An Engineered Self-biomineralized Oncolytic Adenovirus Induces Effective Antitumor Immunity and Synergizes With Immune Checkpoint Blockade

Oncolytic viruses preferentially infect tumor cells and can be utilized for cancer therapy. Through natural tropisms and engineered selectivity, these viruses replicate in tumor cells, leading to tumor cell lysis, the release of pathogen- and damage-associated molecules, and ultimately T-cell priming with tumor and viral antigens. Clinical trials have indicated that oncolytic adenoviruses (oADV) are a promising category of anticancer biotherapeutic agents (1). Many clinical trials with either bare oADV or in combination with chemotherapy or immunotherapy have been conducted (2). However, due to native tropism, severe liver damage, preexisting antiadenovirus immunity, and innate immune response, the clinical efficacy of intravenous injection of the virus is limited (3, 4). Furthermore, numerous studies have shown that primary cancer cells express low levels of coxsackievirus and adenovirus receptor (CAR), which is a cell surface protein to which adenoviruses bind in order to enter cells, and are resistant to oADV infection (5–7). Neutralizing antibodies can also limit oADV efficacy in patients with previous virus exposure or upon repeated virus injections (8).

To address these challenges, oADVs have been physically or chemically modified with various polymers to shield the oADV surface, resulting in increased blood circulation time and decreased immunogenicity and hepatotoxicity (9–11). Virus surface engineering holds the promise of chemically introducing artificial shells on the viral surface, providing a “stealth” cover for evasion of preexisting immunity (12–14). By using reverse genetic technology, the viral capsid can be rationally engineered to improve its physicochemical properties (15–17). In our recent study, a new oADV, OA@CuMnCs, was constructed by combining metal cations Cu²⁺ and Mn²⁺ to form a biomineral layer on the surface of the virus, allowing for intravenous injection and exerting a therapeutic effect on tumors (18).

Many species in nature use biomineralization to enhance their ability to function under challenging conditions. However, most organisms do not possess the ability to self-biomineralize (19). Thus, there are specific challenges in creating a functioning biomineral shell for organisms, including adenoviruses, using biomimetic mineralization techniques. The primary component of bones and teeth calcium phosphate (CaP) has attracted much attention in biogenic mineral research due to its remarkable biocompatibility, and it is currently utilized as an adjuvant and transfer agent (20). CaP represents an excellent mineral shell candidate to stabilize vaccines (17, 21).

In our present study, we aimed to endow adenovirus with the ability to self-biomineralize by adding the nucleopeptide W6p to the surface, enhancing its capacity to initiate CaP mineralization. W6p is an analogue of the amino-terminal 15-residue fragment of salivary statherin and core motifs of dentin matrix protein 1, an acidic protein that can trigger the formation of CaP in vitro by binding calcium ions (22). We generated a biomineralized oADV-W6 with the amorphous CaP coating, hereafter referred to as oADV-W6-CaP, to address the limitation of CAR-dependent cell entry and to increase the transduction efficiency of oADV. We showed that the oADV-W6-CaP system protects the adenovirus from neutralization by preexisting antibodies. Systemic administration of oADV-W6-CaP evaded the innate immune response and achieved significant antitumor effects in several mouse subcutaneous tumor models. We also investigated the efficacy of the combined application of oADV-W6-CaP and PD1 or CD47 blockade in murine tumor models.

The cell lines HEK293 (Cat. # CRL1573), SW1990 (Cat. # CRL2172), BxPC3 (Cat. # CRL1687), MCF7 (Cat. # CRL3435), MDA-MB-231 (Cat. # HTB26), 4T1 (Cat. # CRL2539), and CT26 (Cat. # CRL2638) were purchased from the ATCC. The Panc02 and SCC7 cell lines were maintained in our laboratory. These cell lines were authenticated in 2022 using the PowerPlex 16 System (Promega). 4T1/CAR1 was created by stably transfecting 4T1 cells with a CAR-encoding lentivirus. All cells were regularly validated to be Mycoplasma free. The cell lines were grown in DMEM (Cat. # 11965092, Gibco–Thermo Fisher Scientific) supplemented with 10% FBS (Cat. #16000044, Gibco). All cells were cultivated in an incubator at 37°C and 5% CO₂.

oADVs were generated as previously described (23). Briefly, target sequences containing viral E1A (GenBank: MH629744.1) and enhanced GFP (EGFP) were fully synthesized by GeneScript and subcloned into the pShuttle (pENTER/D-TOPO) plasmid. The ENTR plasmid was further subcloned into the pAd/PL-DEST (Invitrogen) human adenovirus type 5 backbone to generate an oADV vector. These expression vectors were linearized by Pac-I digestion and transfected into HEK293 cells for virus generation. The oncolytic viruses were identified, amplified, purified, and titered according to our previously described method (23). The coding nucleotides of the engineered peptide W6 (RWRLEGTDDKEEPESQRRIGRFG) were cloned into the infectious full-length cDNA of the oADV using standard DNA recombination techniques. oADV-W6 was propagated in HEK293 cells followed by cesium chloride (Sigma) density purification. The infectious titers [plaque-forming unit per milliliter (PFU/mL)] were determined by limiting dilution assay in HEK293 cells. The multiplicity of infection (MOI) was calculated from infectious titers.

Virus solutions (10⁸–10¹⁰ PFU/mL) in DMEM supplemented with 5.5 mmol/L calcium chloride (Merck) were incubated at 37°C for 6 hours. The biomineralized viruses (oADV-W6-CaP) were isolated by centrifugation at 18,000 × g for 10 minutes. The biomineralization efficacy was assessed by the ratio of viral particles in the supernatant and precipitate using plaque assays. Briefly, the virus sample was diluted in a gradient and the diluent added dropwise to the cell culture well for 1 to 2 hours of incubation. Then semisolid agar containing maintenance medium was added to the culture well. After the formation of visible plaque, the cell structure was fixed with formaldehyde while killing the cells and virus; crystal violet stained the nucleus purple, whereas the acellular part was transparent.

