewe) for each pairwise drug combination. Positive scores indicate potential synergy, with a strong likelihood for those scores >10, while negative scores reflect potential antagonism that is more likely if the score is < −10. For increased stringency, we focused on combinations that showed positive values, indicating either additive or synergistic effects, for all four scores. Statistical analysis All experiments were carried out with three biological replicates. Data shown represent the grand means of three biological replicates and standard errors of the mean (SEM). Statistical analysis was performed using Prism 8; GraphPad Software, Inc. Statistical significance was evaluated using one-way ANOVA, followed by Dunnett’s multiple comparison test for AdNTE assay and Fisher’s LSD test for genome replication assay. Results Candidates for combination treatment strategy to inhibit HAdV replication Previous work done by our group and others has identified drugs that interfere with HAdV replication (Dodge et al., 2021; Carretero-Ledesma et al., 2025; Grosso et al., 2017; King et al., 2020; Jans and Wagstaff, 2020; Romanowski et al., 2021; Balsera-Manzanero et al., 2023; Xu et al., 2020a, 2020b, 2021; Marrugal-Lorenzo et al., 2019). We selected ivermectin, digitoxin, deguelin, niclosamide, remdesivir, and rosiglitazone as candidates to evaluate in combination for potential synergistic reduction in HAdV replication. We performed CC50 calculations for each drug to determine the maximum cytotoxic dose we would test in our system (Table 1, Sup Fig. 1). Once the toxicity limit for each compound was established in A549 cells, we examined the potential of each drug individually to reduce HAdV replication over decreasing concentrations (100%, 50%, 10%, or 1% of the CC50 dose; Fig. 1). To improve throughput for these assays as compared to plaque assays, we used a previously described luciferase HAdV-C5 reporter virus, AdNTE to quantitatively measure changes in viral progeny production for each treatment condition (King et al., 2024). Using a two-step infection with the AdNTE reporter virus, infected cells were treated with different concentrations of each individual drug, and the yield of infectious viral progeny produced at 16, 24, 48, and 72hpi was measured through a second round of infection to characterize their effects on viral replication (Fig. 1). Table 1CC50 values, proposed mechanism, and approval status of combination candidate drugs.DrugCC50aProposed anti-HAdV target mechanismClinical approval statusbIvermectin10 μMNuclear transport machinery (King et al., 2020)ApprovedDigitoxin20 nMEarly-stage replication (Grosso et al., 2017)Clinical trials (phase II NCT00782288), related compounds are approvedDeguelin20 μMLate-stage replication (Balsera-Manzanero et al., 2023)Not approvedNiclosamide1 μMEndosomal escape (Marrugal-Lorenzo et al., 2019)Clinical trials in progress (phase IV NCT05087381 and phase I NCT05188170)Remdesivir25 μMUnknown (Romanowski et al., 2021)Approved, clinical trials in progress (phase II NCT07013474, phase III NCT06817889 and phase IV NCT06792214)Rosiglitazone100 μMIFN activation (Wang et al., 2020)Tentative approval, clinical trials in progress (phase II NCT04114136 and phase III NCT04148183)aSee Supplementary Fig. 1 for data.bClinical approval status and trial information acquired from https://www.canada.ca/en/health-canada.html, https://www.fda.gov/drugs, and https://clinicaltrials.gov/. CC50 values, proposed mechanism, and approval status of combination candidate drugs. See Supplementary Fig. 1 for data. Clinical approval status and trial information acquired from https://www.canada.ca/en/health-canada.html, https://www.fda.gov/drugs, and https://clinicaltrials.gov/. Effects of individual drug exposure on HAdV replication. A549 cells were infected with the AdNTE reporter virus at an MOI of 5 and treated with DMSO or the indicated concentrations of ivermectin (A), digitoxin (B), deguelin (C), niclosamide (D), remdesivir (E), or rosiglitazone (F). Lysate from these infections was harvested at 16, 24, 48, and 72hpi and used for a second round of infection of A549 cells, measuring RLU production from the reporter virus after 6 h. In each panel, the different lines represent 100% (purple, squares), 50% (green, hexagons), 10% (magenta, triangles), and 1% (light blue, diamonds) of the respective drug CC50 as well as the DMSO control (black, circles; Sup Fig. 1). Error bars represent standard error of the mean (n = 3). * <0.05, ** <0.01, *** <0.005, NS not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) We observed significant decreases in HAdV progeny production at 48 and 72hpi when cells were exposed to digitoxin, ivermectin, niclosamide, or deguelin at various concentrations. In