Photoactivatable oncolytic adenovirus for optogenetic cancer therapy

Oncolytic adenoviruses are widely used anticancer agents engineered to replicate in and destroy tumor cells, but their tumor selectivity remains suboptimal, raising safety concerns. Introducing optogenetic control could enhance selectivity by restricting viral replication to illuminated tumor regions. The study aimed to create a photoactivatable oncolytic adenovirus (paOAd) in which blue light controls expression of adenoviral E1 genes necessary for replication, thereby improving safety and tumor specificity. Prior work with tumor-specific replication-competent adenoviruses (TRADs) used the hTERT promoter to drive E1A/E1B expression selectively in hTERT-positive cancer cells; however, some normal tissues (e.g., somatic stem cells, testis, bone marrow, duodenum) also express hTERT, risking off-target replication. The objective here was to integrate an optogenetic GAVPO system to achieve light-dependent activation of E1 genes and minimize off-tumor effects.

The paper situates its work within oncolytic adenovirus therapy, noting adenoviruses are the most commonly used oncolytic platform. Previous TRADs leveraged the hTERT promoter to confer cancer selectivity by driving E1A/E1B expression, achieving preferential replication in hTERT-positive cells. Nonetheless, hTERT expression in certain normal tissues undermines absolute specificity, motivating additional control layers. Optogenetic systems such as GAVPO (containing VVD, Gal4(1–65), and p65 activation domain) can be used to regulate transcription with blue light, offering spatiotemporal control of viral replication.

Vector design and construction: An optogenetic transcription system (GAVPO: VVD LOV domain, Gal4 residues 1–65 DNA-binding domain, and p65 activation domain) under the hTERT promoter was assembled with a poly(A) signal and upstream activating sequence for Gal4 (UASG). Adenoviral E1A and E1B19k genes, separated by a 2A peptide, were placed downstream of UASG. The cassette (hTERT-GAVPO-pA-UASG-E1A-2A-E1B19k-pA) was synthesized and cloned into the E1-deleted Ad5 backbone (pAdHM3) via I-CeuI/PsiI, generating pAdHM3-hTERT-GAVPO-pA-UASG-E1A-2A-E1B19k-pA. PacI-digested plasmid was transfected into HEK293 cells with Lipofectamine 2000 to rescue the virus, termed paOAd. Virus was amplified, purified by cesium chloride ultracentrifugation, dialyzed, and stored at −80°C. Infectious units (IFU) were quantified (Adeno-X Rapid Titer Kit); particle-to-infectivity ratio was approximately 12–15:1. A control Ad-hTERT-LacZ was similarly constructed. In vitro illumination: Cells were irradiated with blue light using an Ultra High Power LED (Prizmatix); intensity measured with a Photodiode Power Sensor (PD300). Illumination was from the bottom of culture dishes. Dark conditions were established by wrapping plates in aluminum foil. Typical assays used 0.1 mW/cm² blue light for 5 days after infection. In vivo blue LED device: Implantable devices were fabricated on a polyimide flexible substrate (110–150 μm thick) with copper wiring. Two InGaN/Sapphire blue LEDs (peak 470 nm) were epoxy-mounted at the tip, wire-bonded, and series connected. Ends were epoxy-molded for waterproofing and coated with a 5–10 μm parylene layer for biocompatibility. Devices were inserted between hepatic lobes and secured by fixing the power cord to the peritoneum. Subcutaneous xenograft model: H1299 or HepG2 cells (3×10^6 in 50 μL PBS mixed with 50 μL Matrigel) were injected subcutaneously into female NOD/SCID mice (8 weeks). paOAd was administered intratumorally at 1×10^7 IFU/mouse on day 0 and day 3; tumors were irradiated starting 24 h after infection with 90 mW/cm² blue light for 10 days (8 h/day). Tumors were measured weekly; volume calculated as A×B^2×3.14/6. An additional figure description notes BALB/c nu/nu mice receiving 1×10^8 IFU on days 0 and 3, with 90 mW/cm² irradiation from day 0 to 8. Liver cancer model: Female Rag2/Il2rg double-knockout mice received CCl4 (0.5 mL/kg, i.p., twice weekly for 4 weeks) to induce liver injury. HepG2 cells (1×10^6 in 100 μL PBS) were transplanted intraperitoneally with simultaneous implantation of the hepatic LED device. When serum human AFP reached 50 ng/mL, mice received paOAd intravenously at 5×10^9 IFU/mouse; on day 3, a second injection at 5×10^12 IFU/mouse. Livers were irradiated at 1 mW/cm² blue light for 14 days (6 h/day) using the implanted LED; only livers were illuminated. A figure description alternatively reports 5×10^7 IFU on days 0 and 3 and irradiation at 1 mW/cm² from day 0 to 6 (1 h/day). Cell culture: HepG2 and HEK293 in DMEM high glucose with 10% FBS and penicillin–streptomycin; H1299 in RPMI-1640 with 10% FBS; A549 in DMEM high glucose with 10% FBS. Normal cells: HUVEC cultured with EGM-2; human small intestinal organoids (SIO) derived from human iPSCs. Assays: Cell viability by WST-8 and crystal violet staining. ELISAs measured human AFP (serum), mouse IL-6, and TNFα. Histology: livers fixed in 4% PFA, paraffin-embedded, 5 μm sections stained with H&E. Immunohistochemistry: xenograft sections stained for human AFP with Alexa Fluor 488 secondary; nuclei counterstained with DAPI. Flow cytometry sorting: CD133-positive and -negative HepG2 subsets via FACS Aria II; analysis in FlowJo. Adenoviral genome quantification by qPCR from total DNA (DNeasy kit) on StepOnePlus. Gene expression by RT-qPCR (SYBR Green, 2−ΔΔCT normalized to GAPDH). Statistical analyses included one-way ANOVA with Tukey’s post hoc tests and two-way repeated measures ANOVA with Tukey’s post hoc tests; data reported as mean ± S.E. Sample sizes included n=3 for some in vitro assays; in vivo cohorts included 13–16 mice per group depending on experiment.

