Wild-type C57BL/6 mice or Balb/c mice, aged 5-7 weeks, were purchased from Charles River and kept at Experimental Biomedicine (EBM) at the University of Gothenburg, Sweden. The research received approval from the local Animal Ethics Committee in Gothenburg, Sweden (permit no. Dnr 5.8.18-10595/2023). Some experiments were approved by the Axis Bioservices Animal Welfare and Ethical Review Committee in the United Kingdom (Project license number PPL2862), and all procedures followed the guidelines outlined in the Animal (Scientific Procedures) Act 1986.

E. coli cultures were pelleted and resuspended in 20 mM Tris-HCl (pH 8.0) with 20% sucrose, followed by lysozyme (600 µg/g cells) and 0.1 M EDTA (0.2 mL/g cells) treatment. Spheroplasts were pelleted and sonicated in 10 mM Tris-HCl (pH 8.0) at 4 °C using a Q55 Sonicator (QSonica, Newtown, CT). Membranes were isolated by centrifugation at 40,000 × g for 1 h at 4 °C, then incubated with 0.5% Sarkosyl (Sigma Aldrich, St. Louis, MO) for 20 min and with 200 mM Na₂CO₃ (pH 11) for 1 h at 25 °C to generate outer membrane sheets. Samples were applied to a step gradient (50%, 30%, and 10% iodixanol; Axis-Shield PoC AS, Oslo, Norway) and ultracentrifuged at 100,000 × g for 2 h at 4 °C. Fractions between 10% and 30% were collected, sonicated in an ultrasonic bath (Grant, Cambridge, UK), and designated as SyBV.

OMV derived from E. coli were isolated using a previously described protocol with modifications^[11]. Briefly, bacterial cultures were centrifuged twice at 6,000 × g at 4 °C for 20 min, the supernatant was filtered and concentrated, and OMV were collected by ultracentrifugation at 150,000 × g for 3 h and resuspended in phosphate-buffered saline (PBS).

The isolated SyBV were analyzed by negative staining for TEM. These vesicles were applied to glow-discharged 200-mesh formvar- and carbon-coated copper grids (Electron Microscopy Sciences, Hatfield, PA) for 5 min, rinsed with water, and fixed with 2.5% glutaraldehyde. After two washes in H₂O, samples were stained with 2% uranyl acetate for 1.5 min. Images were acquired using a Talos L120C transmission electron microscope (Thermo Fisher Scientific, Waltham, MA) operated at 120 kV and equipped with a charge-coupled device (CCD) camera.

SyBV were diluted in PBS, and vesicle numbers were quantified using a ZetaView analyzer (Particle Metrix GmbH, Meerbusch, Germany). Analyses were performed in triplicate, with each data point derived from two stationary layers and five measurements per layer. Camera sensitivity was set to 70 for all measurements, and data were processed using ZetaView software version 8.2.30.1.

Bacterial OMV and SyBV were resolved on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred onto a polyvinylidene difluoride (PVDF) membrane. After blocking, the membrane was probed with an anti-OmpF antibody or an anti-lipid A antibody (ThermoFisher Scientific, Waltham, MA). The bound antibodies were detected using a horseradish peroxidase-linked secondary antibody, and the signals were visualized by chemiluminescence.

DNA from SyBV or OMV was extracted using the Qiamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. The quality, concentration, and fragment size of 1 µL of the extracted DNA were assessed via capillary electrophoresis with an Agilent High Sensitivity DNA chip on an Agilent 2100 Bioanalyzer (Agilent Technologies GmbH, Berlin, Germany).

The lipopolysaccharide (LPS) level of SyBV or OMV was measured using the LAL chromogenic endotoxin quantification kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions.

B16F10 (melanoma) cells were seeded in a 96-well plate (2 × 10⁴ cells per well) and treated with SyBV (10⁹, 3 × 10⁹, or 10¹⁰ particles/ mL) for 24 h. MTT (Sigma-Aldrich, St. Louis, MO) was added to the cells, followed by incubation with dimethyl sulfoxide to solubilize the formazan crystals. Absorbance was measured at 570 nm with a reference wavelength of 630 nm using a microtiter plate reader.

