Ultrasound-controllable engineered bacteria for cancer immunotherapy

The study addresses how to safely and effectively deploy engineered bacteria for cancer therapy by restricting therapeutic activity to tumors. Although immune cell therapies are effective in hematologic malignancies, their performance in solid tumors is limited by immunosuppressive microenvironments. Conversely, certain bacteria preferentially colonize hypoxic tumor cores and can be engineered to secrete anti-cancer agents. However, systemically delivered bacteria also engraft in healthy organs (e.g., liver, spleen), risking off-target toxicity if payloads are constitutively expressed. The authors propose to solve this by harnessing focused ultrasound (FUS) to provide noninvasive, spatially precise thermal control of bacterial gene expression. They aim to create a temperature-responsive genetic switch that converts a brief thermal trigger into sustained therapeutic production, thereby localizing therapy to FUS-targeted tumors while minimizing off-tumor activity.

Existing control mechanisms for microbial therapies include chemical inducers and promoters responsive to tumor microenvironments, but these lack precise spatial control. Optogenetic control offers spatial precision but is limited by poor light penetration in tissues. Heat-inducible promoters can be targeted by deeply penetrant energy modalities like FUS. Prior demonstrations used temperature-dependent responses to modulate bacterial genes with FUS, but mainly in cloning strains, with non-therapeutic outcomes and transient activation inadequate for multi-week tumor treatment. Bacteria such as probiotic E. coli Nissle 1917 (EcN) preferentially colonize tumors and have been used to deliver checkpoint inhibitors (e.g., anti-CTLA-4, anti–PD-L1), but systemic dosing of antibodies can cause immune-related adverse events. The work builds on serine integrase-based state switches for permanent DNA inversion to achieve persistent expression post-trigger and leverages RNA thermometers/terminators and circuit tuning (RBS, start codon, degradation tags) to reduce leak and enhance induction.

- Thermal bioswitch characterization: Selected temperature-dependent transcriptional repressors (TipA, Tcl1, Tcl42) and tested in the therapeutically relevant EcN strain. Constructed reporter circuits where repressors control GFP; measured OD-normalized fluorescence from 37–42 °C to quantify fold-change between 37 °C and 42 °C. Tcl42 exhibited effective thermal transduction in EcN. - Heating protocol optimization: Determined that 1 h at 42 °C is needed for robust activation. Assessed pulsatile heating (alternating 37 °C/42 °C; 50% duty cycle) with pulse durations from 1–60 min over 2 h total to optimize induction versus viability, selecting 5-min pulses as a balance of strong induction and improved cell viability. - Permanent state switch design: Placed Bxb1 serine integrase under control of λ pL/pr promoters regulated by Tcl42. Thermal derepression induces Bxb1, which inverts an attP/attB-flanked P7 promoter to turn on a reporter and a downstream payload cassette (tetracycline resistance placeholder, later replaced by therapeutics). Post-inversion sequence is no longer substrate for Bxb1, creating a stable, permanent ON state. Included strong terminators to insulate promoter activity; selected P7 promoter for sustained expression without excessive burden. - Circuit optimization by library screening: Randomized and screened key translation/stability elements affecting Bxb1 (RBS, start codon including non-canonical GUG, and C-terminal sarA/sSRA degradation tag tripeptides) to reduce baseline leak and maximize induction. Conducted high-throughput replica-plating screens: one plate incubated at 37 °C, another heat-pulsed at 42 °C for 1 h then at 37 °C; colonies imaged for GFP. Selected low-leak/high-activation variants and quantified switching. Chose a top performer for further engineering. - Additional leak reduction: (i) Increased plasmid copy number from low-copy pSC101 to medium-copy p15A; (ii) inserted a temperature-sensitive RNA terminator upstream of Bxb1 to suppress translation at 37 °C while permitting expression at ≥42 °C. Combined modifications significantly reduced leak while maintaining high fold-change and conversion upon induction. - Therapeutic payload engineering: Replaced payload placeholder with PelB-tagged nanobodies targeting CTLA-4 (αCTLA-4) or PD-L1 (αPD-L1) to promote secretion. Added Axe-Txe toxin–antitoxin cassette to enhance plasmid stability without antibiotic selection. Validated inducibility: measured percent activation after 1 h induction at 37 °C, 42 °C, or 43 °C; performed Western blots on conditioned media collected 24 h post-induction to confirm secretion at elevated temperatures only. - Focused ultrasound (FUS) platform: Built a closed-loop, feedback-controlled FUS hyperthermia system capable of alternating intratumoral temperatures between 37 °C and 43 °C in live mice in 5-min steps, mimicking clinical systems. Temperature feedback sampled at 1 Hz; PID control adjusted acoustic power to maintain setpoints. - In vivo efficacy study: Established subcutaneous A20 tumors in BALB/c mice (~100 mm³). Systemically injected 1×10^8 engineered EcN (1:1 αCTLA-4:αPD-L1 strains). Allowed 2 days for tumor colonization, then applied FUS for 1 h at 43 °C using 5-min pulses (duty cycle described). Tracked tumor volumes for up to two weeks. Controls included wild-type EcN, therapeutic EcN without FUS, pre-activated therapeutic EcN, and systemic monoclonal antibodies (αPD-L1 + αCTLA-4). Post-study, harvested tumors, livers, and spleens; quantified CFUs and percentage of activated bacteria by GFP to assess localization and durability of activation (up to 2 weeks). Statistical analyses used ANOVA and nonparametric tests as noted in figure captions.

