Oncolytic virus with an immune stimulatory payload for osteosarcoma therapy Authors: Sumbul Khan, Theresa A Higgins, Isabella Shimko-Lofano, Maninder Sandey, Emma Logsdon, Gracie Bunch, Yasin Fatemi, Leah M K Hoffman, Bruce F Smith, Payal Agarwal Published: 2026 ABSTRACT AbstractOsteosarcoma (OS) is a highly aggressive bone malignancy that predominantly affects children and young adults. There have been no major changes in patient survival in the past four decades. Therefore, new therapeutic interventions are needed. We have developed an enhanced conditionally replicative canine oncolytic adenovirus, CAV2-AU-M3, armed with an anti-PD1 heavy-chain antibody (HcAb), and evaluated its efficacy against osteosarcoma across four canine cell lines. CAV2-AU-M3 was characterized for its infectivity, lytic properties, and anti-PD1 Ab production in monolayer and spheroid cultures. Additionally, the impact of intratumoral administration of the virus on tumor growth was measured in a mouse model. Our study demonstrates that CAV2-AU-M3 infects and lyses different canine OS cell lines at varying rates but produces anti-PD1 Ab at similar levels across all osteosarcoma cell lines. Additionally, anti-PD1 Ab produced by CAV2-AU-M3 can effectively bind the cell surface PD1 receptor and inhibit the binding of PDL1 to PD1 receptors. Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00262-026-04458-0.Keywords: Osteosarcoma, Oncolytic Viruses, Immune Checkpoint Inhibitors, Tumor Spheroids, Tumor Microenvironment Introduction Osteosarcoma (OS) is a malignant bone tumor that most commonly occurs in children and adolescents (3–5% of childhood cancers) [1–3]. OS predominantly occurs in long bones, with higher frequency near the knees, hips, and shoulders [4, 5]. Despite multidisciplinary treatment, including surgery, chemotherapy, targeted therapy, radiation therapy, and even combination therapies, the survival rates for patients with OS have not improved in the past four decades, especially for patients with metastatic disease [6–11]. The 5-year survival rate is less than 20% in metastatic patients [8, 10]. Cancer immunotherapy stimulates the host’s own immune system against the tumor and turns the immunologically “cold” tumor microenvironment (TME) into “hot.” Immunotherapies provide long-term defense against tumors and enable the immune system to recognize and target tumor cells in the event of recurrence or metastasis. Immunotherapies, including immune checkpoint inhibitors (ICIs), genetically modified T cells, immune modulators, and cytokines, have shown remarkable improvements in cancer treatment in in vivo models with fewer side effects, including OS [12–14]. Immune checkpoints, such as PD1/PDL1 interaction, cause immune evasion in the tumor microenvironment by inactivating cytotoxic T cells [12–14]; and blocking this interaction leads to tumor growth inhibition, Treg decrease, and increased tumor-infiltrating lymphocytes in vitro and in vivo mice OS model [14]. However, systemic delivery of anti-PD1 mAb can cause immune-related adverse effects [15] in organs such as the liver, lungs, kidneys, and heart [16]. It is also possible that systemic anti-PD1 mAb may not reach the solid tumor site due to the tumor stroma [17]. To address these issues, we have developed a conditionally replicative oncolytic canine adenovirus type 2 vector (CAV2-AU-M3; Fig. 1A) armed with a secretory chimeric anti-PD1 heavy-chain-only antibody. Conditionally replicative oncolytic adenoviruses (CRAds) lyse tumor cells and induce systemic anti-viral immunity (DAMPs and cytokine secretion from tumor cells), along with anti-tumor immunity [18–21]. Additionally, oncolytic virus therapy improves anti-PD1 immunotherapy by promoting intratumoral cell infiltration [22] and localized secretion of anti-PD1 Ab to avoid adverse effects [23]. CAV2-AU-M3 is designed to result in local, sustained production and secretion of anti-PD1 Ab into the TME to avoid systemic exposure and cross the tumor stromal barrier. CAV2-AU-M3’s infectious and cytolytic properties in canine osteosarcoma cells in 2D culture. A Graphical representation of ICOCAV15 and CAV2-AU-M3. B Cytopathic effects of CAV2-AU-M3 at 100 MOI in canine OS cell lines (D17 GLP, CF11 GLL, D22 GLP, and MC-KOS) visualized using inverted fluorescent microscopy (Keyence) at day 5 post-infection at 10 × magnification. C Immunofluorescence using mouse anti-CAV2 (VMRD) primary antibody and anti-mouse IgG Texas-red secondary antibody was performed in D17 GLP, CF11 GLL, and MCKOS cell lines and primary NCF cells at 0, 10, 100, and 500 MOIs. Images were captured using a Keyence microscope at 10X 48 h post-infection. D Luciferase assay for cell viability in OS cell lines at 100, 500, 1000, 5000, and 10,000 MOIs at day 5 post-infection. E TCID50/ml values of CAV2-AU-M3 in OS cell lines. F Tabular representation of the data from (E). G Crystal violet assay to demonstrate TCID50/ml values in all OS cell lines. Each row depicts the virus dilutions Previously, we designed a similar oncolytic virus (CAV2-AU-M2) that encoded anti-PD1 VHH, which was released only after cell lysis. CAV2-AU-M3 encodes and secretes the chimeric anti-PD1 Ab, which is more stable due to the presence of the canine Fc region in the antibody [24]. In this study, we have characterized the CAV2-AU-M3 virus, its cytotoxic properties, and its ability to produce and secrete anti-PD1 Ab. Material and methods Cell lines Canine osteosarcoma cell lines (D17, CF11, and D22) were purchased from ATCC. The additional canine osteosarcoma cell line MCKOS and the human packaging cells HEK293 17SF were kind gifts from Dr. Rimas J. Orentas (Miltenyi). The Human HEK293-K9 PD1 (expressing the canine PD1 receptor on the cell surface) cell line was a generous gift from Dr. Maninder Sandey. Canine normal fibroblast cells (NCF) were isolated and cultured from a donor dog’s subcutaneous fat tissue. The canine lymphoma cell line OSW was a generous gift from Dr. Bruce F. Smith. All cell lines and their respective clones were cultured according to standard laboratory procedures [25]. Adenovirus and lentivirus vector construction and packaging