- Optogenetic control of replication: Blue light induced GAVPO homodimerization, enabling Gal4(65) binding to UASG and robustly increasing E1A/E1B expression in tumor cells, thereby triggering adenoviral replication. - In vitro cytotoxicity is light-dependent: paOAd caused strong cytopathic effects and reduced viability in cancer cell lines (H1299, A549, HepG2) only when followed by blue light irradiation; minimal effects were observed under dark conditions. Crystal violet assays showed pronounced cell lysis with light. WST-8 viability assays confirmed significant decreases under light (p < 0.05 by ANOVA with Tukey’s post hoc). - Reduced off-target toxicity: In normal cells (HUVEC, hTERT-negative; and human small intestinal organoids, hTERT-positive), paOAd did not reduce viability without blue light, indicating low off-tumor toxicity compared to conventional OAd. - Subcutaneous xenograft efficacy: In mice with H1299 or HepG2 xenografts, intratumoral paOAd followed by 90 mW/cm² blue light markedly reduced tumor volumes; tumors nearly disappeared in the paOAd + light group. Across 2–5 weeks post-illumination, mean tumor volumes were significantly lower than controls (n=16 per group; p < 0.01 in post hoc comparisons). - Deep tissue (liver) model efficacy: Using implantable hepatic blue LEDs, paOAd + light significantly improved survival, markedly reduced tumor burden in the liver (H&E and AFP immunostaining showed near absence of HepG2 cells at week 6), and significantly decreased serum human AFP compared with controls. Inflammatory and injury markers (ALT, TNFα, IL-6) were significantly lower with light activation, indicating reduced liver injury/inflammation. - Lower dissemination and off-target activity: paOAd showed less transfer to spleen and intestines than conventional OAd (qPCR), and non-irradiated organs exhibited lower E1A expression with paOAd than with conventional OAd, supporting enhanced safety. - Activity against cancer stem cells: In Rag2/Il2rg mice engrafted with CD133-positive HepG2 cells, paOAd + light significantly reduced serum AFP over time (p < 0.01 at weeks 4–6), whereas Sorafenib did not, indicating paOAd’s efficacy against cancer stem-like populations. - Dosing and illumination parameters: Effective in vitro activation observed at 0.1 mW/cm² for 5 days; subcutaneous tumors responded to 90 mW/cm² (8 h/day for up to 10 days); liver model responded to 1 mW/cm² via implanted LEDs (6 h/day for 14 days).

The study demonstrates that adding an optogenetic GAVPO switch to an oncolytic adenovirus enables light-controllable E1A/E1B expression and thereby confines viral replication to illuminated tumor sites, addressing the central challenge of insufficient tumor selectivity. In vitro and in vivo data show strong antitumor efficacy with minimal effects under dark conditions, improving safety over conventional OAd strategies driven solely by tumor-selective promoters like hTERT. The subcutaneous and liver models confirm that spatially restricted illumination can drive potent oncolysis, lower tumor markers (AFP), reduce inflammatory and injury markers, improve survival, and limit off-target viral activity. The efficacy against CD133-positive cancer stem-like cells suggests potential to overcome resistance seen with standard therapies like Sorafenib. The findings validate optogenetic control as a viable mechanism to increase the therapeutic index of oncolytic adenoviruses. The authors note practical considerations for clinical translation, including improving light delivery to deep tissues and optimizing the transcriptional activation system to reduce required light doses and durations.

The authors engineered a photoactivatable oncolytic adenovirus (paOAd) whose replication is governed by blue light via a GAVPO-UAS system driving E1A/E1B expression. paOAd achieved robust, light-dependent oncolysis across cancer cell lines, demonstrated strong antitumor activity in subcutaneous and deep liver tumor models, reduced off-target effects compared to conventional OAd, and effectively targeted cancer stem-like (CD133-positive) cells. These results indicate that optogenetic control can substantially enhance the safety and efficacy of oncolytic adenoviral therapy. Future work should focus on clinical translation by improving light delivery (e.g., wireless or implantable devices, longer-wavelength or upconversion strategies), enhancing optogenetic activation domains to reduce light exposure, arming the virus with therapeutic payloads, and integrating systems like LACE or Split-CPS7.2 to further regulate replication and transgene expression.

- Blue light has limited tissue penetration, necessitating invasive or implantable illumination devices for deep tumors; clinical translation will require less invasive, deeper-penetrating light strategies (e.g., red/infrared with upconversion nanoparticles). - Long daily illumination (6–8 h/day in mice) raises concerns for phototoxicity and heat generation; stronger activation domains or more light-sensitive systems are needed to reduce exposure. - The GAVPO-p65 system may have limited activation efficiency; alternative activators (SAM, VPR, SunTag) or photoswitches (e.g., Magnet) could improve performance. - Preclinical nature of the data (mouse xenografts, immunodeficient models, implantable devices) limits generalizability to humans. - Some protocol details varied across experiments (e.g., dosing and illumination schedules), and long-term safety, immunogenicity, and biodistribution require further study.