The particle size and potential of oADV-W6-CaP and oADV-W6 were determined by dynamic light scattering (Malvern Instruments). The morphology of the samples was observed using an 80 kV transmission electron microscope (JEM-2100F) with energy-dispersive spectrometer elemental mappings (Super-X EDS, Thermo Fisher), and the surface elements and their valence states were analyzed by X-ray photoelectron spectroscopy (Thermo Fisher Scientific).

Electron microscopy characterizations: The viral solutions were added onto carbon-coated copper transmission electron microscopy (TEM) grids (400 mesh; Agar Scientific) by dip coating, and the samples were then dried at room temperature before observation. The samples were also stained with phosphotungstic acid. TEM was performed using a JEM-1200EX microscope (JEOL). Scanning electron microscopy (SEM) was performed using a JSM-35CF microscope (JEOL). Samples were prepared for SEM by dispersing 50 μL of solution on the surface of silica specimen stubs. They were dried at 30°C for at least 24 hours and were sputter-coated with gold before examination.

To evaluate the cancer cell killing efficacy of naked oADV and oADV-W6, various cancer cell lines (SW1990, BxPC3, MCF7, MDA-MB-231, Panc02, CT26, SCC7, 4T1, and 4T1/CAR) were grown in 96-well plates until they reached 70% to 80% confluence. Cells were then infected with either naked oADV or oADV-W6 at various MOIs, ranging from 0.01 to 10. After 24, 48, 72, or 96 hours of infection, 100 μL MTT (2 mg/mL; Sigma-Aldrich) in PBS was added to each well. After 4 hours of incubation at 37°C, MTT formazan formed by metabolically viable cells was solubilized with 200 mL DMSO. Absorbance at 540 nm was read on a microplate reader. All assays were performed in triplicate.

For immunologic detection of oADV coat proteins, naked oADV, oADV-W6, oADV-CaP, or oADV-W6-CaP (10³ PFU in 20 μL) was spotted onto a nitrocellulose membrane (Millipore) and then air-dried at room temperature. The membrane was blocked using 5% skimmed milk in PBS. The membrane was then probed with an antibody specific to adenovirus hexon (ab316852, Abcam), followed by a horseradish peroxidase–conjugated anti-mouse antibody (ab6728, Abcam). Both antibodies were diluted in blocking solution (0.05% BSA, Beyotime). Signals were generated by the addition of the enhanced chemiluminescence developer (Pierce) to visualize the immunoreactive bands.

Western blotting was carried out in accordance with a previously reported methodology (23). The cell pellet was washed once with cold PBS and then lysed with RIPA (Thermo Fisher Scientific), which contained a protease inhibitor (P1005, Beyotime) and a PhosSTOP phosphatase inhibitor cocktail (Roche). The lysates were centrifuged at 4°C for 15 minutes at 14,000 rpm, and the clear supernatant was collected. SDS-PAGE was used to separate the proteins, which were then electroblotted onto polyvinylidene fluoride membranes (Millipore) and probed with primary antibodies (ab6728, Abcam). The main antibody for Western blotting was directed against CAR (Abcam). An enhanced chemiluminescence reagent (Pierce) was used to visualize the immunoreactive bands.

MDA-MB-231, MCF7, SW1990, and BxPC3 cells (5 × 10⁴ cells) were plated in a 24-well plate. After 24 hours, the cells were incubated in FBS-free DMEM with PBS or CAR-specific antibody (5 mg/mL, ab272711, Abcam) at 4°C for 1 hour. Naked oADV, oADV-W6, or oADV-W6-CaP was added to the media and incubated at 37°C for 1 hour. The cells were then washed three times with PBS and incubated in DMEM containing 5% FBS. After incubating for 24 hours, the cells were observed using a fluorescence microscope (Nikon), and EGFP expression was quantified by flow cytometry.

MDA-MB-231 cells (1 × 10⁵ cells) were pretreated for 30 minutes with chlorpromazine (S5749, Selleck Chemicals), dynasore (S8047, Selleck), methyl-β-cyclodextrin (S6827, Selleck), genistein (S1342, Selleck), nocodazole (S2775, Selleck), dihydrocytochalasin B (HY16928, MedChemExpress), or amiloride (S1811, Selleck) diluted in FBS-free DMEM at the desired concentrations in each well of a 12-well plate. oADV-W6 or oADV-W6-CaP was added in the presence of the inhibitors at an MOI of 5 for an additional 2 hours. Afterward, the cells were washed three times with PBS and incubated in DMEM containing 5% FBS. After incubating for 24 hours, EGFP expression was quantified by flow cytometry.

Antiadenovirus serum was diluted in DMEM at ratios of 1:100, 1:50, and 1:10. oADV-W6 or oADV-W6-CaP was mixed with 50 μL of anti-Ad5 serum at an MOI of 1, and the mixtures were incubated at 37°C for 1 hour. The mixtures were then added to 60% confluent MDA-MB-231, MCF7, SW1990, and BxPC3 cells. Cells were observed by fluorescence microscopy, and GFP expression was quantified by flow cytometry.

oADV, oADV-W6, oADV-CaP, or oADV-W6-CaP (1 × 10⁹ PFU) was incubated at 26°C or 37°C, and samples were collected periodically. The remaining infectivity was determined by plaque assays, as described above.

qRT-PCR was used to detect the expression levels of C11r, Ifnα, Ifnβ, Ifnar1, Ifnar2, Ccl5, Cxcl9, Cxcl10, and Cxcl11. Total RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer’s protocol, and cDNA was generated from 1 μg of total RNA using a PrimeScript RT Reagent Kit (TaKaRa). The Gapdh gene was used as the reference gene to normalize the expression level between samples, and samples were calculated using the 2^−ΔΔCT method. qRT-PCR was performed using a 7500 system (Thermo Fisher Scientific). The primer pairs used to detect the mRNA levels are listed in Supplementary Table S1.