contrast, rosiglitazone and remdesivir showed little effect in reducing HAdV replication, even at their respective CC50 concentrations, limiting their utility for further investigation. For this reason, we focused our proof-of-concept studies on drug synergy using pairwise combinations of digitoxin, ivermectin, niclosamide and deguelin. Effects of ivermectin and digitoxin combinations on HAdV replication We first examined the effects of combining ivermectin and digitoxin on inhibiting HAdV infection at 72hpi. We noted that digitoxin alone caused a dose-dependent decrease in viral progeny production, producing between 0.2 and 65% viral progeny compared to that of the DMSO control (Fig. 2A). The combination groups also showed a dose-dependent decrease with increasing digitoxin concentrations. Conversely, ivermectin monotreatment caused viral progeny production to be between 63 and 78% of the DMSO control levels (Fig. 2A). To determine potential synergy between combination groups, we then employed the widely-used SynergyFinder (Ianevski et al., 2017, 2020; Zheng et al., 2022) to calculate four different synergy scores (ZIP, HSA, Bliss, Loewe) for each combination. These scores identify synergistic drug responses as a decrease in viral progeny greater than the decrease caused by each drug individually. We aligned the drug concentrations, reduction in viral progeny production, and synergy scores across all combinations (Fig. 2B). For stringency and to account for the assumptions and calculations used to generate the four different synergy scores, we chose to investigate only those combinations that showed positive values (indicative of synergy) for all four scores. Specifically, ivermectin at 1 μM and 100 nM with digitoxin at 2 nM showed positive ZIP, HSA, Bliss, and Loewe synergy scores (red bars; Fig. 2B). Effects of combining ivermectin and digitoxin on HAdV replication. A549 cells were infected with AdNTE reporter virus at an MOI of 5 and treated with the indicated concentrations of ivermectin, digitoxin, or DMSO solvent control before being harvested for luciferase assay. A) Combinations of ivermectin and digitoxin at 1, 10, 50, or 100% of their respective CC50 values were tested for their ability to reduce HAdV replication at 72hpi. Antiviral effects are represented as % viral replication, with stronger reduction being more intense red colour (scale bar shown). Comparing each combination to the DMSO control, there is a decrease in viral replication across the different drug treatments. Underlined values represent those that show synergy by all four synergy scores indicated in panel B. B) Four synergy scores (ZIP, HSA, Bliss, Loewe) were produced using SynergyFinder (Ianevski et al., 2017, 2020; Zheng et al., 2022). These scores reflect the likelihood of antagonism (<0), additivity (=0), or synergy (>0) for the respective combinations. Drug concentrations and synergy scores are aligned to viral progeny production. Two combinations show positive values for all four synergy scores (red bars). C) Time course of infected cells treated with selected synergistic ivermectin and digitoxin combinations (red), respective single drug concentrations (grey) and DMSO control (black). Lysate from these infections was harvested at 16, 24, 48, and 72hpi and used for a second round of infection of A549 cells, measuring RLU production from the reporter virus after 6 h. Error bars represent standard error of the mean (n = 3). Statistical significance was determined with a one-way ANOVA (Dunnett’s multiple comparison test) comparing groups to the DMSO control. * <0.05, ** <0.01, *** <0.005, NS not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) We next looked at the effects of these specific combinations across an infection time course with the AdNTE reporter virus (Fig. 2C). There was no reduction in viral progeny when infected cells were treated with ivermectin or digitoxin individually (grey), compared to DMSO (black). However, the combination of ivermectin at 100 nM or 1 μM with digitoxin at 2 nM led to decreased viral progeny at 48hpi and 72hpi for the 2 nM digitoxin-100 nM ivermectin combination (red; Fig. 2C). Finally, we normalized 72hpi viral progeny production to 16hpi and expressed it as a percentage of the DMSO control levels for these combinations, to control for any changes in viral progeny early in infection (Sup Fig. 2A). There was no statistically significant difference in viral progeny production across all drug combinations (red) or monotreatments (grey), compared to the DMSO (black). However, there was a trend towards a decrease in viral replication when normalizing 72hpi to 16hpi. Overall, while there is apparent synergy of ivermectin and digitoxin at the tested