Engineered Cancer Targeting Microbes and Encapsulation Devices for Human Gut Microbiome Applications

Abstract

The gut microbiota produce specialized metabolites that are important for maintaining host health homeostasis. Hence, unstable production of these metabolites can contribute to diseases such as inflammatory bowel disease and colon cancer. While fecal transplantation or dietary modification approaches can be used to correct the gut microbial community’s metabolic output, this Perspective focuses on the use of engineered bacteria. We highlight recent advances in bacterial synthetic biology approaches for the treatment of colorectal cancer and systemic tumors and discuss the functionality and biochemical properties of novel containment approaches using hydrogel-based and electronic devices. Synthetic circuitry refinement and incorporation of novel functional modules have enabled more targeted detection of colonic tumors and delivery of anticancer compounds inside the gastrointestinal (GI) tract, as well as the design of tumor-homing bacteria capable of recruiting infiltrating T cells. Engineering challenges in these applications include the stability of the genetic circuits, long-term engraftment of the chosen chassis, and containment of the synthetic microbes’ activity to the diseased tissues. Hydrogels are well-suited to the encapsulationo of living organisms due to their matrix structure and tunable porosity. The matrix structure allows a dried hydrogel to collect and contain GI contents. Engineered bacteria that sense GI tract inflammation or tumors and release bioactive metabolites to the targeted area can be encapsulated. Electronic devices can be enabled with additional measuring and data processing capabilities. We expect that engineered devices will become more important in the containment and delivery of synthetic microbes for diagnostic and therapeutic applications.

The gut microbiome plays a crucial role in human health by controlling digestion, supporting the immune system, and preventing colonization by certain pathogens. The microbiota absorb energy and nutrients from their host’s diet that they use for growth and the production of specialized metabolites. Several of these microbiome metabolites benefit the host by maintaining homeostasis, and a disruption of their production has been associated with diseases such as colon cancer.^1,2 Strategies for reestablishing host homeostasis after a disruption include dietary modifications, fecal transplantation, and the delivery of probiotic or engineered bacteria to the GI tract. These strategies are not currently used for clinical cancer treatment, and further development and clinical trials will be required to demonstrate their efficacy in patients. Most of these strategies are nontargeted with potential outcomes that are difficult to predict from the onset. Use of synthetic microbial therapeutics allows them to be tailored to the treatment of specific disease conditions with minimal impact on the microbial community.

Genetic engineering and the development of synthetic biological parts such as memory circuits, heterologous production modules, and logic gates enabled the construction of bacterial strains with unique diagnostic and therapeutic characteristics.³ This is achieved via heterologous introduction of transgenes into a plasmid or insertion into the bacterial chromosome. Bacteria are attractive tools due to their sensing capabilities and various complementary molecular mechanisms for cell–cell communication. While the most commonly used chassis is still Escherichia coli, commensal or probiotic microbes (e.g., Bacteroides or lactic acid bacteria) and attenuated pathogens (e.g., Salmonella or Listeria) are gaining popularity as they might be better tailored to specific applications. Diagnostic synthetic bacteria can be engineered to sense autoinducers or measure specific metabolite levels and report on these detection events by triggering the expression of fluorescent proteins, β-galactosidase- or integrase-based memory switches, and CRISPR-Cas recording devices.⁴⁻⁸ The use of engineered therapeutic bacteria allows tailored control over a specific functionality within the microbiome (e.g., delivery of antibiofilm enzymes by E. coli), regardless of whether this functionality was naturally present in a subject’s microbiota prior to the introduction of engineered bacteria.⁹ Moreover, both diagnostic and therapeutic functionalities can function in parallel and be incorporated into engineered bacteria capable of diagnosing a specific disease and produce therapeutic molecules as a response.¹⁰