Six-month-old male white rabbits and 6-week-old C57BL/6 or BALB/c mice were procured from the Shanghai Laboratory Animal Resource Center and kept in a laminar air flow cabinet under pathogen-free conditions. The Association for Assessment and Accreditation of Laboratory Animal Care accredited all facilities. All animal experiments were approved by the Animal Care and Use Committee of the Zhejiang Provincial People’s Hospital and carried out in accordance with the guidelines provided by Zhejiang Provincial People’s Hospital’s Ethics Committee. To generate antiadenovirus serum, we injected inactivated adenovirus into the rabbit muscle for a total of five immunizations to enhance the immune response of antibodies. Blood was collected from the ear vein of a rabbit to isolate serum containing antibodies. Antibodies were separated and purified from the collected serum. 4T1, CT26, Panc02, and SCC7 cancer cells were subcutaneously implanted in the left abdomen of each mouse (see figure 4 for the number of cells implanted per mouse). When the tumor grew to 50 to 100 mm³, intravenous injections of PBS, oADV, oADV-W6, or oADV-W6-CaP were initiated. Every other day for 4 days (days 0, 2, 4, and 6), an intravenous dose of oADV, oADV-W6, or oADV-W6-CaP was administered at the concentration of 2.5 × 10⁹ PFU/mouse. Tumor growth was monitored by measuring the length and width of each tumor using a caliper every 2 or 3 days until the end of study. Tumor volume was calculated as volume = 0.5 × length × width². The body weight of the mice was measured every 2 or 3 days. Kaplan–Meier survival analysis of tumor-bearing mice treated with PBS (control), oADV-W6, or oADV-W6-CaP was performed using Prism 8.2.1 software.

In some experiments, CD8⁺ T or NK cells were depleted. To achieve this, each mouse was injected intraperitoneally with 500 μg of anti-CD8 (cone YTS 169.4, Cat. # BP0117, Bio X Cell) or anti-NK1.1 (clone PK136, Cat. # BP0036, Bio X Cell). Flow cytometry was used to determine whether the in vivo depletion was effective.

For the combination therapy with oADV-W6-CaP and anti-mouse PD1 (clone RMP1-14, Cat. # BE0146, Bio X Cell), a subcutaneous Panc02 model was established as described above. Each mouse was injected intraperitoneally with 200 μg of anti-PD1, which was initiated from the next day of viral treatment and continued five times every 2 days. For the rechallenge assay, mice with complete tumor clearance and seven age-matched treatment-naïve mice were subcutaneously rechallenged with Panc02 cells (5 × 10⁵ cells per mouse). For the combination therapy with oADV-W6-CaP and anti–mouse CD47 (clone MIAP410, Cat. # BE0283, Bio X Cell), a subcutaneous 4T1 model was established as described above. Each mouse was injected intraperitoneally with 100 μg of anti-CD47, which was initiated 1 day after the first viral injection and performed six times every 2 days. Statistical significance of tumor volume was calculated by one-way ANOVA, followed by the Dunnett multiple comparisons test. The log-rank test of the survival curve was performed to determine statistical significance among groups.

The following antibodies were purchased from BioLegend: PE anti–mouse CD45 (clone 30-F11, Cat. # 103106), APC anti–mouse CD3 (clone 17A2, Cat. # 100236), FITC anti–mouse CD4 (clone GK1.5, Cat. # 100406), PE anti–mouse CD4 (clone GK1.5, Cat. # 100408), FITC anti–mouse CD8α (clone 53-6.7, Cat. # 100706), PE/Dazzle 594 anti–mouse CD279 (PD1; clone 29F.1A12, Cat. # 135227), PE/Cyanine7 anti–mouse CD152 (CTLA4; clone UC10-4B9, Cat. # 106313), Brilliant Violet 711 anti–mouse CD366 (TIM3; clone RMT3-23, Cat. # 119727), Alexa Fluor 647 anti–mouse CD233 (LAG3; clone C9B7W, Cat. # 125241), APC/Fire 750 anti–mouse IFNγ (clone XMG1.2, Cat. # 505860), Biotin anti–mouse CD107a (clone 1D4B, Cat. # 121603), APC/Fire 750 anti–mouse CD49a (clone Miap301, Cat. # 142609), APC/Fire 750 anti–mouse CD11c (clone N418, Cat. # 117352), APC anti–mouse/human CD11b (clone M1/70, Cat. # 101212), FITC anti–mouse Ly6G/Ly6C (Gr1; clone RB6-8C5, Cat. # 108406), PerCP anti–mouse F4/80 (clone BM8, Cat. # 123126), PerCP anti–mouse CD86 (clone GL1, Cat. # 105026), APC anti–mouse CD80 (clone 16-10A1, Cat. # 104713), FITC anti–mouse CD206 (clone C068C2, Cat. # 141704), PE/Fire 700 anti–mouse CD206 (clone C068C2, Cat. # 141741), FITC anti–mouse CD25 (clone 3C7, Cat. # 101907), PE anti–mouse Foxp3 (clone MF14, Cat. # 126404), PE anti–mouse/human CD44 (clone IM7, Cat. # 103008), FITC anti–mouse CD62L (clone MEL14, Cat. # 104406), PE anti–mouse CD47 (clone MP6-XT22, Cat. # 506322), APC anti–mouse PDL1 (clone 10F.9G2, Cat. # 124311), and APC anti–mouse NK1.1 (clone S17016D, Cat. # 156505).