Genetically modified adenoviruses CAV2-AU-M3 (CAV2-E1AΔ21-RGD-sec.anti-PD1; Fig. 1A) were generated from the pICOCAV15 plasmid as a backbone using CRISPR/Cas9 technology, as published previously [26]. The anti-PD1 HcAb was developed by conjugating the anti-PD1 sdAb [25] to canine Fc and adding a secretory signal. The sequence of secretory anti-PD1 Fc-conjugated antibody includes FLAG and His-tags at the C-terminus of the sequence. We chose the sdAb clone with the highest binding affinity for canine PD1 (APD-1D5). Anti-PD1 HcAb sequence was then inserted under the control of the CMV promoter between the fiber and E4 sequences [26]. GFP/ffluc lentivirus was produced using the pALD-Lenti system (Aldevron), employing the pGFP/ffluc plasmid and packaging plasmids. Lentivirus packaging plasmid, pALD-Rev, pALD-gagpol, and pALD-VSV-G were purchased from Aldevron. pGFP/ffluc was a kind gift from Dr. Rimas J Orentas (Miltenyi). pGFP/ffluc, along with packaging plasmids, was transfected into HEK293/17SF cells using PEIpro (Polyplus) transfection system. Sodium butyrate (8mg/ml) was added 24 h post-transfection. On day 4, cells were harvested and centrifuged at 2000 rpm for 10 min at room temp. The supernatant was collected and filtered through a 0.45 μm filter, then centrifuged at 10,000 × g for 4 h with slow acceleration and a brake at 4 °C. The pellet, post-centrifugation, was suspended in 400 ul serum-free media and stored in small aliquots at -80°C. For lentivirus titration, canine OSW cells were cultured and transduced with GFP/ffluc lentivirus at different concentrations. The transduced cells expressing GFP were quantified using flow cytometry. Titer was calculated using the following formula: Generation of GFP + osteosarcoma cell lines Canine osteosarcoma cell lines D17, CF11, and D22 were transduced with GFP/ffluc lentivirus (MOI 1). Five days post-infection, GFP-positive cells were sorted (CytoFLEX, Beckmann) into 96-well plates at a density of one cell per well. Plates were monitored daily for growth and GFP expression. Selected clones from the 96-well plates were expanded gradually to T75 flasks. GFP expression was confirmed via fluorescent microscopy on a Keyence BZ-X800 microscope. Luciferase expression was confirmed via Steady-Glo luciferase assay (Promega cat #E2510) as per the manufacturer’s instructions. D17-GLP, CF11-GLL, and D22-GLP clones were selected to perform further experiments. CAV2-AU-M3 viral infections 2D (monolayer) cell culture: Canine osteosarcoma cell lines, D17-GLP, CF11-GLL, MCKOS, and D22-GLP, were cultured in a 12-well plate overnight. On the next day, media was removed, and cells were washed once with 1X PBS (Phosphate Buffer Saline, Corning: cat#21–040-CV). Cells were infected with CAV2-AU-M3 at multiple multiplicities of infection (MOI; 0, 10, 100, 500, 1000, 5000, and 10,000) in low-serum medium (500 ul DMEM, 2% FBS). After 2 h of incubation at 37 °C, 500 ul of additional DMEM (5% FBS) was added. Infected cells were maintained in DMEM (5% FBS), and the media was changed on days 3 and 5. The Keyence BZ-X800 imaging system was used to capture bright phase images on days 0, 3, and 5. 3D (spheroid) cell culture: Spheroids were formed by plating 5000 canine osteosarcoma or NCF cells on a low attachment round-bottom 96-well plate (Corning ref #7007) in 50 ul of complete media and 50 ul of extracellular matrix (ECM) (30 µg/mL; Collagen I, Thermo #5056-20ML). The plates were centrifuged at 500 g for 5 min and incubated at 37 °C for 72 h with regular monitoring of spheroid formation. Infections were performed at 72 h using CAV2-AU-M3 (MOI: 0, 100, and 1000). Partial media changes were preformed, and images were captured (Keyence BZ-X800 imaging system) on days 0, 4, 7, 10, and 13. The diameters of the spheroid images were calculated using ImageJ. The calculated spheroid area on days 7 and 10 was normalized to the spheroid size of day 0. In vivo model: D17-GLP 5X106 cells in 0.2 ml liquid Matrigel (Corning ref#354230) were subcutaneously injected into the left flanks of 6-week-old NOD.Cg‐Prkdcscid Il2rgtm1Wjl/SzJ female mice. Mice were randomly divided into two groups of five each. When the tumor size reached 100 mm^3, group 1 (n = 5) and group 2 (n = 5) were treated with 1 × 10^9 CAV2-AU-M3 virus particles and 1X PBS in a volume of 50 ul. Virus/1X PBS was injected intratumorally into the tumors. The status of the mice and tumor size were monitored daily during the 3 days after virus administration and then 3 times a week for 30 days. Tumor volume was defined by the equation V (mm3) = (W2 × L)/2, where W and L are the width and the length of the tumor, respectively. When animals from each group displayed uncontrolled tumor growth, the mice were euthanized. Spheroid histology Spheroids were infected with 0, 100, and 1000 MOIs of CAV2-AU-M3 and cultured for 13 days. On day 13, spheroids were collected, washed with 1 X PBS, and fixed in 4% PFA (Paraformaldehyde) at room temperature for 42 min. For better visualization of the spheroids, they were pre-stained with erythrosine B (0.2%) (Sigma cat#200964-5G) in 1X PBS and incubated for 3 min at room temperature. Spheroids were washed 3 times with 1X PBS and deposited in a cryomold containing liquid HistoGel (Thermo cat#HG10144). The gel was cured on an ice block for 10 min. The encased spheroids were subsequently transferred to mesh cassettes and placed in 70% ethanol until processing. Samples were processed using the routine cycle program in a Leica TP1040 processor and embedded in Leica Paraplast X-tra paraffin using a Sakura Tissue Tek IV embedding station. The blocks were sectioned at 4 µm using a Leica 2125RM microtome, with sections placed on Matsunami APS-charged slides (Cancer Diagnostics cat#SUAPS11). Slides were air-dried at room temperature overnight. Prior to staining, the slides were “baked” in an incubator at 60 °C for 10 min. Slides were manually stained using Cancer Diagnostics (CDI) Gill’s II hematoxylin (cat#CM5951) and eosin phloxine B (cat#EM00PG) working solutions. Slides were cover-slipped with #1 24–50 mm slips using Richard Allan CytoSeal XYL. Immunofluorescence (IF) 2D: Cells were