concentrations at 72hpi, when normalized to 16hpi this effect is diminished. Effects of ivermectin and niclosamide combinations on HAdV replication The maximum decrease we observed from combinations of ivermectin and niclosamide reduced viral progeny production to 33% of the DMSO control, while most showed 60-80% progeny production at 72hpi (Fig. 3A). Comparing all synergy scores for these conditions, all ivermectin combinations showed synergy with 100 nM niclosamide (purple; Fig. 3B). Examining the effects of these combinations (purple) across an infection time course at 16, 24, 48, and 72hpi, we observed no significant decrease in viral progeny production compared to the DMSO control (black), though progeny production in the presence of 10 μM ivermectin and 10 nM niclosamide (purple) was lower compared to monotreatment groups (grey; Fig. 3C). Normalizing the 72hpi data to 16hpi, we saw no significant reduction in HAdV viral progeny production, though there was an apparent trending decrease for all ivermectin monotreatments (grey) and the respective combinations with 10 nM niclosamide (purple) compared to DMSO (black; Sup Fig. 2C). This suggests that ivermectin alone may reduce HAdV replication at the concentrations tested, without any synergistic benefit when in combination with niclosamide at the concentrations we tested. Effects of ivermectin and niclosamide on HAdV replication. A549 cells were infected with the AdNTE reporter virus at an MOI of 5 and then treated with the indicated drug combinations at 1hpi. A) Combinations of ivermectin and niclosamide at 1, 10, 50, or 100% of their respective CC50 values were tested for their ability to reduce HAdV replication at 72hpi. Antiviral effects are represented as % viral replication, with stronger reduction being more intense purple colour (scale bar shown). Underlined values represent those that show synergy by all four synergy scores indicated in panel B. B) Four synergy scores (ZIP, HSA, Bliss, Loewe) were produced using SynergyFinder (Ianevski et al., 2017, 2020; Zheng et al., 2022). These scores reflect the likelihood of antagonism (<0), additivity (=0), or synergy (>0) for the respective combinations. Drug concentrations and synergy scores are aligned to viral progeny production. Combinations with a positive value for all four synergy scores (purple bars) were chosen to examine further. C) Time course of infected cells treated with selected synergistic ivermectin and niclosamide combinations (purple), respective single drug concentrations (grey) and DMSO control (black). Lysate from these infections was harvested at 16, 24, 48, and 72hpi and used for a second round of infection of A549 cells, measuring RLU production from the reporter virus after 6 h. Error bars represent standard error of the mean (n = 3). Statistical significance was determined with a one-way ANOVA (Dunnett’s multiple comparison test) comparing groups to the DMSO control. * <0.05, ** <0.01, *** <0.005, NS not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Effects of niclosamide and digitoxin combinations on HAdV replication Treating HAdV infected cells with 20 nM digitoxin produced less than 3% viral progeny compared to the DMSO group, regardless of combination with any niclosamide concentration at 72hpi (Fig. 4A). This effect of digitoxin was dose-dependent, while the majority of niclosamide monotreatments led to between 65% and 80% viral production (Fig. 4A). Combining these drugs showed further reduction and positive synergy scores. Specifically, 2 nM digitoxin treatments paired with 100 nM, 500 nM, or 1 μM niclosamide showed all 4 synergy scores to be positive (blue), indicating these combinations may be synergistic (Fig. 4B). Indeed, combinations of 10 nM digitoxin with 100 nM or 1 μM niclosamide (blue) caused a decrease in viral progeny production at 72hpi to a greater extent than effects seen in the monotreatment groups (grey), when compared to the DMSO control (black) (Fig. 4C). Finally, normalizing the change in viral progeny production at 72hpi to 16hpi, the combinations of 2 nM digitoxin with either 500 nM or 1 μM niclosamide (blue) caused a significant reduction in HAdV progeny production, while the monotreatments (grey) were not significantly reduced compared to control-treated cells (black; Sup Fig. 2D). These data suggest that niclosamide and digitoxin are candidates for an antiviral synergistic combination, as their combined effects caused greater inhibition of viral replication than the individual drug conditions. Effects of niclosamide and digitoxin on HAdV replication. A549 cells were infected with the AdNTE reporter at an MOI of 5 and then treated with the indicated combinations of niclosamide and digitoxin or a DMSO control. A) Combinations