- Tcl42 functions as an effective thermal transducer in EcN, enabling strong induction at 42 °C with low baseline at 37 °C. A 1-hour induction at 42 °C is sufficient for robust activation. Pulsatile heating with 5-min pulses preserves activation while improving viability relative to continuous heating. - A Tcl42→Bxb1 integrase state switch converts a brief thermal input into a permanent ON state, driving sustained payload expression after heat is removed. - High-throughput screening of Bxb1 translation/stability elements (RBS, start codon, sarA/ssrA tag tripeptides) identified variants with low leak and high inducibility. Additional incorporation of a temperature-sensitive RNA terminator and increased plasmid copy number further reduced baseline activity while preserving high fold-change. - Therapeutic circuits secreting PelB-tagged αCTLA-4 or αPD-L1 nanobodies show minimal activation at 37 °C and robust activation at 42–43 °C. Western blots of culture media detect nanobody secretion only after thermal induction. - In vivo, feedback-controlled FUS reliably toggles tumor temperature between 37 °C and 43 °C in 5-min steps. Systemically delivered therapeutic EcN activated by FUS significantly suppress tumor growth compared to controls (two-way ANOVA with Dunnett’s multiple comparisons; reported p values include p<0.004, 0.038, 0.003, and p<0.0001 across time points). Tumor growth inhibition is comparable to systemic checkpoint antibody therapy and to pre-activated bacteria. - Bacterial activation is spatially localized: CFU and GFP analyses indicate high activation within FUS-targeted tumors with minimal activation in liver and spleen; activation persists for at least two weeks post-FUS. A reported comparison of activation levels shows a significant difference between tumors and bystander tissues (e.g., one-tailed Mann–Whitney p=0.0005 in a summarized measure). One of nine FUS-treated tumors completely regressed in this cohort.

The work demonstrates that engineered probiotic bacteria can be remotely and locally activated in tumors using FUS hyperthermia, converting a transient temperature cue into sustained therapeutic protein secretion via an integrase-based state switch. This addresses the key challenge of off-tumor toxicity by restricting activation to the insonated tumor volume while allowing systemic administration for tumor homing. Optimized circuit elements minimize basal leak at physiological temperature and improve the induction window, making long-term in situ therapy feasible without repeated daily inductions. In vivo results confirm significant, localized tumor growth suppression comparable to systemic checkpoint blockade but with the potential for reduced systemic exposure. The approach integrates spatial (FUS targeting) and temporal (permanent switch) control, offering a generalizable platform for precise delivery of potent payloads within solid tumors.

The study introduces a focused ultrasound–controllable genetic state switch in tumor-homing EcN that enables sustained, localized secretion of checkpoint inhibitor nanobodies after a brief, noninvasive thermal trigger. Through rational design and high-throughput screening, the authors minimized baseline leak and maximized inducibility, achieving durable in vivo activation and significant tumor growth suppression with spatial confinement to FUS-targeted tumors. This platform may be extended to other therapeutic payloads and microbial chassis, and integrated with clinical FUS systems for precise, on-demand control of bacterial therapies in oncology. Future work could expand to additional tumor models, evaluate long-term safety and immunological consequences, refine heating protocols to balance efficacy and tissue safety, and explore multiplexed payloads or alternative energy modalities for activation.

- The efficacy was demonstrated in a single murine tumor model (A20) and in immunocompetent mice; broader tumor types and hosts remain to be tested for generalizability. - FUS hyperthermia required intratumoral temperatures up to 43 °C, which may cause some thermal damage; while acceptable in tumors, the therapeutic window and potential collateral effects need further evaluation. - Some background activation in bystander tissues was measured, though substantially lower than in targeted tumors; complete elimination of off-target activation has not been established. - The approach depends on sufficient bacterial colonization of tumors after systemic delivery, which can vary with tumor type, size, and host factors. - Long-term persistence, horizontal gene transfer risk, and comprehensive safety/toxicity profiles of engineered EcN expressing immunomodulatory payloads require additional study for clinical translation.