For the preparation of single-cell suspensions of tumor tissues, mice were anesthetized and sacrificed, and the tumor tissues were harvested and placed in serum-free medium (Cat. # H740KJ, Basal Media) with 0.2% collagenase IV (Cat. # C4-22-1G, Sigma-Aldrich, Germany). The tumor tissues were then cut into 1 to 2 mm pieces, digested for 2 hours, and passed through 70-µm nylon filters (Cat. # CSS013070, JetBiofil) to obtain single-cell suspensions. Then, collagenase was removed by centrifugation, and the cell pellets were suspended in serum-free medium and adjusted to 2 × 10⁷ cells/mL.

For the preparation of splenocytes, spleens were obtained from sacrificed mice and ground into single-cell suspensions using syringe plungers on 70-µm nylon filters (Cat. # CSS013070, JetBiofil). The cells were then counted and adjusted to 5 × 10⁶ cells/mL.

For extracellular staining, single-cell suspensions were incubated with the fluorescent mAbs for 15 minutes at room temperature. After incubation, the cells were fixed with 4% (w/v) paraformaldehyde (Cat. # G1101, Servicebio) solution and directly analyzed using a NovoCyte cytometer (Agilent). For intracellular staining, fixed cells were ruptured with 1× permeabilization buffer (Cat. # 00-8333-56, eBioscience), and then the corresponding antibodies were added, respectively, and incubated for 30 minutes in the dark. After washing with PBS once, an appropriate amount of PBS was added to resuspend the cells and was immediately analyzed by flow cytometry. Data analysis was performed using FlowJo software (version 10.8.1; TreeStar).

4T1 tumor–bearing mice were euthanized 12 days after injection, and the heart, liver, spleen, lungs, and kidneys were collected. The organs were fixed with 4% paraformaldehyde solution, embedded in paraffin, and cut into 5-μm sections for hematoxylin and eosin staining. In brief, the sections underwent baking, dewaxing, hematoxylin staining, eosin staining, and sealing to observe the internal morphology and structure of the cells. The hematoxylin staining solution is alkaline, which mainly causes the chromatin in the nucleus and nucleic acids in the cytoplasm to turn purple-blue; eosin is an acidic dye, which primarily causes components in the cytoplasm and extracellular matrix to turn red. For IHC, formalin fixed paraffin embedded sections were incubated with a secondary horseradish peroxidase goat anti–rabbit IgG antibody (G1213, Servicebio) and a goat anti–hexon antibody (AB1056F, Millipore). Then, the slides were stained with 3′,3-diaminobenzidine (G1212-2, Servicebio) and counterstained with 37% (w/v) hematoxylin (G1004, Servicebio).

Following a rechallenge experiment with previously cured mice, the mice were sacrificed under anesthesia to collect the spleens and prepare splenocytes. Panc02, 4T1, CT26, and SCC7 cells were plated at a density of 2 × 10⁵ cells per well in a six-well plate, and 1 × 10⁶ splenocytes were added to make the splenocyte-to-tumor cell ratio 5:1. After 48 hours of coculture, samples were centrifuged at 600 × g for 5 minutes at room temperature, and the supernatants were collected. To assess the concentration of the respective cytokines, ELISA Max Standard Sets of IFNγ (Cat. # 430801, BioLegend), TNFα (Cat. # 430901, BioLegend), and IL2 (Cat. # 431001, BioLegend) were used. The data were collected on an SpectraMax Mini Microplate Reader (Molecular Devices), according to the manufacturer’s instructions.

Blood samples were obtained from mice and centrifuged at 3,000 to 4,000 rpm for 10 minutes, and the supernatant was collected. The reagent kits for the detection of serum biochemical indicators were all obtained from Ningbo Purui Bo Biological Technology Co. Ltd.: aspartate aminotransferase (Cat. # AST01), alanine aminotransferase (Cat. # ALT01), alkaline phosphatase (Cat. # ALP01), total bilirubin (Cat. # TBI01), direct bilirubin (Cat. # DBI01), thiobarbituric acid (Cat. # TBA01S), γ-glutamyl transferase (Cat. # GGT01), blood urea nitrogen (Cat. # URE01), creatinine (Cat. # CR01), and uric acid (Cat. # UA01). The assays were performed according to the kits’ instructions using a fully automatic biochemical analyzer LW C400 (Landwind Medical Holdings Ltd.).

Prism 8.2.1 (GraphPad Software) was used for all statistical analyses. ANOVA was used to investigate the statistical differences between the groups. The survival curve was generated using the Kaplan–Meier method, and the log-rank test was performed to determine statistical significance among groups. P < 0.05 was considered statistically significant in all statistical studies.

The data generated during the study are available in the article and its Supplementary Materials or from the corresponding author on reasonable request.

An oADV was constructed by integrating a calcium chelating agent (W6p) onto the virion surface to boost its ability to induce CaP mineralization. The location at the 3′-end of IX protein has been shown to be an acceptable insertion site for the viral surface deployment of the designed peptide (Fig. 1A). Using typical DNA recombination methods, the W6p-coding nucleotides were cloned into the infectious, full-length cDNA of attenuated oADV. The MTT assay revealed no statistically significant difference in oncolytic activity between oADV-W6 and oADV (Fig. 1B). Furthermore, oADV-W6 and oADV exhibited similar replication competence in tumor cell lines (Fig. 1C). These results suggest no alteration in the reproductive and oncolytic capacities of the recombinant virus.