of niclosamide and digitoxin at 1, 10, 50, or 100% of their respective CC50 values were tested for their ability to reduce HAdV replication at 72hpi. Antiviral effects are represented as % viral replication, with stronger reduction being more intense blue colour (scale bar shown). Underlined values represent those that show synergy by all four synergy scores indicated in panel B. B) Four synergy scores (ZIP, HSA, Bliss, Loewe) were produced using SynergyFinder (Ianevski et al., 2017, 2020; Zheng et al., 2022). These scores reflect the likelihood of antagonism (<0), additivity (=0), or synergy (>0) for the respective combinations. Drug concentrations and synergy scores are aligned to viral progeny production. Only 3 combinations produced positive values for all four scores (blue bars). C) Time course of infected cells treated with selected synergistic niclosamide and digitoxin combinations (blue), respective single drug concentrations (grey) and DMSO control (black). Lysate from these infections was harvested at 16, 24, 48, and 72hpi and used for a second round of infection of A549 cells, measuring RLU production from the reporter virus after 6 h. Error bars represent standard error of the mean (n = 3). Statistical significance was determined with a one-way ANOVA (Dunnett’s multiple comparison test) comparing groups to the DMSO control. * <0.05, ** <0.01, *** <0.005, NS not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Synergy evaluation of deguelin combinations at EC50 doses We performed a similar analysis of deguelin combined with either ivermectin, digitoxin, or niclosamide using fractions of their CC50 value (Sup Fig. 3, Fig. 4, Fig. 5). However, the potent antiviral activity of deguelin at even 1% of its CC50 value made it impossible to determine if there was evidence of synergy. As a result, we performed EC50 calculations of these drugs to determine what concentrations of each drug resulted in a 50% reduction in infectious viral progeny production (Sup Fig. 6). Indeed, deguelin showed an EC50 dose of 80 nM, which is far lower than its CC50 of 20 μM (Sup Fig. 6). In comparison, the EC50 value for digitoxin was 75 nM, which was almost 4x its CC50 value (20 nM). Ivermectin and niclosamide showed EC50 values resembling their CC50 values. Synergy evaluation for deguelin combinations at EC50-based doses. A549 cells were infected with the AdNTE reporter at an MOI of 5 and then treated with the indicated combinations of ivermectin, digitoxin, or niclosamide and deguelin or a DMSO control. A,C,E) Combinations of ivermectin (A), digitoxin (C), or niclosamide (E) at 1, 10, 50, or 100% of its CC50 value were tested in combination with deguelin at the same percentages of its EC50 value to examine their ability to reduce HAdV replication at 72hpi. Antiviral effects are represented as % viral replication, with stronger reduction being more intense colour (green: ivermectin-deguelin; yellow: digitoxin-deguelin; orange: niclosamide-deguelin; scale bar shown). Underlined values represent those that show synergy by all four synergy scores indicated in panel B, D, and F. B,D,F) Four synergy scores (ZIP, HSA, Bliss, Loewe) were produced using SynergyFinder (Ianevski et al., 2017, 2020; Zheng et al., 2022). These scores reflect the likelihood of antagonism (<0), additivity (=0), or synergy (>0) for the respective combinations. Drug concentrations and synergy scores are aligned to viral progeny production. Deguelin paired with ivermectin, digitoxin, and niclosamide showed 2, 15, and 11 concentration pairs, respectively, with positive values for all four scores. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) We further evaluated the synergistic potential of combinations of ivermectin, digitoxin, and niclosamide with deguelin based on 100%, 50%, 10%, and 1% of its EC50 dose (Fig. 5). Both ivermectin and deguelin produced a dose-dependent reduction in viral progeny production, yielding 35 and 30% of control levels respectively at their highest doses (Fig. 5A). Combinations produced even further reduction, causing replication to be less than 20% of the DMSO control levels for some concentration pairs. However, SynergyFinder analysis revealed that only two of these pairs had positive values for all four synergy scores (Fig. 5B). We also tested digitoxin at its EC50 dose, despite being higher than its CC50 value. Again, we observed a dose-dependent effect down to a minimum progeny production of 35% of the control from both monotreatments (Fig. 5C). Combining these drugs we observed further reduction in HAdV replication, with the maximum dose of both drugs yielding 11% of control levels. SynergyFinder revealed 15 groups with positive synergy scores, a dramatic difference from our observations at