B16F10 (melanoma), EMT6 (breast cancer), LL/2 (Lewis lung cancer), and RENCA (renal cancer) cell lines were obtained from ATCC (Manassas, VA), while the MC38 (colon cancer) cell line was purchased from Sigma-Aldrich (St. Louis, MO). A total of 5 × 10⁵ cells from each mouse cancer cell line were subcutaneously inoculated into the flanks of C57BL/6 or Balb/c mice. Once tumors reached approximately 10-100 mm³, mice were randomly assigned to either sham or SyBV-treatment groups. In the B16F10 melanoma model, mice received intratumoral injections of SyBV (10⁹ or 10¹⁰) five or six times at 3-day intervals. In the other tumor models, SyBV (10¹⁰) was injected intratumorally five or six times at 3-day intervals. For all models, anti-mouse PD-1 antibody (100 µg; BioXcell, West Lebanon, NH) was administered intraperitoneally 1 day prior to immunization with SyBV. Tumor growth was monitored every 2-3 days with calipers, and tumor volume was calculated as (width)² × (length)/2. Mouse serum was diluted 1:500 in PBS containing 1% bovine serum albumin (BSA) and applied to 96-well plates coated with 200 ng of tumor vesicular proteins. After a 2-h incubation, bound antibodies were detected using a peroxidase-conjugated anti-mouse IgG antibody (Thermo Fisher Scientific, Waltham, MA).

Created with BioRender.com.

All data are presented as mean ± standard error of the mean (SEM). Independent replication of each tumor experiment was not performed. All data were tested for normal distribution using the Shapiro-Wilk test. For comparisons between two groups, an unpaired two-tailed Student’s t-test was used when data were normally distributed with similar variance. Otherwise, a nonparametric Mann-Whitney test was applied. For experiments with multiple groups or two independent variables, two-way analysis of variance (ANOVA) was applied, followed by Tukey’s post hoc test for multiple comparisons when normality and equal variance assumptions were met. When data did not meet normality assumptions, the nonparametric Kruskal-Wallis test was used, followed by Dunn’s post hoc test for pairwise comparisons. Effect sizes and variance estimates were considered to determine whether parametric or nonparametric tests were appropriate. Multiple comparison corrections were applied consistently according to the chosen post hoc tests. A significance level of P < 0.05 was considered statistically significant. All the statistical analyses were performed using GraphPad Prism 10 (GraphPad Software Inc., San Diego, CA).

SyBV were efficiently produced from E. coli membranes using a stepwise purification procedure [Figure 1A]. Briefly, bacterial pellets were enzymatically digested with lysozyme, followed by sonication in the presence of sarkosyl detergent to disrupt the inner membrane. Subsequent exposure to high pH conditions removed residual cytosolic contents, yielding enriched outer membrane fractions. These fractions were subjected to buoyant density gradient ultracentrifugation (10%-30% iodixanol), which enabled efficient separation from cellular debris and protein aggregates. The resulting outer membrane sheets were converted to SyBV through mild sonication, producing nanosized vesicles with high structural uniformity. NTA [Supplementary Figure 1A] and TEM [Figure 1B] analyses confirmed that SyBV were spherical, with diameters ranging between 50-200 nm (average size: 153.6 ± 1.2 nm; n = 3 independent preparations). Importantly, the preparations were devoid of detectable contaminants such as pili-like structures or cell fragments, indicating high purity [Figure 1B]. Moreover, OmpF, a major constituent of the outer membrane implicated in immune responses^[12], was identified in SyBV [Supplementary Figure 1B]. SyBV were generated in substantially greater amounts than naturally derived OMV from an equivalent culture volume [Supplementary Figure 1C]. In addition, SyBV exhibited higher purity, as indicated by a greater particle-to-protein ratio [Supplementary Figure 1D] and the absence of DNA contamination [Supplementary Figure 1E], in agreement with the TEM observations. Interestingly, LPS levels in SyBV were found to be comparable to those in OMV [Supplementary Figure 1F and G]. The electron microscopy further confirmed that the structural integrity of SyBV is maintained for at least two weeks, indicating a high degree of stability [Supplementary Figure 1H]. Collectively, these findings demonstrate the successful production of highly pure SyBV, establishing a reliable foundation for subsequent in vivo functional studies.