The self-biomineralization capacity of the modified oADV-W6 was tested in calcium-enriched DMEM. The biomineralization efficacy was then assessed using plaque tests based on the ratio of virus particles in the supernatant and precipitate following normal-speed centrifugation. Dot blot assays revealed that the biomineralized oADV-W6 was barely detectable using adenovirus-specific antibodies. oADV-W6-CaP is shown in Fig. 1D. Because of mineral inclusion, centrifugation may separate out more than 80% of infectious virions. Almost no infectious viral particles were precipitated from native oADV, and the majority of virions remained in the supernatant (Fig. 1E). This result implies that the capsid proteins were shielded by the precipitated mineral phase.

SEM and TEM revealed that there was a slight increase in average particle size from 80.3 to 108.7 nm following CaP biomineralization (Fig. 1F and G; Supplementary Fig. S1). The zeta potential of oADV-W6-CaP was approximately 2.68 mV (Fig. 1H). TEM mapping was used to analyze the elemental distribution of oADV-W6-CaP nanoparticles, revealing a uniform distribution of elements C, O, P, and Ca (Fig. 1I). Additionally, the presence of the CaP coating layer on the oADV-W6-CaP nanoparticles was confirmed by surface X-ray energy-dispersive spectra (Fig. 1J). The release of the enclosed virus from the mineral coating was also demonstrated by dot blot assays in acidic solutions (Fig. 1K). This pH-sensitive feature is ideal for biomedical applications, as slightly acidic conditions are frequently involved in the intracellular processing of internalized particles.

To evaluate the transduction efficiency of oADV-W6-CaP, we selected eight cell lines with different endogenous CAR expression levels (SW1990, BxPC3, MDA-MB-231, MCF7, Panc02, CT26, SCC7, and 4T1) as well as 4T1/CAR (4T1 cells transfected with CAR antigen) for further study (Fig. 2A and B). Cells were transduced with five MOIs of naked oADV-W6 or oADV-W6-CaP for 24 hours, and the EGFP intensity was quantified by flow cytometry analysis (Fig. 2C). These results demonstrate that oADV-W6-CaP was transduced in most cancer cell lines at significantly higher efficiency levels than naked oADV-W6. These data suggest that oADV-W6-CaP enhanced transduction efficiency in cancer cells regardless of CAR expression level.

We then investigated whether oADV-W6-CaP caused an increase in cancer cell death due to rapid cellular uptake. An MTT assay was performed with various MOIs of naked oADV-W6 or oADV-W6-CaP in different cancer cell lines. As shown in Fig. 2D, oADV-W6-CaP induced a higher level of cancer cell death than naked oADV-W6 at the tested MOIs. We observed that increased cancer cell death was more pronounced in the CAR low expression cancer cell lines (Panc02, CT26, SCC7, and 4T1) compared with the CAR-positive lines (SW1990, BxPC3, MDA-MB-231, MCF7, and 4T1/CAR). Cell death was inversely proportional to the expression level of CAR, which was consistent with the enhanced transduction efficiency of oADV-W6-CaP. These results demonstrate that the ability of oADV-W6-CaP to induce cancer cell death was enhanced with CaP biomineralization.

To further explore whether oADV-W6-CaP could bypass the requirements for CAR-mediated endocytosis, we performed a competition assay with a CAR-specific antibody in MDA-MB-231 cells. Pretreatment with anti-CAR decreased EGFP expression in cells transduced with naked oADV or oADV-W6 by 53.4% and 59.2%, respectively. By contrast, EGFP expression in cells transduced with oADV-W6-CaP was only partially inhibited, with a decrease in EGFP expression of 9.97% with the anti-CAR (Fig. 3A). Similar changes in EGFP expression occurred in other tumor cell lines expressing high levels of CAR (Supplementary Fig. S2A). These data suggest that cellular uptake of oADV-W6-CaP was not completely blocked by competition with a CAR-specific antibody. Thus, entry or internalization of oADV-W6-CaP was not predominantly dependent on CAR-mediated endocytosis.

We further investigated the internalization mechanism of oADV-W6-CaP in comparison to that of naked oADV-W6. Cells were pretreated with chlorpromazine, dynasore (inhibitor of clathrin-mediated endocytosis), methyl-β-cyclodextrin, genistein (inhibitor of caveolae-mediated endocytosis), nocodazole, dihydrocytochalasin B (inhibitor of microtubule-mediated endocytosis), or amiloride (inhibitor of macropinocytosis-mediated endocytosis). EGFP expression levels were measured after transduction with naked oADV-W6 or oADV-W6-CaP. As shown in Fig. 3B, methyl-β-cyclodextrin, genistein, and amiloride significantly reduced the transduction efficiency of oADV-W6-CaP, achieving 32.71%, 30.53%, and 39.8% inhibition, respectively. In combination with the results of the competition assay with the anti-CAR, we conclude that oADV-W6-CaP primarily entered cells through caveolae-mediated endocytosis and macropinocytosis-mediated endocytosis.