higher doses of deguelin (Fig. 5D, Sup Fig. 4B). Finally, testing these lower deguelin concentrations in combination with niclosamide, we observed the same dose-dependent decrease in viral progeny production from deguelin, but saw no effect of niclosamide monotreatment (Fig. 5E). Despite this lack of monotreatment efficacy, we did observe superior reduction in the combination groups, with less than 1% replication observed from combining both drugs at the highest doses tested. Of the concentration pairs tested, 11 showed all synergy scores to be positive, indicating synergy between the niclosamide and deguelin at those concentrations (Fig. 5F). Synergistic combinations are effective at reducing HAdV genome replication and protein expression To expand on our findings of the synergistic combinations of ivermectin, digitoxin, deguelin, and niclosamide at reducing HAdV progeny production, we examined their combined impact on genome replication at 24hpi, when substantial DNA replication was occurring (Fig. 6A–E). We focused on combinations that produced positive synergy values by all four scores (ZIP, Bliss, HSA, Loewe) and were below the respective drug CC50. Digitoxin 2 nM either alone (grey) or in combination with 100 nM or 1μM ivermectin (red) caused a significant reduction in genome replication below 20% of the DMSO control level (black; Fig. 6A). Conversely, we observed no statistically significant differences in genome replication when cells were treated with ivermectin or deguelin, either alone (grey) or in combination (green), compared to the DMSO control (black; Fig. 6B). Similarly, ivermectin and niclosamide showed no significant reduction in HAdV genome replication in any of the single (grey) or combination (purple) treatments versus DMSO control (black; Fig. 6C). Combinations of niclosamide and digitoxin (blue) caused a significant reduction in genome replication compared to the control (black; Fig. 6D). Similar decreases were also observed for monotreatments of 2 nM digitoxin and 100 nM niclosamide, but not for 500 nM niclosamide (Fig. 6D). Finally, niclosamide and deguelin treatment revealed a significant reduction in viral genome replication, for all combinations (orange) and monotreatments (grey) of 10 μM deguelin and 100 nM niclosamide (Fig. 6E). Deguelin 2 μM or niclosamide 500 nM alone did not significantly reduce HAdV genome replication (Fig. 6E). Effects of drug combinations on HAdV genome replication and protein expression. A549 cells were infected with wildtype HAdV-C5 at an MOI of 5 and treated with the indicated drug concentrations. A-E) Genome replication was measured by qPCR of the left end of the HAdV-C5 genome normalized to cellular GAPDH at 24hpi and compared to their 6hpi baseline levels. Synergistic combinations of pairwise drug combinations are shown: A) ivermectin and digitoxin (red), B) ivermectin and deguelin (green), C) ivermectin and niclosamide (purple), D) niclosamide and digitoxin (blue), and E) niclosamide and deguelin (orange), compared to their respective monotreatments (grey) and DMSO control (black). Genome replication efficiency was expressed as a percentage of the DMSO control group. Statistical significance was determined with a one-way ANOVA (Dunnett’s multiple comparison test) comparing groups to the DMSO control. * <0.05, ** <0.01, *** <0.005. F-J) Western blot was used on cells infected for 48hrs to detect HAdV E1A, DBP, and viral structural proteins (hexon, penton, pV, and pVI). β-actin was used as a loading control. The same monotreatment samples were run in multiple experiments for both qPCR and western blotting. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Overall, most of these observations agree with the effects on HAdV progeny production we saw with the AdNTE reporter (Figs. 2–5). However, we did note a difference for digitoxin treatment combinations. For digitoxin combinations with both ivermectin and niclosamide, we noted that the digitoxin 2 nM monotreatment group caused a significant reduction in genome replication. We did not observe a similar effect of digitoxin alone in our AdNTE reporter assay experiments (Figs. 2 and 4). We also assessed effects of those antiviral combinations on HAdV replication by evaluating viral protein production at 48hpi, when both early and late proteins are produced (Fig. 6F–J). Infected A549 cells treated with ivermectin and digitoxin, alone or in combination, revealed similar levels of viral early region 1 A (E1A) proteins (13S and 12S isoforms) and single-stranded DNA binding protein (DBP), as well as HAdV structural proteins (hexon, penton, fiber, pV, pVI; Fig. 6F). Niclosamide at 100 or 500 nM when combined with digitoxin at 2 nM yielded no change to HAdV proteins for most conditions (Fig. 6I). Conversely, for combinations of ivermectin with deguelin, we observed a substantial decrease in E1A, DBP, and HAdV structural proteins when cells were treated with 2 μM deguelin alone or in combination with ivermectin at 48hpi (Fig. 6G). It also appears that increasing concentrations of ivermectin alone caused a dose-dependent decrease in E1A expression (Fig. 6G). While there was no decrease in hexon and penton when infected cells were treated with ivermectin and niclosamide combinations, pV, E1A, and DBP expression were reduced (Fig. 6H). In particular, 100 nM ivermectin alone as well as 5 μM ivermectin alone or combined with 10 nM niclosamide showed substantial decrease in E1A, DBP, pV, and pVI levels. Finally, we assessed combinations of niclosamide and deguelin. We observed substantial decreases in E1A, DBP, and viral structural proteins when cells were treated with either 2 μM or 10 μM deguelin alone, while treatment with niclosamide alone had minimal effects. Combinations of deguelin with niclosamide, particularly with niclosamide 0.5 μM, lead to a further reduction in DBP and HAdV structural proteins beyond differences seen with deguelin alone, with most being undetectable compared to the DMSO control-treated cells (Fig. 6J). Thus, except for those synergistic effects of deguelin and niclosamide, reductions in viral protein expression largely appeared to be caused by a single drug. Discussion HAdV outbreaks have been reported with increasing frequency in both immunocompetent and immunocompromised populations (Cai et al., 2024; Killerby et al., 2019; Dudding et al., 1972; Muller et al., 2018; Hysmith et al., 2018; Kajon et al., 2018; Miro et al., 2021; Lamson et al., 2018; Kujawski et al., 2021; Lee et al., 2024; Giardina et al., 2024; Marafungana et al., 2024; Fukuda et al., 2024; Servellita et al., 2023; Montgomery et al., 2024; Hannestad et al., 2023; Wang et al., 2023; Abdullah et al., 2024). Severe HAdV infections are a significant cause of mortality in immunocompromised populations and there is no clinically approved therapy to offer these patients (Lion, 2014, 2019). Despite having a lower incidence of severe infections than other viruses, HAdVs can be fatal when these pathologies develop (Tollefson et al., 2025). The disease severity and propensity of HAdVs to cause outbreaks drives the need to develop safe and effective treatment options. The study serves as a proof of concept, illustrating that combinations of existing drugs can be used to create antiviral synergy against HAdVs. Further, we focused on testing concentrations below the CC50 of each drug to illustrate the utility of synergistic inhibition of HAdV replication. It is also important to consider the EC50 doses of candidate drugs in these antiviral assays. Two of the drugs tested, ivermectin and niclosamide, showed EC50 values close to their CC50 values, verifying our choice in dose testing (Sup Fig. 1). However, the EC50 dose of deguelin was determined to be 80 nM, 250 times lower than its CC50 dose, highlighting its antiviral potency and necessitating synergy evaluation at those lower doses (Fig. 5). Conversely, digitoxin showed a greater EC50 dose than its respective CC50 value. While we did observe synergetic effects when using this higher dose of digitoxin, its impacts on cell viability restrict its use in future therapeutic exploration, at this concentration (Sup Fig. 1). Indeed, our work suggests that deguelin has a strong antiviral effect at interfering with HAdV replication, and that this may be improved when combined with niclosamide. This strategy follows in the footsteps of many other fields, including efforts to develop combination anti-viral, anti-parasitic, and anti-cancer drug therapies (Gibas et al., 2022; Naggie and Muir, 2017; Wagoner et al., 2022; Abdelnabi et al., 2022; Herring et al., 2021; Dyall et al., 2018). An important potential benefit of combination therapy and drug synergy is that treatment doses can be reduced in cases of synergistic efficacy. Reduced dosage may minimize toxicities and reduce the likelihood of viral escape mutants, while enabling previously unachievable treatments in clinical settings (Dodge et al., 2021; Abdelnabi and Delang, 2020; Verwimp et al., 2025). We chose these six drugs as candidates because they potentially target different cellular mechanisms and pathways that HAdV usurps. Niclosamide, a clinically approved anti-parasitic drug, has shown broad-spectrum activity against other diseases, including many viral infections (Xu et al., 2020b; Chen et al., 2018; Wang et al., 2022). With respect to HAdV, niclosamide has been shown to interfere with endosomal escape of infecting virions and cause a decrease in viral protein expression as well as nuclear