The use of engineered bacteria to perform in vivo diagnostic and therapeutic functions raises certain concerns regarding safety and potential spread to the environment. To address these concerns, genetic kill switches can be integrated to remove the chassis when it is no longer needed¹¹ or the engineered strains can be physically contained by encapsulation. Common encapsulation methods include the use of alginate, polymers, or hydrogels.¹² An increase in the number of polymer layers and combination of various polymers increase the robustness and biocontainment, thereby decreasing the risk of bacterial escape. To ensure therapeutic functionality, metabolites could be delivered across hydrogel pores that can be made small enough to prevent passage of the bacteria themselves. A hydrogel is synthesized using hydrophilic polymers, which are cross-linked to form a two-dimensional or a three-dimensional (3D) mesh network. This allows the hydrogel to hold water and swell while maintaining its structure.¹³ The polymers that act as the hydrogel backbone are combined with an initiator, which makes the polymer reactive and subsequently allows a cross-linking reagent to link the polymer chains together.¹⁴ Hydrogels are widely used in environmental and biomedical applications due to their biocompatibility and tunable characteristics, including porosity, biodegradation, mechanical flexibility, and chemical functionalization. Hydrogels are typically classified as “naturally derived” or “synthetic”, depending on the polymers that are used for their synthesis. Natural hydrogels are those that are found naturally in the body or can be extracted naturally from animals or plants. These include chitosan, alginate, cellulose, gelatin, collagen, and hyaluronic acid. Natural hydrogels are preferred over synthetics for biomedical applications due to their enhanced biocompatibility, typically low immunogenicity, and increased biodegradation rate. Synthetic polymers include poly(vinyl alcohol), poly(ethylene oxide), and poly(propylene fumarate). These polymers can be used to form synthetic hydrogels, with highly tunable properties and more control of the degradation rate.

Synthetic biology approaches to design bacteria as drug delivery devices have been reviewed extensively in the past.^3,11,15,16 This Perspective focuses on recent advances in the use of bacteria for tumor targeting and the development of engineered containment devices to address potential biosafety concerns.

Synthetic Bacteria for In Vivo Tumor Targeting Applications

Synthetic Circuitry Enabling Targeted Delivery in the Gastrointestinal Tract

Delivery of anticancer agents is often complicated by their inherent toxicity to healthy, nontarget tissues. Recent advances in synthetic microbe design incorporate modules for specific cancer microenvironment detection combined with targeted therapeutic delivery, allowing the synthetic strains to act where and when needed. Intestinal inflammation is associated with several GI cancers and inflammatory bowel diseases. The often chronic and recurring nature of this inflammation poses a major challenge for the majority of synthetic chassis, which are typically unstable weeks after the final administration. The main reasons for instability are genetic pressure on the engineered circuits or the inability of the bacterial host to stably engraft in the microbiota. To overcome these challenges, Riglar et al. developed a strategy for engineering a murine commensal E. coli chassis to respond to the inflammation biomarker tetrathionate as a long-term diagnostic.⁴ Tetrathionate is formed as a transient product of thiosulfate that is oxidized by reactive oxygen species in the inflamed gut lumen. Introduction of the TtrR/TtrS two-component system from Salmonella enterica serotype Typhimurium enabled the synthetic murine E. coli to sense tetrathionate and induce expression of the Cro regulatory protein from Escherichia phage λ. Cro expression in turns triggers a second synthetic memory circuit that maintains Cro expression in concert with continuous expression of β-galactosidase, including after exposure to tetrathionate is ceased. This results in recording of exposure memory, which can be read out as blue colonies after plating feces from inflamed animals on agar media containing 4-chloro-3-indoyl-β-d-galactopyranoside (X-gal). The murine E. coli chassis exhibited stable colonization in vivo, and the engineered synthetic circuits were found to be remain stable and functional for >6 months after initial colonization.⁴

Ho et al. designed an elegant synthetic strategy that combines specific adherence to colorectal cancer cells with the delivery of an anticancer agent produced from dietary precursors.¹⁷ A commensal E. coli strain was engineered to express histone-like protein A (HlpA) from Streptococcus gallolyticus, allowing the bacteria to adhere to the heparin sulfate polysaccharides displayed on colorectal cancer cells. In addition, a secretable myrosinase was introduced, which is capable of producing the anticancer compound sulforaphane from the glucosinolates present in cruciferous vegetables. Sulforaphane exerts its anticancer activity by interfering with the cell cycle and inducing apoptosis. Upon in vitro supplementation of this engineered strain with glucosinolates, it successfully attenuated the growth of colorectal adenocarcinoma cell lines. In combination with dietary input, this strain achieved significant reductions in tumor size and numbers in an azoxymethane/dextran sodium sulfate (AOM/DSS) colitis-associated mouse model for colorectal cancer.¹⁷ In a separate study, Chung et al. engineered Pediococcus pentosaceus, a lactic acid bacterial chassis that is generally regarded as safe for human use.¹⁸ The bacterium expressed and secreted the therapeutic protein P8, and this was effective to reduce tumor volumes in the AOM/DSS colorectal cancer model. Interestingly, this study also takes into account the effect of the synthetic microbe on the microbiome composition of the experimental animals. Whereas AOM/DSS-induced colitis expectedly alters the microbial community, the orally introduced P. pentosaceus strain partially alleviated this chemically induced dysbiosis.¹⁸