plated at 50,000 cells per well in eight-chambered slides (Corning) and cultured overnight at 37 °C. Slides were infected with 0, 10, 100, and 500 MOI as per the 2D infection protocol above. Forty-eight hrs post-infection, slides were washed with PBS and fixed for 20 min in 4% PFA. Slides were then washed 1 × with PBS and incubated in block buffer (10% normal horse serum (Vector Laboratories cat #S-2000-20), 0.4% triton, and 1X PBS) for 1 h at room temperature. Primary antibody (Mouse anti-CAV2: VMRD; rabbit anti-VHH-iF647: GenScript cat #A01994-200) incubation was performed for 1 h at room temperature in 1X wash buffer (1% normal horse serum, 0.4% Triton X-100 in 1X PBS). Slides were washed 1–2 × with PBS. Slides were incubated with secondary antibody (Texas-red goat anti-mouse IgG: Invitrogen ref #T6390) for 1 h at room temperature. The slides were washed 1–2 × with PBS, the slide chambers were removed, and the slides were mounted with DAPI mounting medium (Vector Laboratories; cat# H-1800-Z). Slides were imaged on the Keyence BZ-X800 microscope. 3D: Slides were incubated at 60 °C for 10 min. Slides were deparaffinized and rehydrated with the following washes for 5 min each: xylene × 3, 100% ethanol × 2, 90% ethanol, 75% ethanol, 50% ethanol, and distilled water. After washing, slides were transferred to 10 mM sodium citrate and boiled in an Instant Pot for 10 min on the sterilization cycle. After completion, the slides were incubated in 10 mM sodium citrate at room temperature for 30 min. Slides were then washed in distilled water for 5 min. A PAP pen (Vector Laboratories) was used to draw a wide circle around the periphery of the tissue sample without allowing the sample to dry. Slides were incubated for 15 min at room temperature in wash buffer (1% normal horse serum, 0.4% triton × 100, in 1X PBS) and blocked using block buffer (10% normal horse serum, 0.4% triton 100X in 1X PBS) for 1 h at room temperature in a humid chamber and at 4 °C until the end of the day. At the end of the day, 1:500 primary antibody was added and incubated overnight at 4 °C in a humid chamber. The following day, slides were washed 3 × with wash buffer for 10 min each wash. Secondary incubation was done for 1 h in a humid chamber at room temperature. Slides were washed 1 × with 1X PBS for 10 min and mounted with DAPI mounting medium. Slides were imaged on the Keyence BZ-X800 microscope. PBMC isolation and T-cell activation Peripheral blood was collected from healthy donor dogs in 10 mL Monoject EDTA (K3) 0.10 mL 15.0% solution tubes (Covidien cat#8881311743). Blood was diluted with 25 mL wash buffer (PBS + 2% FBS + 0.5% EDTA) per 10 mL of blood. Thirty-five mL of diluted blood was then layered onto 15 mL of lymphocyte separation medium (Corning cat#25-072-CV) in a SepMate tube (StemCell cat#15450) and centrifuged at 1200 × g for 10 min as per the manufacturer’s instructions. The top layer was collected in a 50 mL conical tube, and wash buffer was added to a final volume of 50 mL and then centrifuged at 200 × g for 10 min at 50% acceleration and 0% deceleration to remove platelets. The supernatant was removed, and the pellet was washed twice with 10 mL of wash buffer and centrifuged at 500 × g for 10 min. To remove remaining red blood cells (RBCs), the cell pellet was resuspended in 3 mL 10% RBC lysis buffer (Invitrogen cat#00-4300-54) in sterile water and incubated at 4 °C for 4 min. Twelve mL of wash buffer was then added to stop lysis, and the cells were centrifuged at 500 × g for 10 min. If visible RBCs remained, this step was repeated. Cell pellet was then resuspended in 10 mL of wash buffer and centrifuged at 500 × g for 10 min. For T-cell purification, the cell pellet was resuspended in 80 μL MACS buffer (PBS + 0.5% FBS + 0.4% EDTA; kept cold during processing) per 10 million cells. Then, 10 μL of dog gamma globulins (Jackson cat#004-000-002) were added per 10 million cells, and the mixture was incubated on a rotator for 15 min at room temperature. Next, primary antibodies to CD11b (Biorad cat#MC1777S) 5 uL per 10 million cells, CD11c (Biorad cat#MC1778S) 5 uL per 10 million cells, CD21 (Biorad cat#MCA1781R) 1 uL per 10 million cells, and CD14 (Biorad cat#MCA1568GA) 1 uL per 10 million cells were added and incubated on a rotator for 30 min at 4 °C. Cells were then washed with 2 mL MACS per 10 million cells and centrifuged at 500 × g for 10 min at 4°C. The cell pellet was then resuspended in 80 μL MACS per 10 million cells. Next, 20 μL anti-mouse IgG (Miltenyi cat#130-048-401) beads per 10 million cells were added and incubated on a rotator for 15 min at 4 C. Cells were then washed twice with 2 mL MACS per 10 million cells and centrifuged at 500 g for 10 min at 4 °C. An LS column (Miltenyi cat# 130-042-401) was attached to the magnet (Miltenyi; midi MACS multistand) and rinsed with 3 mL of MACS. The cell pellet was resuspended with 50 μL MACS per 10 million cells and ran through the LS column. The column was rinsed with 500 μL MACS. Purified T cells were diluted to 10 mL in MACS and counted. A total of 100,000 cells were set aside for flow cytometry. Cells were then centrifuged at 500 × g for 10 min at 4 °C. To confirm T-cell purity, flow cytometry using anti-CD5 antibody (Invitrogen cat#25-0250-42) was performed. For activation, purified T cells were cultured in a 6-well plate coated with anti-dog CD3 (Biorad cat#MCA1774GA) and anti-dog CD28 clone 5B8 (generously gifted by Dr. Maninder Sandey). Activation was confirmed after 72 h by flow cytometry with anti-dog CD25 (Invitrogen cat#25-0250-42) primary antibody. Activated T cells were then maintained at 1 million T cells per mL in IL-2-supplemented T-cell media (TCM; IMDM (Gibco) 10% FBS (Sigma cat#16,000–044), penicillin (100 IU/ml, Corning cat#30-002-CI), streptomycin (100 μg/ml, Corning cat#30–002-CI), 1% non-essential amino acids (Corning cat#25-025-CI), and 1% sodium pyruvate (Gibco cat#113600-070) for 30 min for 10 days, with media supplemented every 2–3 days and cell counts every 3–4 days. Luciferase cytotoxicity assay Cell cytotoxicity post-virus infection in 2D culture was quantified using Luciferase cytotoxicity assay (Promega). Canine osteosarcoma cell lines D17-GLP, CF11-GLL, MCKOS, and D22-GLP were cultured in a 96-well plate (10,000 cells/well). CAV2-AU-M3 infections were performed as described above. On days 0, 3, and 5, Steady-Glo® Luciferase Assay reagent (Promega cat #E2510) was added in equal amounts of media as per the manufacturer’s directions. Cells and luciferase reagent were incubated at room temperature for up to 30 min. Cell lysis was monitored by luminescence using a microplate reader (BioTek Synergy H1). The media was changed before adding the assay reagent on each day. TCID50/ml All four OS cell lines and NCF were plated in a 96-well plate (10,000 cells/well). The next day, each well was infected with tenfold dilutions of virus with a starting inoculum of 2.5 ul in the top row. After 5 days of infection, crystal violet staining was performed as described previously [27]. TCID50 was calculated using the Reed and Munich assay. CAV2-AU-M3 Infections and nickel column purifications CAV2-AU-M3 infections and anti-PD1 Ab purification were done as per our standard laboratory techniques [25]. Briefly, D17-GLP, CF11-GLL, MCKOS, and D22-GLP cell lines were cultured in T75 culture flasks up to 70–80% of confluence. The cells were infected with CAV2-AU-M3 at 100 MOI in DMEM with 5% FBS. Seventy-two hrs post-infection, media and cell lysate were collected. For cell lysate, cells were collected and washed with 1X PBS, then lysed with RIPA buffer (Thermo ref #89900) supplemented with 1% Halt Protease Inhibitor Cocktail (Thermo). Cells were frozen at − 80 °C and centrifuged at 14,000 X g. Supernatant was collected and stored at -20 °C. Cell culture media was concentrated using protein concentrators (Thermo ref #88528; 10K MWCO). His-tagged anti-PD1 Ab was purified using Ni–NTA spin Kit (Qiagen) as per the manufacturer’s protocol. The purified samples were stored at − 20 °C. Western blot Anti-PD1 Ab identification Purified anti-PD1 Ab samples were boiled for 10 min in Lane marker reducing buffer ((Thermo: cat#39000)) and loaded onto NuPAGE 4–12% Bis–Tris gel (Thermo ref #NP0323BOX). The chameleon Duo Pre-stained ladder (Licor ref #928–60000) was used for molecular weight control. SDS-PAGE Electrophoresis was performed using 1X NuPAGE MOPS SDS running buffer (Thermo #ref B0001) for ~ 45 min at 210 Volts. Protein transfer on PVDF membrane post-electrophoresis was performed using 1X Bolt transfer buffer (Thermo ref #BT00061) at 20 V for 1 h. The membrane was blocked for 2 h at room temperature using block buffer (1 × PBS, 0.1% Tween20, 5% non-fat milk powder). Incubation with primary antibody was performed overnight at 4 °C. The membrane was washed the next day 3 times using wash buffer (1 × PBS and 0.1% Tween20) for 20 min each and incubated with secondary antibody for 1 h at room temperature. The membrane was washed 3 times using the wash buffer after secondary antibody incubation. The membrane was scanned using LICOR Odyssey system. Mouse anti-His-Tag (Thermo ref #MA1-21315) and rabbit anti-VHH-HRP (GenScript cat #A01861-200) primary antibody and IRDye 800 CW goat anti-mouse IgG (LicorBio ref #926–32,210) were used. Purified anti-PD-1-D5 sdAb (~ 15 kDa) was used as a positive control (a generous gift from Dr. Maninder Sandey). Flow cytometry PD1 Binding Human HEK293-K9PD1 cell line cells and T cells were blocked for 1 h at room temperature using blocking buffer (FACS + 10% normal mouse serum; ImmunoReagents cat #SP-002-VX5). Cells were labeled with primary antibody (purified anti-PD1 Ab from each cell line and NCF cells) for 1 h at room temperature post-blocking. Unlabeled cells were used as negative control. Cells were washed with FACS buffer (1X PBS + 1% B.S.A.; bovine serum albumin; VWR cat #0332, + 200 mM EDTA, Thermo ref #15,575–038) and labeled with mouse anti-His-Tag APC-conjugated secondary antibody (RandD: cat #IC050A) for 1 h at room temperature. Cells were washed and resuspended in FACS buffer before analyzing them with flow cytometry (CytoFLEX LX, Beckman Coulter; LSR-II; B.D. Biosciences). Data analysis was done using the FlowJo software version 10.7.1 (BD Biosciences). PD1/PDL1 Inhibition Human HEK293-K9PD1 cell line cells were blocked for 1 h at room temperature using blocking buffer (FACS + 10% normal mouse serum; ImmunoReagents cat#SP-002-VX5). Cells were labeled with primary antibody (purified anti-PD1 Ab from each cell line and NCF cells) for 1 h at room temperature post-blocking. Unlabeled cells were used as a negative control. Cells were washed with FACS buffer (1X PBS + 1% B.S.A.; bovine serum albumin; VWR cat #0332, + 200 mM EDTA, Thermo ref #15,575–038) and labeled with Zenon AF647 (Thermo ref #Z25408) labeled canine PDL1 protein conjugated with human Fc (Thermo: cat #P21462) for 1 h at room temperature. Cells were washed and resuspended in FACS buffer before analyzing via flow cytometry (CytoFLEX LX, Beckman Coulter; LSR-II; B.D. Biosciences). Data analysis was done using the FlowJo software version 10.7.1 (BD Biosciences). Statistical analysis Statistical analysis was performed using Python and GraphPad. Linear regression was used to evaluate the relationship between viral dose (MOI) and cell viability in the 2D luciferase assay, with log-transformed MOI as the independent variable. Two-way ANOVA was applied to spheroid size data to assess the effects of time, MOI, and their interaction on growth for each cell line separately. For tumor volume data, linear regression was used to examine changes over time and to evaluate the interaction between time and treatment (virus-treated vs. PBS). Results Cytotoxic and cytopathic effects of CAV2-AU-M3 2D Four canine osteosarcoma cell lines, D17-GLP, CF11-GLL, D22-GLP, and MCKOS that express GFP/ffluc, were used to test the inter-line tumor heterogeneity. Cytotoxicity was observed using the luciferase cytotoxicity assay (cell viability assay). There were no significant differences in cell viability at 0, 10, and 100 MOI in D22-GLP and MCKOS cell lines on day 5, but some cytotoxicity was evident (rapid cell rounding and progression to complete cell death) in D17-GLP and CF11-GLL (Figs. 1B; S1–12). However, we observed that