Although there are emerging innovations of molecular imaging probes to detect and image tumors, most of these molecular dyes and nanoparticles have limitations of low targetability in tumors and fast clearance when administered systemically. In contrast, some bacteria, such as Escherichia coli MG1655, can selectively proliferate in a hypoxic environment inside of a tumor for several days, which highlights the potential for the development of a genetically encoded multimodal imaging probe to monitor the progress of the tumor. Here, we developed bimodal imaging tumor-homing bacteria (GVs-miRFP680 MG1655) that allow both optical and acoustic imaging in tumor-bearing mice. An in vivo optical image system and a Vevo 2100 imaging system were applied to detect different imaging properties of the engineered bacteria in vivo. Our results show that the GVs-miRFP680 MG1655 bacteria can effectively integrate the advantages of low tissue absorbance from near-infrared fluorescent proteins and non-invasiveness from gas vesicles. We successfully developed GVs-miRFP680 MG1655 bacteria, which have both acoustic and optical imaging abilities in vitro and in vivo. The acoustic signal can last for up to 25 min, while the near-infrared fluorescence signal can last for up to 96 h. The combination of different imaging modalities in the tumor-homing bacteria may contribute to the non-invasive monitoring of the therapeutic effect of bacterial therapy in the future.

Molecular imaging has demonstrated promise for evaluating the expression levels of biomarkers for the early prediction of tumor progression and metastasis. However, most of the commonly used molecular imaging modalities are relatively single and have difficulties imaging complex biological processes. Here, we fabricated αvβ3-integrin-targeted quantum-dots-loaded liposome–microbubble (iRGD-QDLM) complexes that combined ultrasound imaging with optical imaging. The resulting iRGD-QDLM has excellent binding capability to 4T1 breast cancer cells. Ultrasound molecular imaging of 4T1 tumors demonstrated that significantly enhanced ultrasound molecular signals could be observed in comparison with non-targeted QDLM. Importantly, our study also suggested that iRGD-QDL on the surface of microbubbles could be delivered into a tumor by ultrasound-mediated microbubble destruction and adhered to αvβ3 integrin on breast cancer cells, achieving transvascular fluorescent imaging. Our study provides a novel approach to dual-modality molecular imaging of αvβ3 integrin in the tumor tissue.

Nanoscale ultrasound contrast agents have attracted considerable interest in the medical imaging field for their ability to penetrate tumor vasculature and enable targeted imaging of cancer cells by attaching to tumor-specific ligands. Despite their potential, traditional chemically synthesized contrast agents face challenges related to complex synthesis, poor biocompatibility, and inconsistent imaging due to non-uniform particle sizes. To address these limitations, bio-synthesized nanoscale ultrasound contrast agents have been proposed as a viable alternative, offering advantages such as enhanced biocompatibility, consistent particle size for reliable imaging, and the potential for precise functionalization to improve tumor targeting. In this study, we successfully isolated cylindrical gas vesicles (GVs) from Serratia. 39006 and subsequently introduced the GVs-encoding gene cluster into Escherichia coli using genetic engineering techniques. We then characterized the contrast imaging properties of two kinds of purified GVs, using in vitro and in vivo methods. Our results demonstrated that naturally isolated GVs could produce stable ultrasound contrast signals in murine livers and tumors using clinical diagnostic ultrasound equipment. Additionally, heterologously expressed GVs from gene-engineered bacteria also exhibited good ultrasound contrast performance. Thus, our study presents favorable support for the application of genetic engineering techniques in the modification of gas vesicles for future biomedical practice.

Gas vesicles (GVs) are hollow protein nanostructures derived from microorganisms and show significant potential for ultrasound imaging applications. However, the direct production of gas vesicles (GVs) from their native hosts faces several limitations: poor contrast imaging performance, insufficient yield, and high costs associated with extraction and purification. These challenges heavily hinder their clinical translation and application. The heterologous expression of GV genes varies significantly among different chassis strains due to their distinct intracellular environments, which ultimately affects GV performance and yield. Therefore, it is crucial to select an appropriate chassis cell that can produce GVs with excellent imaging performance. In this study, the GV gene cluster from Serratia sp. ATCC 39006 was heterologously expressed in five different bacterial chassis strains: Escherichia coli BL21 (AI), Escherichia coli K-12 MG1655, Escherichia coli Nissle 1917, Salmonella YB1, and Vibrio natriegens. By systematically comparing the yield, particle morphology, and ultrasound imaging performance of GVs produced by these strains, we elucidated the impact of chassis cells on GV synthesis and function. This work provides experimental evidence and theoretical support for screening robust GV-producing strains and facilitates future biomedical applications of GVs.

Biosynthetic gas vesicles (GVs), as novel nanoscale ultrasound contrast agents, exhibit unique potential in biomedical ultrasound imaging. For example, they are expected to have better tissue penetration through the tumor vasculature for detecting tumor cells by the design of GV-based acoustic probes. Of all these GVs, GVs from Halobacterium sp. NRC-1 possess the largest size (over 200 nm) and are nearly spherical in shape, endowing them with stronger acoustic signals and better tumor penetration. However, their genetic manipulation is relatively difficult due to the requirement of a high-salt cytoplasmic environment for their expression and assembly, limiting the application of biosynthetic technology for modulating their structural features in heterologous host cells. In this study, we cloned the gene cluster encoding GVs from Halobacterium sp. NRC-1 and transformed it into Haloferax volcanii, an archaeal species naturally incapable of producing GVs. The genetically engineered Haloferax volcanii successfully synthesized functional GVs (GV_vol) with a similar size and shape to naturally synthesized GVs from Halobacterium sp. NRC-1 (GV_halo). The ultrasound imaging properties of GV_vol heterologously synthesized in Haloferax volcanii were compared with naturally synthesized GV_halo in vitro and in vivo, showing that GV_vol could achieve a mean signal intensity of 113.6 ± 0.9 a.u. in vitro and a peak intensity of 121.5 ± 0.8 a.u. in vivo in the kidney, compared with 115.7 ± 0.5 a.u. and 119.0 ± 0.5 a.u. for GV_halo, respectively. These findings confirm the functional integrity of heterologously synthesized GV_vol and its potential for biomedical applications. Our study provides a solid experimental foundation for genetically tailoring Halobacterium GV properties to optimize biomedical imaging performance.