localization of HAdV genomes (Marrugal-Lorenzo et al., 2019). Importantly, a related, newly developed niclosamide analog, JMX0312, has recently been shown to be effective and safe in the immunocompromised Syrian hamster HAdV infection model, supporting the potential for clinical translation of this family of drugs (Carretero-Ledesma et al., 2025). Digitoxin is suggested to exert its effects on HAdV at early, pre-DNA replication stages, but no pathway has been specifically linked with this effect (Grosso et al., 2017). However, cardiac steroids like digitoxin have been shown to exert antiviral activity against HIV through the cellular MEK/ERK signalling pathway, so this pathway may be one to pursue for targeting HAdV (Wong et al., 2018). Alternatively, deguelin targets a later phase of HAdV replication through an unknown mechanism (Balsera-Manzanero et al., 2023). Previous reports have linked deguelin to neurotoxicity that results in Parkinsons disease-like symptoms in rats (Wang et al., 2013; Caboni et al., 2004). Despite these concerns, some derivative compounds have shown promising reduction in toxicity (Lee et al., 2016; Kim et al., 2008). Further work on deguelin and its derivatives will be necessary to fully understand their toxic adverse effects, and thus their potential as clinically viable agents. Ivermectin was shown to disrupt the nuclear transport machinery that HAdV relies heavily on during infection (King et al., 2020). Rosiglitazone and remdesivir have been suggested to function against HAdV through the interferon pathway and an unknown pathway (Romanowski et al., 2021; Wang et al., 2020). However, we did not observe effects of those drugs on viral replication in our initial experiments and so chose not to include them in subsequent experiments with drug combinations. We used a replication-competent reporter virus, AdNTE, as a high throughput means to investigate the effects of antiviral combinations on overall viral replication through progeny production (King et al., 2024). Due to the enzymatic nature of the AdNTE reporter virus, we were more easily able to examine multiple key timepoints throughout the HAdV replication cycle and across a range of drug concentrations (Figs. 2–4, Sup Fig. 3, Fig. 5). Initially, we tested ivermectin, digitoxin, deguelin, and niclosamide at four concentrations and potential synergistic interactions were identified using the tool SynergyFinder. Deguelin resulted in robust inhibition of viral replication, prompting us to test additional, lower concentrations, focusing on the EC50 dose. SynergyFinder calculates four different synergy scores for each combination, with each score taking different considerations into account when determining potential treatment synergy (Ianevski et al., 2017, 2020; Zheng et al., 2022). Many other groups have used this tool to similarly investigate combination treatments (Herring et al., 2021; Giammarino et al., 2025; Pérez-Vargas et al., 2023; Tang et al., 2024; Fiaschi et al., 2024). Overall, we identified 17 combinations with positive scores across the four synergy score matrices, 15 of which used concentrations below the CC50 dose (Figs. 2–4, Sup Fig. 3, Fig. 5). Alternatively, using deguelin at doses at and below its EC50 value, we identified 28 combinations with all-positive synergy scores (Fig. 5). We took all four synergy matrix scores into consideration as a more stringent approach for determining potential synergy. It is important to note that many of the combinations we initially tested revealed some potentially antagonistic scores through SynergyFinder. At high concentrations, some synergistic drug pairs can exhibit levels of antagonism; however, as we are focused on lower concentrations to avoid potential toxicities, the potential for antagonism is less concerning (Foucquier and Guedj, 2015; Meyer et al., 2019; Cokol et al., 2011). While these combinations all had different magnitudes of antiviral effects on HAdV replication, niclosamide and deguelin exerted a uniquely strong impact on HAdV replication. Overall viral titre yield was dramatically reduced throughout viral replication, as seen with our AdNTE reporter assay (Sup Fig. 5). Further exploration into viral genome replication for these combinations identified a 75-85% reduction in genome replication efficiency (Fig. 6E). Similarly, we observed substantial reductions in viral protein expression when cells are treated with combinations of niclosamide and deguelin (Fig. 6J). Interestingly, many of our results suggest that deguelin plays a predominant role in inhibiting HAdV replication. Though we selected 2 and 10 μM deguelin based off our SynergyFinder results from our CC50-based experiments, we still observed potent antiviral activity from