at MOIs of 1000 or higher, CAV2-AU-M3 lysed all cell lines efficiently by day 5, but at varying percentages (Fig. 1D). D17-GLL and CF11-GLL cell lines showed clear cell lysis at MOIs greater than 500 (Figures AD; S1-12). D17 GLP and CF11 GLL cell lines have higher as well as TCID50 (Fig. 1G) values, also in comparison with MCKOS and D22 GLP cell lines. CAV2-AU-M3 showed CPE in MCKOS and D22-GLP cell lines at MOIs above 5000, compared to the negative control (Fig. 1D). Linear regression analysis using log-transformed MOI demonstrated a significant negative association between viral dose and cell viability across all cell lines (p < 0.001). Increasing MOI was associated with a reduction in viability in D17-GLP (β = − 21.40, p < 0.001), CF11-GLL (β = − 22.22, p < 0.001), D22-GLP (β = − 8.48, p < 0.001), and MCKOS (β = − 14.40, p < 0.001). CF11-GLL and D17-GLP cell lines exhibited the steepest declines in viability, indicating higher sensitivity to increasing viral dose. In contrast, D22-GLP showed a more gradual reduction, suggesting relative resistance, while MCKOS demonstrated an intermediate response. Evidently, CAV2-AU-M3 produces more CAV2 virus particles in D17-GLP and CF11-GLL in comparison with MCKOS and NCF cells (Fig. 1C). D17-GLP and CF11-GLL cell lines had the highest TCID50/ml (1.26 × 105) in comparison with MCKOS (1.26 × 104) and D22-GLP (1.26 X 103) (Fig. 1E, F, and G). 3D Spheroid growth and cell lysis D17-GLP, CF11-GLL, and MCKOS spheroids were infected at 0, 100 (S13–23), and 1000 MOIs of CAV2-AU-M3 for 13 days. All four cell lines responded differently to CAV2-AU-M3 infections in 3D. A two-way ANOVA incorporating time, MOI, and their interaction revealed heterogeneous growth responses across the osteosarcoma cell lines. There were no or minimal changes in spheroid volume between 0 and 1000 MOIs in D22-GLP and MCKOS spheroids (Fig. 2A, B, and D), even at day 13. The MCKOS cell line showed a significant increase in size over time (β = 14.73, p = 0.017), with no significant interaction between time and MOI (p = 0.695 for 1000 MOI). D22-GLP cell line spheroids exhibited no significant change in size over time (β = − 0.89, p = 0.706), and no significant effects of MOI or interaction were detected (p > 0.05), suggesting minimal responsiveness to treatment. CAV2-AU-M3’s infectious and cytolytic properties in canine osteosarcoma cell lines, spheroids, and a xenograft murine model. A OS spheroids were infected with 1000 MOIs of CAV2-AU-M3, and spheroid size was visualized using inverted fluorescent microscopy (Keyence) after 13 days of infection. B Spheroid size on days 0, 7, and 13 at 0 and 1000 MOIs. C Histological sections of D17GLP, CF11GLL, and MCKOS spheroids infected with 0, 100, and 1000 MOIs after 13 days of infection. D Spheroid volumes as a percentage of the same OS cell line at day 0 shown at days 0, 7, and 13, at 0 and 1000 MOIs. E Xenograft tumor volumes (mm3) in PBS (β = 30.73, p < 0.001) and CAV2-AU-M3-treated (β = − 15.98, p < 0.001) groups. Tumors were monitored up to 30 days In the D17-GLP cell line, the infected spheroids were bigger in size in comparison with day 0, but spheroids infected with CAV2-AU-M3 had less growth in comparison with non-infected spheroids. There is a significant difference in spheroid volume between 0 and 1000 MOIs of infections (Fig. 2A, B, and D) on day 7 (p = 0.0005) and almost significant on day 13 (p = 0.0711). CF11-GLL spheroids were significantly smaller in comparison with non-infected spheroids both on days 7 and 13 at 1000 MOIs (Fig. 2A, B, and D). CF11-GLL cell line cells demonstrated a strong time-dependent increase in spheroid size (β = 101.91, p < 0.001) and a significant negative interaction between time and 1000 MOI (β = − 66.03, p = 0.001), indicating that high viral dose significantly suppressed spheroid growth over time compared to control. These findings indicate that, while spheroid growth increased over time in most cell lines, only D17-GLP and CF11-GLL showed a statistically significant reduction in growth rate with increasing MOI, highlighting cell line-specific sensitivity to viral treatment. Spheroid histology Infected (100 and 1000 MOIs) and non-infected spheroids were collected on day 13. Spheroids were processed following the protocol described in the methods section, sectioned (4 um), and stained with hematoxylin and eosin phloxine B stain. At 0 MOI, no tumor cell death was observed in D17-GLP, CF11GLL, or MCKOS cells. D17-GLP and CF11-GLL spheroids infected with CAV2-AU-M3 at 100 or 1000 MOI showed extensive tumor cell lysis. Tumor cell lysis is evident by cell shrinkage, loss of plasma membrane, cytoplasmic eosinophilia, and karyolitic or pyknotic nuclei. However, MCKOS spheroids were relatively resistant to CAV2-AU-M3-induced tumor lysis (Fig. 2C). Animal model We tested the cytolytic effects of CAV2-AU-M3 in a murine xenograft osteosarcoma model in NOD.Cg‐Prkdcscid Il2rgtm1Wjl/SzJ mice. Five million D17-GLP cells were injected subcutaneously into the left flank of 10 mice. A single intratumoral injection of 1 X 109 VPs was performed when the tumor size reached 100 mm3 in five mice. Five control mice were injected with 1X PBS. Tumor size was measured every alternate day for 30 days in both the control and treatment groups. Both groups had constant tumor growth in all mice. However, mice treated with CAV2-AU-M3 had slower growth in comparison with the control group, as seen in the spheroid model (Fig. 2E). Linear regression analysis demonstrated a significant increase in tumor volume over time in the control (PBS) group (β = 30.73, p < 0.001). Importantly, a significant interaction between time and treatment was detected (β = − 15.98, p < 0.001), indicating that tumor growth over time was significantly reduced in the virus-treated group compared to PBS. Specifically, the estimated growth rate in the virus-treated group (~ 14.75 units/day) was substantially lower than in the control group (~ 30.73 units/day), demonstrating that viral treatment effectively suppresses tumor progression. Characterization of His-Tagged Anti-PD1 Ab produced and secreted by OS cells infected by CAV2-AU-M3 