The adverse impacts of preexisting immunity on oADV include both reduced transgene expression and compromised antitumor immune responses (16). The most effective characteristic for an engineered oADV to have to circumvent preexisting immunity is the ability to stealthily bypass inactivation. In this context, we evaluated the ability of oADV-W6-CaP to evade neutralizing anti-Ad5 serum in vitro, as monitored by the cellular fluorescent signals expressed by the EGFP transgene. As we expected, EGFP expression of native oADV-W6 in MDA-MB-231 cells was blocked when the cultures were exposed to high levels of anti-Ad5 serum. However, the EGFP expression of oADV-W6-CaP remained essentially unaffected in the presence of anti-Ad5 serum, even at high titers (Fig. 3C). Similar changes in EGFP expression occurred in other tumor cell lines (Supplementary Fig. S2B). This result implies that the CaP shell masked the oADV from serum neutralization in vitro, suggesting direct ablation of the recognition between the virus’ immunodominant epitopes and neutralizing serum. Consequently, the CaP shell functioned to circumvent anti-Ad5 immunity in vitro, resulting in efficient cellular uptake and transgene expression.

The thermostability of oADV-W6-CaP was also studied. oADV-W6-CaP, oADV-W6, and parental oADV were incubated at room temperature for varying number of days, and their subsequent infectivity was titrated by plaque assays. The results showed that after self-biomineralization, the resulting oADV-W6-CaP exhibited a significantly slower inactivation rate, and its storage could be prolonged to more than 8 days at 26°C (Fig. 3D). Accelerated degradation tests with samples subjected to physiologic temperature (37°C) were performed to estimate the loss of oncolytic virus potency (Supplementary Fig. S3A). Unlike oADV-W6, the exposure of native oADV to CaP did not improve thermostability due to its poor capacity to self-biomineralize (Supplementary Fig. S3B). These results indicate that the thermal stability of the oADV-W6 was greatly improved by the biomineralization of CaP, and this improvement was directly related to the biomineralization efficacy.

Infection with oncolytic viruses leads to the activation of the type I IFN signaling pathway, which is crucial in oncolytic virus–mediated antitumor immunity. Our results showed that oADV-W6-CaP treatment significantly promoted Ifna and Ifnb expression in mouse breast cancer 4T1 cells (Fig. 3E). The levels of IFNAR subunits Ifnar1 and Ifnar2, type I IFN receptors, were unchanged in response to infection with oADV-W6-CaP (Fig. 3E). Furthermore, oADV-W6-CaP infection in 4T1 cells resulted in a significant increase in proinflammatory chemokines, such as CCL5, CXCL9, CXCL10, and CXCL11 (Fig. 3F), which are crucial for regulating T-cell recruitment and activity. Similar results were observed in the other three mouse tumor cell lines, Panc02, CT26, and SCC7 (Supplementary Fig. S4). These data suggest that oADV-W6-CaP infection triggers a strong proinflammatory response in mouse cancer cells.

We next evaluated the antitumor activity of oADV-W6-CaP in subcutaneous tumor models (4T1, CT26, Panc02, and SCC7) in immunocompetent BALB/c or C57BL/6 mice (Fig. 4A). Although the intravenous administration of oADV-W6 showed superior antitumor effects compared with PBS in the CT26 model in terms of inhibiting tumor growth and prolonging the survival of mice, it did not show any superior antitumor effects compared with PBS in the 4T1, Panc02, and SCC7 models (Fig. 4B; Supplementary Fig. S5). In all models, mice treated with oADV-W6-CaP had significantly lower tumor volume and prolonged survival compared with mice treated with oADV-W6, oADV, or PBS (Fig. 4B and C). There was no significant difference in body weight among the several groups of mice (Fig. 4D). Evaluation of liver and kidney functions of 4T1 tumor–bearing mice treated with oADV-W6-CaP included analysis of histopathologic changes by hematoxylin and eosin assay and serum biochemical analysis of aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, total bilirubin, direct bilirubin, thiobarbituric acid, γ-glutamyl transferase, blood urea nitrogen, creatinine, and uric acid (Supplementary Fig. S6). These data indicate that oADV-W6-CaP is superior to oADV-W6 and parental oADV in reducing tumor burden and prolonging the survival of tumor-bearing mice.

We evaluated the infiltration and activation of immune cells 1 day after the last intravenous injection of oADV-W6-CaP in the 4T1 subcutaneous tumor model (Supplementary Fig. S7). Flow cytometry analysis showed that treatment with oADV-W6-CaP significantly increased tumor infiltration by CD3⁺ T cells, CD4⁺ T cells, and CD8⁺ T cells compared with either oADV-W6 or PBS but did not cause obvious changes in NK cells (Fig. 5A). We then evaluated the expression of immune checkpoints (PD1, CTLA4, TIM3, and LAG3) and activation markers (IFNγ and CD107A) on T cells. Compared with PBS, the injection of oADV-W6 increased the frequency of only TIM3⁺ cells in the tumor-infiltrating CD8⁺ T-cell population, whereas injection of oADV-W6-CaP significantly increased the frequency of PD1⁺, CTLA4⁺, TIM3⁺, LAG3⁺, IFNγ⁺, and CD107A⁺ cells in the tumor-infiltrating CD8⁺ T-cell population (Fig. 5B and D) and increased the frequency of PD1⁺ cells in the tumor-infiltrating CD4⁺ T-cells population (Fig. 5C). Moreover, oADV-W6-CaP significantly increased the proportion of CD8⁺ T and CD4⁺ T cells in splenocytes (Supplementary Fig. S8A).