deguelin at lower EC50-based doses, alone and in combination with the other three antivirals. Interestingly, combinations of digitoxin with deguelin at higher doses revealed no synergistic pairs by our criteria of all scores being positive, where 15 pairs were synergistic at lower doses of deguelin (Fig. 5D, Sup Fig. 4). Many of the drugs we tested in this proof-of-concept study are clinically approved for other conditions (Table 1), unlocking the potential of using drug repurposing to fast-track development of an anti-HAdV therapeutic (Dodge et al., 2021; Adalja and Inglesby, 2019). Traditional de novo drug development demands a high financial cost that can be attenuated through drug repurposing allowing an accelerated transition of therapy (Sahoo et al., 2021; He et al., 2024; Verwimp et al., 2025; Rodrigues et al., 2022). This strategy is becoming more frequently implemented for novel therapeutic assessment. Even the compassionate use of cidofovir and ganciclovir for treatment of HAdV infections exemplified the utility of drug repurposing (Adalja and Inglesby, 2019). In order to extend our findings and further develop strategies for combination drug therapy for HAdV infection, future work should expand to other HAdV species, especially those commonly associated with severe infection and outbreaks (Lynch and Kajon, 2016; Cai et al., 2024; Saint-Pierre Contreras et al., 2023; Gray and Chorazy, 2009). Furthermore, understanding the true biological mechanism of action behind how these drugs interfere with HAdV, both alone and in combination, is paramount to therapeutic development and should form a focus for future work. By identifying the specific host cell aspects that HAdV infection is uniquely sensitive to, these pathways could be targeted with more precise drugs. Future work should involve primary human airway epithelial or human lung fibroblasts as non-transformed models for HAdV infection as more representative cellular models for HAdV infection and to better understand mechanistic impacts of drugs and combinations (Yang et al., 2024; Sequeira et al., 2025; Prusinkiewicz et al., 2020). Beyond these, the use of organoid models would provide a more relevant context of tissue infection to improve our understanding of infection spread and potential toxicities of these drugs/combinations (Zamzamy et al., 2026). Finally, future evaluation of synergistic drug combinations should implement small animal models like the immunocompromised Syrian hamsters to ensure this approach is viable in vivo (Toth et al., 2008; Bertzbach et al., 2021). In summary, over the past 15 years, many compounds, including the ones we investigated here, have been evaluated for their ability to disrupt HAdV replication (Dodge et al., 2021). Our findings contribute to the growing body of evidence identifying HDTs with efficacy against HAdVs, supporting previous research on the antiviral capacity of these drugs (Carretero-Ledesma et al., 2025; Grosso et al., 2017; King et al., 2020; Jans and Wagstaff, 2020; Romanowski et al., 2021; Balsera-Manzanero et al., 2023; Xu et al., 2020a, 2020b, 2021; Marrugal-Lorenzo et al., 2019). Further research into the clinical efficacy of these drug combinations for acute and severe HAdV infection is required. Overall, this work represents a proof of concept that synergetic combinations of repurposed drugs at low concentrations are capable of inhibiting HAdV-C5 replication, suggesting that this could be a viable approach to develop new therapeutic strategies for patients with HAdV infections. Supplementary Material Funding statement This work was supported by a grant from the Canadian Institutes of Health Research (MOP-148689 and PJT-204110) to J.S.M. J.B.W is supported by two National Institute of Health grants (R21 AI163720 and R01 AI184463). K.M.M. holds a Natural Sciences and Engineering Research Council CGS-D award. A.M.E holds a Canadian Institutes of Health Research CGS-D award. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.antiviral.2026.106463. Footnotes CRediT authorship contribution statement Mackenzie J. Dodge: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. Katelyn M. MacNeil: Conceptualization, Writing – review & editing. Tanner M. Tessier: Writing – review & editing. Andris M. Evans: Writing – review & editing. Kaitlyn R. Palmateer: Writing – review & editing. Jason B. Weinberg: Writing – review & editing. Joe S. Mymryk: Conceptualization, Funding acquisition, Supervision, Writing – review & editing. Declaration of competing interest The authors have no competing interests to declare. Data availability Data will be made available on request. References Associated Data Supplementary Materials Data Availability Statement Data will be made available on request.