Anti-PD1 Ab size, production, and secretion verification Western blot was performed to verify the correct size, production, and secretion of His-tagged anti-PD1 Ab from CAV2-AU-M3 infected OS cells (Fig. 3A and B). The secreted heavy-chain antibody (HcAb) was labeled with anti-His-tag (Fig. 3A) and an antibody against camelid VHH (Fig. 3B) for verification. All four OSA cell lines (D17-GLP, CF11-GLL, MCKOS, and D22-GLP) produced and secreted anti-PD1 Ab in cell culture media. Both monomers (~ 42KD) and dimers (~ 84KD) were observed (Figs. 3A and B; S24). We observed that the amount of anti-PD1 Ab was higher in the media than cell lysates across all four OS cell lines (Figure S24). There was no visible monomer anti-PD1 Ab secreted by primary NCF (normal canine fibroblasts) cells; however, NCF produced very low levels of anti-PD1 Ab dimers in comparison with OSA cell lines (Figs. 3A and B; S24). Functional properties of anti-PD1 HcAb isolated from cell culture media of D17GLP, CF11GLL, D22GLP, and MCKOS cell lines and NCF cells 72 h after CAV2-AU-M3 infection at 100 MOI. A Western blot analysis. Anti-PD-1 HcAb protein levels (monomers) were measured using mouse anti-6x-His (Thermo) primary antibody and IRDye 800 CW goat anti-mouse IgG (LI-COR) secondary antibody, and B rabbit-anti-VHH-HRP primary antibody (GenScript) LI-COR. C Immunofluorescence (Keyence; 10X) demonstrating the presence of anti-PD1 HcAb in monolayer cultures was performed in D17 GLP, CF11 GLL, and MCKOS cell lines, and primary NCF cells, at 0, 10, 100, and 500 MOIs (CAV2-AU-M3 infections for 48 h) using rabbit anti-VHH-iFluor 647 (GenScript) primary antibody. D Immunofluorescence (Keyence; 10X) demonstrating the presence of anti-PD1 HcAb using the rabbit anti-VHH-iFluor 647 (GenScript) primary antibody was performed in D17 GLP, CF11 GLL, and MCKOS spheroids at 0 and 1000 MOIs after 13 days of infection. Images were captured using a Keyence 48 h post-infection. E Flow cytometry analysis of anti-PD-1 HcAb binding to PD-1 receptor on HEK293-K9 PD1 cells, with cell count on the y-axis and fluorescence on the x-axis. The binding of anti-PD-1 HcAb purified from cell media collected from osteosarcoma cell lines (D17 GLP, CF11 GLL, MC-KOS, and D22 GLP) infected with CAV2-AU-M3 was evaluated. F Flow cytometry analysis of anti-PD-1 HcAb binding to PD-1 receptor on T cells, with cell count on the y-axis and fluorescence on the x-axis. The binding of anti-PD-1 HcAb purified from cell media collected from osteosarcoma cell lines (D17 GLP, CF11 GLL, MC-KOS, and D22 GLP) infected with CAV2-AU-M3 was evaluated. G Flow cytometry analysis of PD-1/PD-L1 binding inhibition by anti-PD-1 HcAb. Inhibition of the binding of PD-L1 to PD-1 by anti-PD-1 HcAb purified from cell media collected from osteosarcoma cell lines (D17 GLP, CF11 GLL, MC-KOS, and D22 GLP) infected with CAV2-AU-M3. H Percentage population of HEK293 K9PD1 cells labeled by anti-PD1 HcAb isolated from OS cell lines. This table is the average of three independent experiments. I Percentage population of canine T cells labeled by anti-PD1 HcAb isolated from OS cell lines. This table is the representation of three independent experiments. J Percentage population of HEK293 K9PD1 cells labeled by xenon AF647 labeled PDL1 protein due to anti-PD1 HcAb inhibition. This table is the average of three independent experiments Immunofluorescence analysis of anti-PD1 secretion from OS cells OSA cell lines (D17-GLP, CF11-GLL, and MCKOS) and NCF were infected with CAV2-AU-M3 for 48 h in monolayers before being stained with anti-camelid VHH antibodies. All three OS cell lines showed anti-PD1 Ab production, as evidenced by the red fluorescence (Fig. 3C). There was no or minimal production of anti-PD1 Ab in NCF cells (Fig. 3C). Similarly, all three OS cell lines showed anti-PD1 Ab secretion in spheroids after 13 days of CAV2-AU-M3 infection (Fig. 3D). Confirmation of secreted anti-PD1 Ab binding with its ligand PD1 + HEK293-K9PD1 cells: Flow cytometry was performed to analyze the binding capacity of anti-PD1 Ab secreted by CAV2-AU-M3-infected OS cells to its PD1 ligand. Anti-PD1 antibody released from all four infected cell lines demonstrated similar binding to PD1 cell surface receptor. A shift in the fluorescence in the histogram to the right shows anti-PD1 Ab binding canine PD1 surface receptor in HEK293-K9PD1 cells (Fig. 3E and H). D17-GLP, CF11-GLL, D22-GLP, and MCKOS media samples have shown binding to 91%, 92.61%, 89.97%, and 92.01% respective cell population of HEK293-K9PD1 cells expressing canine PD1 ligand on the cell surface (Fig. 3E and H). T cells: T cells isolated and purified from peripheral blood from a donor dog were activated using costimulatory antibodies (anti-CD3 and anti-CD28) and cultured for 10 days before analysis. Anti-PD1 HcAb samples from all four infected cell lines (D17-GLP, CF11-GLL, D22-GLP, and MCKOS media) showed binding to canine T cells (Fig. 3F and I). Assessment of the Inhibition of PDL1 Binding to PD1 HEK293-K9PD1 cells: The functionality of anti-PD1 Ab secreted by CAV2-AU-M3 infected OS cells was assayed for its ability to block the PDL1 protein binding to PD1 receptor. HEK293 cells expressing canine PD1 receptors were labeled with AF647 Zenon labeled canine PDL1 protein with and without the presence of anti-PD1 Ab (CAV2-AU-M3) (Fig. 3G). A shift in the fluorescence in the histogram to the right after PDL1 labeling reverts back to the left, below the threshold when blocked with anti-PD1 antibody, demonstrating inhibition of PDL1 binding with PD1 receptor (Fig. 3G). 