Tumor-Homing Synthetic Bacteria

Facultative anaerobic bacteria are known to target solid tumors and thrive in their characteristic immunosuppressive and hypoxic microenvironments.¹⁹ These properties make tumor-homing bacteria an attractive chassis for delivery of immunostimulatory agents to the tumor microenvironment, which recruits infiltrating T cells that can kill the cancer cells. Canale et al. developed a unique approach to allow synthetic bacteria to modulate tumor metabolism.²⁰ Their strategy is designed on the principle that the successful outcome of treatments with programmed death-ligand 1 (PD-L1) targeting antibodies is dependent on local l-arginine levels. The authors rewired the regulatory and metabolic restrictions of the l-arginine biosynthesis pathway, enabling an E. coli Nissle 1917 chassis to produce l-arginine from ammonia, which typically accumulates in the tumor environment. Upon colonization of subcutaneously implanted MC38 colon adenocarcinoma tumors with this engineered strain, Canale et al. observed increased numbers of tumor-infiltrating CD4⁺ and CD8⁺ T cells, which in turn increased the PD-L1 therapy efficacy and tumor clearance. Interestingly, bacterial survival and continuous delivery of l-arginine were required to achieve the synergistic effect of the PD-L1 therapy, because direct intratumoral injection of l-arginine did not reduce tumor growth. Intravenous injection of the engineered bacteria could be used to target larger (>100 mm³) non-accessible tumors, while smaller tumors were not consistently colonized. Systemic administration also caused decreases in body weight for the animals, which could indicate the bacteremia had adverse side effects.²⁰

Targeting of systemic tumors is a vital component of strategies for combating pancreatic cancer, which is characterized by high levels of metastasis at the time of diagnosis. Selvanesan et al. used a Trojan Horse strategy to have an intracellular bacterium deliver an immunostimulatory payload.²¹ The authors engineered an attenuated strain of the intracellular pathogen Listeria monocytogenes as a chassis to produce the immunogenic tetanus toxoid (TT) protein fused to a truncated, noncytolytic form of the secreted protein listeriolysin O. L. monocytogenes infects myeloid-derived suppressor cells (MDSCs), which in turn are attracted to tumor microenvironments. Upon arrival in these immunosuppressive environments, the Listeria can spread from the MDSCs into the tumor cells. Intratumoral accumulation of TT resulted in the reactivation and recruitment of preexisting TT-specific memory CD4⁺ T cells. In combination with a treatment with the nucleoside analogue Gemcitabine, Listeria-TT-treated mice resulted in a CD4⁺ T cell-dependent reduction in tumor burden and metastases, along with increased survival compared to nontreated controls.²¹

Qin et al. employed an original strategy that targets the vascular networks that supply solid tumors with oxygen and nutrients to sustain their growth.²² They use an attenuated E. coli chassis that is engineered to express the cytotoxic pore-forming protein cytolysin A from S. typhimurium under the control of the acid sensitive promoter adiA. The construct was fused to green fluorescent protein to allow for tracking of the bacteria in cell cultures as well as in vivo using fluorescent imaging. Upon infiltration, the acidic tumor environment is sensed by the bacteria and switches on expression of cytolysin A. This triggers thrombus formation and disruption of the blood vessels that supply oxygen and nutrients to the tumor tissue, thereby attenuating growth and metastasis. Using this strategy raises concerns regarding potential thrombus formation in healthy peripheral tissues, which would be detrimental to host health. The authors investigated the spread of the synthetic chassis to heart, liver, spleen, lung, and kidneys and observed colonization of these tissues that persisted up until their end point at 96 h, albeit at much lower levels compared to those of the bacteria recovered from the tumor tissues. While no lesions were observed in peripheral tissues after a 2-week treatment, the long-term effects of peripheral tissue colonization must still be evaluated. Building in additional fail-safe mechanisms could help address concerns regarding leaky or inappropriate expression of potentially toxic proteins. Magaraci et al. generated a similar cytolysin A-expressing E. coli therapeutic strain but drove expression with a tightly controlled light-inducible system.²³ While cytolysin production was assessed only in vitro, these types of systems could be used as an additional control measure where the toxin is produced only when the patient is treated with long-wavelength, tissue-penetrating light.²³