We evaluated the anti-tumor efficacy of SyBV in the subcutaneous B16F10 melanoma model, a well-established “cold” tumor that lacks T cell infiltration and is resistant to checkpoint blockade^[13]. We previously demonstrated that a maximum dose of 2.5 × 10¹¹ SyBV caused no changes in body temperature and did not induce systemic inflammation^[11]. On the basis of these findings, a higher dose of 10¹⁰ SyBV per administration was selected and given six times. When tumors reached 10-30 mm³, mice received intratumoral injections of SyBV at two doses (10⁹ or 10¹⁰ particles) directly into the tumor, and tumor growth was monitored [Figure 2A]. Individual tumor growth for each mouse from Figure 2A is presented in Figure 2B-E. Anti-PD-1 monotherapy was included to confirm the model’s inherent resistance. As expected, PD-1 blockade alone failed to significantly inhibit tumor growth compared to the sham-treated group [Figure 2B and C]. In contrast, SyBV administration in combination with anti-PD-1 robustly suppressed tumor progression at both doses [Figure 2D and E]. Even at the lower dose (10⁹), SyBV plus anti-PD-1 reduced tumor volume by 92%, (P < 0.001 compared to sham) with two of nine mice exhibiting complete regression following repeated treatment [Figure 2D]. The higher dose (10¹⁰) achieved nearly 99% tumor reduction (P < 0.001), and six of nine mice displayed complete tumor clearance by three weeks [Figure 2E]. Moreover, SyBV alone significantly inhibited melanoma growth, consistent with previous reports^[14], and the combination of SyBV with an anti-PD-1 antibody exhibited an even more pronounced effect [Supplementary Figure 2]. These results demonstrate that SyBV can convert anti-PD-1-resistant tumors into responsive ones, highlighting its potential as a sensitizing agent for checkpoint blockade. Notably, the higher dose (10¹⁰) of SyBV did not induce any increase in systemic pro-inflammatory cytokines during either the acute phase (6 h) or the chronic phase (Day 21), whereas OMV provoked substantial cytokine release [Supplementary Figure 3]. Moreover, no local skin reactions were observed in any SyBV-treated mice, even at the higher dose and with repeated administration, indicating a favorable safety profile (Data not shown for representative images). We next examined the viability of melanoma cells treated with SyBV and found no direct effect [Supplementary Figure 4], suggesting that their anti-tumor effects are indirect and likely mediated through immune cells.

We have hypothesized that intratumoral administration of SyBV might elicit systemic effects that modulate the growth of distal, non-injected tumor tissues. To test this, mice were engrafted with B16F10 melanoma cells on both the left and right flanks, and only one tumor was treated locally with two different doses of SyBV [Figure 3A]. The growth of the injected primary tumor was shown in Figure 2. Remarkably, a significant reduction was observed in the untreated contralateral tumors at the higher dose of SyBV plus anti-PD-1 (10¹⁰; Figure 3A). Although distal tumors exhibited a notable inhibition (approximately 50%-60% reduction; P < 0.01 compared to sham), this inhibition was slightly weaker than the potent suppression at the directly treated tumor sites [Figure 2]. In contrast, systemic administration of anti-PD-1 antibodies alone had no measurable effects on either injected or contralateral tumors. We also confirmed significantly elevated serum IgG antibody levels against tumor-associated proteins in SyBV-immunized mice (approximately 3-fold increase; P < 0.01; Figure 3B), which may partly explain the systemic anti-tumor effects observed at distal sites. These findings highlight the potential of SyBV plus anti-PD-1 as a local immunotherapeutic agent with broader systemic efficacy. However, the precise mechanisms underlying this effect remain to be elucidated.

Since early intervention significantly improves cancer prognosis^[15], we examined whether SyBV efficacy depends on initial tumor volume. Mice were inoculated with B16F10 melanoma and grouped based on tumor size at treatment onset: small tumors (10-30 mm³) and larger tumors (50-100 mm³). SyBV plus anti-PD-1 induced near-complete inhibition of tumor growth in mice with small, palpable tumors (P < 0.001 compared to sham), whereas efficacy was less in larger tumors, indicating greater sensitivity at early stages (P < 0.01 for small vs. large tumors; Figure 4). However, when treatment was initiated in mice with larger tumors, SyBV plus anti-PD-1 still achieved approximately 50% inhibition of tumor growth by day 10 of treatment (P < 0.05 compared to sham). These results demonstrate that the combination retains potent anti-tumor activity across tumor stages, with effective tumor reduction observed in both small and more established tumors.

We next evaluated whether SyBV plus anti-PD-1 exhibit broad therapeutic potential across different tumor types. To address this, we employed four syngeneic mouse tumor models, MC38 (colon), EMT6 (breast), LL/2 (Lewis lung), and RENCA (kidney), for in vivo experiments. Among the tested models, colon cancer showed a greater response than the other tumor types, with the treatment resulting in over 50% growth inhibition (P < 0.01 compared to sham; Figure 5A). Breast cancer showed a more moderate but significant reduction in growth, indicating partial sensitivity of this type of tumor (P < 0.05; Figure 5B). In contrast, Lewis lung and kidney tumor models showed minimal response to SyBV treatment [Figure 5C and D]. These observations highlight the selective efficacy of SyBV across certain tumors types. Furthermore, SyBV plus anti-PD1 might not be sufficient to induce robust anti-tumor effects in all tumor types tested.