55.07 percent of the population of HEK293-K9PD1 cells were labeled by Zenon-tagged canine PDL1 protein. The percentage population of HEK293-K9PD1 cells decreased in the presence of D17-GLP, CF11-GLL, D22-GLP, and MCKOS media samples containing secreted anti-PD1 Ab from 55.07 to 2.31%, 2.19, 2.39, and 2.38%, respectively (Fig. 3J). Discussion Osteosarcoma is a highly metastatic tumor with no major changes in survival outcomes in patients over the past four decades [28, 29]. There is a pressing need to explore new therapeutic approaches to treat osteosarcoma. Cancer immunotherapies, such as oncolytic viruses and immune checkpoint inhibitors, are a promising avenue for more patient-tailored and effective therapeutic responses [30]. Conditionally replicative oncolytic virus (CRADs) immunotherapy is a work-in-progress that has shown promising results in some tumors, including osteosarcoma [31, 32]. T-VEC is an FDA-approved oncolytic virus currently used to treat melanoma [33]. CRADs infect and specifically lyse tumor cells [34] and induce systemic anti-viral immunity along with anti-tumor immunity [35]. Additionally, they can activate innate immunity, thereby triggering adaptive immunity in the TME [18]. We have developed a conditionally replicative oncolytic virus, CAV2-AU-M3, which is armed with an immune checkpoint inhibitor, anti-PD1 HcAb (Fig. 1A). ICOCAV15 [36] was genetically modified to insert a secretory canine hinge and an Fc-conjugated camelid anti-canine PD1 single-domain antibody between the fiber and E4 to produce CAV2-AU-M3. The resulting virus is ~ 107% the size of the CAV2 Toronto strain. Our hypothesis was that an armed oncolytic virus will be able to lyse cancer cells as well as produce and secrete the anti-PD1 HcAb payload in the tumor microenvironment. We demonstrated the cytolytic activity of CAV2-AU-M3 in four different osteosarcoma cell lines and its ability to produce and secrete a functional anti-PD1 antibody. Virus-induced cell lysis results in visible cytopathic effects (CPE), such as cell rounding, detachment, and syncytia in a cell monolayer in cell culture [37]. We observed that at a low multiplicity of infection (MOI) of 100, there was a visible CPE in D17-GLP and CF11-GLL; however, no such effect was observed in D22-GLP and MCKOS cell lines (Fig. 1B). Immunofluorescence in OS cell lines and primary canine fibroblast cells (NCF) showed that the D17-GLP and CF11-GLL cell lines produced more CAV2-AU-M3 virus particles than MCKOS and NCF cells (Fig. 1C). All four cell lines, when infected at higher MOIs of 500–10,000, showed varying levels of cell lysis (as measured by luciferase assay luminescence). D17-GLP and CF11-GLL showed significant cell lysis at 500 MOI and nearly 90% killing at 10,000 MOI after 5 days. MCKOS cells began showing significant cell lysis at 5000 MOI, reaching nearly 70% at 10,000 MOI. D22-GLP cells also showed significant cell lysis at 5000 MOI; however, unlike MCKOS, at 10,000 MOI, they showed only 40% cell lysis. We have observed that CAV2-AU-M3-induced cell lysis is more robust in fast-growing cell lines such as D17-GLP and CF11-GLL. Our data also demonstrated that CAV2-AU-M3 packages and lyses at different rates in different cell lines. TCID50 values of CAV2-AU-M3 in the MCKOS cell line are tenfold more than D17 GLP and CF11-GLL (Fig. 1D, E, and F). Similarly, the TCID50 value in D22-GLP is tenfold lower than in MCKOS cell lines. Hence, CAV2-AU-M3 can infect and lyse tumor cells in monolayer cell cultures, however, at different ratios in different cell lines. The differences in cell lysis across cell lines could result from inter-individual heterogeneity. Further, we cultured spheroids from all four OS cell lines to demonstrate the effect of CAV2-AU-M3 on spheroid growth. Tumor spheroids provide the 3D architectural features of solid tumors that are absent in 2D monolayers, making them a better model for studying viral penetration, replication, and therapeutic efficiency [38]. We observed that the D17-GLP and CF11-GLL cell lines’ spheroids infected with 1000 MOI of CAV2-AU-M3 had a reduced growth over the course of time (13 days) compared to uninfected/0 MOI spheroids (Fig. 2A). There was a significant size difference between infected and uninfected spheroids at days 7 and 13 in D17-GLP cell lines. The size difference is significant only on day 13 in CF11-GLP cell lines. However, no significant reduction in growth was observed in the D22-GLP and MCKOS cell lines (Fig. 2B and D). As in 2D cultures, the virus effect was different across cell lines in 3D cultures as well. Lastly, we demonstrated the effect of CAV2-AU-M3 in D17-GLP xenograft tumors in immunocompromised mice (Fig. 2E). CAV2-AU-M3 was injected once intratumorally after tumors reached the size of 100 mm3. Tumors were measured every other day. A single injection of CAV2-AU-M3 did not clear the tumors completely but reduced the tumor growth. We monitored tumor growth for up to 30 days. A significant difference in tumor size between the control and virus groups was observed starting on day 22. Based on the spheroid and xenograft tumor data, to attain a profound effect, we likely need to administer multiple doses of the virus to achieve complete tumor remission. It should be noted that the immunoregulatory aspect of the virus has little to no effect in the immunosuppressed murine strain required for xenograft survival, and thus, efficacy will likely be enhanced in an immunocompetent model. The second aspect of our virus is its ability to produce and secrete anti-PD1 HcAb. CRADs can be engineered to express the immunomodulatory proteins that can shape the tumor microenvironment (TME) against the tumor [39–41]. Immune checkpoint inhibitors (ICIs), such as antibodies against PD1 and PDL1 [42–46], can release the immune inhibitory “brakes” on T cells to restore anti-tumor immunity [47, 48]. However, systematic delivery of ICIs can lead to off-target side effects and autoimmunity [49]. Therefore, we designed next-generation CAV2-AU-M3, which will produce sustained levels of anti-PD1 Ab directly in the TME upon infection. CAV2-AU-M3 is an enhanced version of our previously published CAV2-AU-M2 oncolytic virus [50]. CAV2-AU-M3 produces secretory chimeric anti-PD1 Ab (Fig. 1A). The anti-PD1 Ab comprises anti-canine camelid VHH and hinge and Fc domain of canine IgG D. The chimeric version of anti-PD1 Ab is bigger in size than nanobodies and has a higher half-life in TME [51]. CAV2-AU-M3 is designed to produce and secrete anti-PD1 Ab in the TME upon tumor cell infection, along with viral replication, which is important for enhanced T-cell function in the TME [52]. We infected all four cell lines with CAV2-AU-M3 and analyzed the production and secretion