We also assessed the activation of dendritic cells (DC), and the results showed that oADV-W6-CaP treatment significantly increased the expression of CD80, CD86, and MHC class II in DCs (Fig. 5E), suggesting enhanced maturation and a greater capacity of DCs to present antigens. The number of macrophages, especially the proportion of M1-like macrophages, was increased significantly in tumors with oADV-W6-CaP treatment (Fig. 5F). We also found that intratumoral CD11b⁺Gr1⁺ myeloid-derived suppressor cells and CD4⁺CD25⁺FoxP3⁺ regulatory T cells were reduced by oADV-W6-CaP treatment (Fig. 5G). As we expected, abundant virus titers were observed in tumors of mice treated with oADV-W6-CaP, whereas much lower virus titers were observed in tumors of mice treated with oADV-W6, and no virus was detected in the tumor tissues of PBS-treated mice and in the sera of the three groups of mice (Supplementary Fig. S8B). The viral content in tumor tissues of 4T1 and CT26 mice was quantified by qPCR detection of vaccinia virus C11R gene. These results showed that virus titers were significantly increased in oADV-W6-CaP–treated tumors compared with those treated with oADV-W6 or oADV control (Supplementary Fig. S8C). Altogether, these results indicate that oADV-W6-CaP can reshape the tumor microenvironment (TME) by recruiting immune cells and activating the tumor-infiltrating CD8⁺ T cells.

To determine the role of CD8⁺ T cells and NK cells in mediating the antitumor activity of oADV-W6-CaP, we depleted these two types of lymphocytes in the 4T1 subcutaneous syngeneic model and analyzed whether their depletion had an effect (Fig. 6A). We confirmed that the CD8⁺ T cells and NK cells in the peripheral blood were completely depleted 48 hours after injection of the corresponding mAbs, and their depletion was maintained for 2 weeks (Fig. 6B). Consistent with previous results in this study, the treatment of mice with oADV-W6-CaP significantly decreased tumor volume (Fig. 6C and D) and increased the proportion of CD3⁺ T cells, CD4⁺ T cells, and CD8⁺ T cells in the TME (Fig. 6E). Depletion of CD8⁺ T cells reversed all of these increases in oADV-W6-CaP–treated mice, resulting in levels at or lower than those measured in PBS-treated mice (Fig. 6E). Although the depletion of NK cells also significantly decreased CD3⁺ T cells, CD4⁺ T cells, and CD8⁺ T cells, these cells still showed a trend toward higher frequencies in oADV-W6-CaP–treated mice than in PBS-treated mice (Fig. 6E). Consistent with these results, the depletion of CD8⁺ T cells completely abrogated the antitumor efficacy of oADV-W6-CaP, whereas the depletion of NK cells had no obvious impact. These results indicate that CD8⁺ T cells are a major immune cell type that mediates the antitumor immunity of oADV-W6-CaP.

We next investigated whether infection with oADV-W6-CaP affected the expression of PDL1 and CD47 in 4T1, Panc02, CT26, and SCC7 cells. After 24 hours of treatment, oADV-W6-CaP showed more infection efficiency in the four tumor cell lines than did oADV-W6 (Supplementary Fig. S9). In response to oADV-W6-CaP infection, the expression of PDL1 and CD47 on the surface of the four tumor cell lines was upregulated (Supplementary Fig. S9). Furthermore, we evaluated the antitumor activity of the combined application of oADV-W6-CaP and anti-PD1 in the Panc02 subcutaneous syngeneic model (Fig. 7A). Monotherapy with oADV-W6-CaP and the combination of oADV-W6-CaP and anti-PD1 significantly inhibited tumor growth (P < 0.0001 and P < 0.0001, respectively) when compared with PBS. Compared with mice treated with oADV-W6-CaP or anti-PD1 alone, the combination therapy showed a trend of smaller tumor volume (Fig. 7B). One tenth of oADV-W6-CaP–treated mice and seven tenths of combination-treated mice achieved tumor complete regression (CR), whereas none of the mice treated with PBS or anti-PD1 alone achieved CR (Fig. 7C). Consistent with these results, monotherapy with oADV-W6-CaP or anti-PD1 significantly prolonged the survival time of tumor-bearing mice (P < 0.05 or P < 0.01, respectively) when compared with PBS. Moreover, combination therapy further prolonged the survival time of tumor-bearing mice (P < 0.05 vs. oADV-W6-CaP; P < 0.05 vs. anti-PD1; P < 0.001 vs. PBS; Fig. 7D).

To examine tumor-specific immunologic memory, 7 of 10 mice previously cured by oADV-W6-CaP and anti-PD1 combination therapy were subcutaneously rechallenged with the same tumor cells. All age-matched, treatment-naïve mice developed syngeneic tumors and eventually died after implantation with Panc02 cells, whereas all the previously cured mice did not develop syngeneic tumors after Panc02 rechallenge (Fig. 7E). Consistent with these results, the proportion of naïve CD8⁺ T cells (CD62⁺CD44⁻) in splenocytes of mice that had rejected Panc02 tumor rechallenge decreased (P < 0.001 vs. treatment-naïve mice), whereas total CD8⁺ T cells and their central memory (CD62L⁺CD44⁺) and effector memory (CD62⁻CD44⁺) subsets increased (P < 0.001, P < 0.0001, and P < 0.0001, respectively; Fig. 7F). Similar proportional changes also occurred in the CD4⁺ T cells and in the naïve CD4⁺ T-cell, central memory CD4⁺ T-cell, and effector memory CD4⁺ T-cell subsets (Fig. 7F). Furthermore, coculturing Panc02, 4T1, CT26, or SCC7 cells with splenocytes from the Panc02-rechallenged mice showed that splenocytes stimulated with Panc02 cells secreted higher concentrations of IFNγ, TNF-α, and IL2 than did those stimulated with 4T1, CT26, or SCC7 cells, whereas splenocytes of treatment-naïve mice showed no detectable response to any stimulation (Fig. 7G). These data indicate that the treatment of tumor-bearing mice with oADV-W6-CaP and anti-PD1 combination therapy established long-term, tumor-specific immunologic memory and systemic antitumor immunity.