of anti-PD1 Ab in cell culture media. The anti-PD1 Ab produced makes monomers of ~ 42KD (Fig. 3AandB) and dimers of ~ 84KD (Supplementary D3). We observed that MCKOS and D22-GLP cell lines, although not sensitive to CAV2-AU-M3 cell lysis, did produce more anti-PD1 Ab at a lower MOI of 500 in comparison with D17-GLP and CF11-GLL (Fig. 3A and B). D22-GLP showed similar cell lysis to NCF (TCID50; Fig. 1E and F). However, NCF cells produce a lower level of anti-PD1 antibody than OS cell lines. There was no visible monomer band in NCF cells’ media (Fig. 3A and B) and a minimal dimer band at the same MOI in comparison with all OS cell lines. These data demonstrate that CAV2-AU-M3 may be impaired in lysing some cell lines (D22-GLP and MCKOS compared to D17-GLP and CF11-GLL) but can still produce anti-PD1 Ab at similar levels. To further test this hypothesis, we performed immunofluorescence against anti-PD1 VHH in both 2D and 3D cultures. Compared with Fig. 1B, it is reasonable to assume that although CAV2-AU-M3 does not produce and package virus particles equally across all OS cell lines, its genome is replicating and thus producing a comparable amount of anti-PD1 in low packaging cell lines (MCKOS and D22-GLP; Fig. 3B and C). ICIs, such as anti-PD1 antibodies, competitively inhibit the PD1-PDL1 interaction, thereby reversing T-cell exhaustion in the TME and facilitating T-cell killing of tumor cells [53, 54]. It is not possible to demonstrate the effect of the enhanced CAV2-AU-M3 virus on the host’s immune response in vitro. However, we were able to demonstrate that anti-PD1 Ab produced by OS cell lines can bind to its intended receptor, PD1, and can also inhibit PD1 and PDL1 interaction (Fig. 3E–H). As is evident in Fig. 3F and I, anti-PD1 secreted by all OS cell lines binds the PD1 receptor expressed on HEK293 K9PD1 cells and on T cells isolated and activated from a healthy dog donor. In addition to efficient PD1 binding, anti-PD1 Ab produced by CAV2-AU-M3 can effectively inhibit the PD1-PDL1 interaction (Figure G and J). Thus, CAV2-AU-M3 will be able to reverse T-cell exhaustion in the TME upon infection and facilitate the stimulation of innate and adaptive immune responses against osteosarcoma. Based on the data presented, we can deduce that CAV2-AU-M3 can infect and produce anti-PD1 Ab in multiple OS cell lines. It is apparent that CAV2-AU-M3 has a slight packaging disadvantage in slow-growing cell lines MCKOS and D22-GLP, but it can still efficiently produce anti-PD1 in cell culture media. This may be due to the viral genome replicating in the presence of free E2F in immortal cell lines, so more copies are available to produce more anti-PD1 Ab. Production of anti-PD1 Ab is under the CMV promoter in CAV2-AU-M3, and the more gene copies, the more protein is produced. The proliferative cells have a high capacity for protein synthesis, which viruses can exploit to produce capsid proteins [55]. Therefore, CAV2-AU-M3 packaging may be lower in D22-GLP and MCKOS due to their low proliferation rates. Based on this evidence, it is suggestive that one treatment does not fit all tumors, but it will likely be useful in lysing aggressive tumors. However, the virus still has the potential to release anti-PD1 Ab into the TME, even if it is not efficiently lytic. We acknowledge that CAV2-AU-M3 lyse cells at higher MOIs (> 500 MOI). This effect could also be the result of a higher genome size (~ 107%) than its wild type. The design of this virus will serve as a framework for future generations with immunostimulatory payloads and a smaller genome. We would also like to add that higher MOI viruses will, in general, have higher production costs and a higher antigen load in the host. However, since virus production is less expensive than protein production, it will still have a better reach for ICIs in TME. And by limiting virus administration to the tumor site (intratumorally), the virus’s systemic effects can also be limited. In conclusion, CAV2-AU-M3 will be able to infect, lyse, produce, and secrete anti-PD1 Ab in the host TME upon infection. However, in some slow-growing tumors, CAV2-AU-M3 may be inefficient in lysing tumor cells such as MCKOS and D22-GLP; however, it can still initiate and enhance an immune response in the TME against the tumor by producing a functional anti-PD1 antibody. While we were able to produce an oncolytic virus with an immunostimulatory payload, further improvements are still needed for a better lytic oncolytic virus. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (PDF 8976 KB) Acknowledgements The authors would like to acknowledge Dr. Rie Watanabe, Auburn University Flow Cytometry and High-Speed Cell Sorting Laboratory Core Facility (RRID: SCR_025507), and Dr. James Gillespie. Author contributions SK drafted the manuscript. TAH drafted a part of the manuscript and revised the whole document. PA conceived and designed the study. YF and MS performed the statistical analysis. SK, TAH, ISL, EL, and GB conducted experiments and curated data. MS and BFS contributed to data analysis. LH performed tissue histology. PA critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript. Funding This work was supported by the National Institutes of Health [R15 CA283777-01]. Data availability All data generated and analyzed during this study are included in the article and its Supplementary Materials. Further supporting data is available from the corresponding author upon reasonable request. Declarations Conflict of interest The authors declare no conflict of interests. Ethical approval The animal experiment was approved by the Institutional Animal Care and Use Committee (IACUC #2023–5251) of Auburn University and was conducted in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Sumbul Khan and Theresa A. Higgins have equally contributed to the manuscript. References Associated Data Supplementary Materials Supplementary file1 (PDF 8976 KB) Data Availability Statement All data generated and analyzed during this study are included in the article and its Supplementary Materials. Further supporting data is available from the corresponding author upon reasonable request.