We next evaluated the antitumor effect of oADV-W6-CaP and anti-CD47 combined therapy in the 4T1 breast cancer model (Fig. 7H). Consistent with the previous results, intravenous injection of oADV-W6-CaP significantly inhibited tumor growth and prolonged the survival time of the tumor-bearing mice, whereas monotherapy with anti-CD47 showed no change compared with treatment with PBS (Fig. 7I). Six of 10 mice treated with oADV-W6-CaP plus anti-CD47 achieved tumor CR (Fig. 7J). These results indicate that adding treatment with oADV-W6-CaP improved the antitumor efficacy of anti-PD1 and anti-CD47 therapy.

In recent decades, oADVs have been extensively explored for cancer gene therapy. The advantages of using oADVs in cancer gene therapy include not only cancer cell–specific replication, the infection of neighboring cancer cells, and the destruction of infected cells but also the high expression of inserted therapeutic genes, leading to potent antitumor effects (24). In addition, oADVs can infect both dividing and nondividing cells, and high viral titers can be achieved. Several human clinical trials have reported the successful application of oADVs in local cancer gene therapy (25). However, the efficacy of intravenous administration of oADV is compromised by nonspecific sequestration in the liver, preexisting antiadenovirus immunity, and the innate immune response (26).

To tackle these problems, two main strategies have been employed. First, genetic engineering strategies have been extensively used, including switching serotypes and modifying fibers with heterologous peptides. Replacing an Ad5 fiber knob with Ad3 can lead to resistance against neutralization by an Ad5-induced antibody (27). A chimeric Ad5/Ad48 system reduces liver tropism after systemic administration (28). However, the removal or exchange of virulence factors usually limits replication in target tissues, and serotype switching is difficult due to the multiple surface-exposed capsid proteins that contain neutralizing epitopes. Fiber modification of adenovirus has been observed to delay viral uptake, which possibly impairs infectivity (29). Consequently, there is an urgent need for alternative strategies to solve these issues. The second strategy that has been used is to construct a hybrid vector system combining viral and nonviral carriers that can overcome some of the challenges that are associated with genetic engineering strategies.

Recent approaches in the modification of Ad have focused on the use of cationic polymers or lipids, such as polyethylenimine, poly-L-lysine, arginine-grafted bioreducible polymer, chitosan, and 1, 2-dioleoyl-3-trimethylammonium-propane chloride salt (9). The cationic moieties in the polymers or lipids can form hybrids with the negatively charged Ad, thereby protecting the Ad from preexisting neutralizing antibodies and facilitating Ad interaction with the cell membrane and increasing in vitro transfection efficiency (29, 30). However, the cationic polymer/liposome-based strategy faces other problems, such as toxicity, low tissue specificity, short in vivo circulation time due to nonspecific interactions with the negatively charged cell membrane, and incompatibility with negatively charged macromolecules that are present in the physiologic environment (29). These problems prompted us to explore other approaches for the systemic administration of oADV for use in cancer gene therapy.

In this study, by genetically engineering potential nucleating peptide W6p onto the surfaces of living oADV, we rationally designed a self-biomineralized virus with improved thermostability and immunogenicity. This CaP biomineralization ability was given to oADV-W6 by genetically engineering a CaP nucleating peptide onto the viral capsid. We showed that after self-biomineralization, oADV-W6-CaP was protected from neutralization by preexisting neutralizing antibodies. oADV-W6-CaP induced efficient tumor-targeting ability and antitumor effects after systemic administration.

The variable level of CAR expression in tumor cells is the main factor impeding effective Ad delivery in vitro and in vivo. In this study, we demonstrated that the transduction efficiency of oADV-W6-CaP was dramatically increased in both CAR-positive and CAR-negative cancer cells. The fold increase in gene transfer efficiency of oADV-W6-CaP was inversely correlated with the level of CAR expression in cancer cells, suggesting that the oADV-W6-CaP complex overcomes the low gene transfer efficiency in CAR-deficient tumors.

Notably, oADV-W6-CaP upregulated the expression of PDL1 in cancer cells, which may be induced by intratumoral IFNγ. Although PDL1 generally downregulates T-cell activity, many reports have shown that patients with high baseline expression of immune-related genes, including PD1/PDL1, or preexisting intratumoral CD8⁺ T cells seem to respond preferentially to anti-PD1/PDL1 and anti-CTLA4 therapies (31). We demonstrated that oADV-W6-CaP administration made 4T1 or Panc02 tumors sensitive to anti-PD1 and anti-CD47, suggesting that oADV-W6-CaP switched from “cold” tumors to “hot” tumors, eliciting a synergistic lethal effect in combination with immune checkpoint blockade.

In conclusion, our data indicate that oncolytic virotherapy using engineered oADV-W6-CaP is an effective strategy for cancer immunotherapy. The intravenous injection of oADV-W6-CaP causes profound reshaping of the TME from “cold” to “hot” status. Furthermore, oADV-W6-CaP also synergizes with PD1 or CD47 blockade to achieve a complete response to tumors with poor response to oADV or immune checkpoint blockade monotherapy (Supplementary Fig. S10).

No disclosures were reported.

S. Wang: Conceptualization, software, supervision, funding acquisition, project administration. X. Yang: Project administration. Y.-Y. Ma: Software. J. Wu: Project administration. K. Jin: Resources, supervision, funding acquisition, validation, project administration. R. Zhao: Resources, supervision, funding acquisition, validation, project administration. H. Zou: Resources, supervision, funding acquisition. X. Mou: Validation, project administration.

We thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript. This study was supported by grants from the Natural Science Foundation of Zhejiang Province (LGF22H080012 and LY19H160034), National Natural Science Foundation of China (82102330 and 81602706), and Zhejiang Provincial Medical Technology Plan Project (No. 2022KY569).