Systemic delivery of bacteria is often hampered by their natural clearance from the bloodstream by the host immune system. One commonly used strategy for overcoming this challenge is to increase the payload or increase the efficiency of its delivery by the synthetic microbes. Harimoto et al. used an alternative strategy focusing on stabilizing the bacterial delivery vehicle.²⁴ They introduced a dynamically inducible isopropyl β-d-1-thiogalactopyranoside (IPTG) circuit to control capsular polysaccharide (CAP) production in the commonly used E. coli Nissle 1917 chassis. Expression of the CAP allows the bacteria to temporarily evade the host immune system, resulting in 10-fold larger systemic doses. In addition, the inducible CAP system allowed the bacteria to translocate to distal tumors in CT26 syngeneic, 4T1 orthotopic, and MMTV-PyMT genetically engineered animal models. Bacteria injected into a flank tumor would travel to and colonize multiple distal tumors in the animal upon IPTG induction of capsule expression. Introducing the Clostridium perfringens Theta toxin as a payload under the control of the luxI promoter allowed for delivery induced by acyl-homoserine lactone autoinducers. This approach resulted in an attenuation of tumor growth for both injected and distal tumors.²⁴

Engineered Devices for Microbial Containment

Encapsulating Microbes in Hydrogels

Hydrogels have a distinguished potential for acting as a living matrix for cells or bacteria due to their easily adjustable parameters, such as their mechanical properties, porosity, and biodegradability. These parameters control the size of molecules that can permeate the hydrogel and the rate at which the hydrogel is degraded inside the body (Table 1). Another crucial hydrogel property is the potential to respond to various stimuli, including pH, enzymes, metabolites, temperature, or pressure and shear forces. Hydrogels can be designed such that a combination of these stimuli can be used to initiate swelling, shrinking, or drug delivery.

Table 1. Summary of Hydrogel-Based Devices Based on Their Purpose, Hydrogel Composition, and Biodegradability.

	purpose	hydrogel composition	hydrogel biodegradability
Li et al.25	investigate encapsulated bacterial viability and autoinducer diffusion	PDB-alginate-Ca	Alginate-based hydrogels are biodegradable.
	detection of P. aeruginosa autoinducers by sensor bacteria		The rate of degradation increases with a decrease in molecular weight.26
Li et al.30	investigate encapsulated bacterial viability, bacterial escape, and autoinducer diffusion	alginate-methacrylate microbeads	Alginate’s molecular weight can be adjusted by γ-irradiation.26
	impact of ionic and covalent cross-linking on the swelling properties		An increased calcium concentration decreases the degradability rate, due to cross-link formation.27
Courbet et al.31	detect pathological biomarkers in urine and/or serum	polyvinyl alcohol and sodium alginate	Methacrylate allows degradation through an enzymatic process or hydrolysis.28
Tang et al.32	examine cell–cell communications across the hydrogel to sense heavy metals in water samples	chemical containment (core): alginate-Ca	Polyvinyl alcohol degrades to fatty acids, ketones, and alcohols.29
		physical containment (shell): acrylamide, alginate, ammonium persulfate, and methylene bisacrylamide	Shell containment is tough and semipermeable. A slow degradation rate is predicted.
Cheng et al.33	drug delivery system for inflammatory bowel disease and colitis	mesoporous silica sphere: macrocyclic CB[8], TRP-CS, and Azo-HA	Chitosan and HA, naturally derived, have a high biodegradation rate.
Waimin et al.34	non-invasive, target sampling of GI contents and microbiota	acrylic acid and acrylamide	Nonbiodegradable due to its superabsorbent properties
Liu et al.35	to sense chemical stimuli and provide feedback through fluorescence	hydrogel ink on elastomer adhesive: Pluronic F127-DA and Irgacure 2959	The elastomer adhesive is nonbiodegradable, thereby preventing the hydrogel ink from degrading.

Li et al. used a hydrogel strategy to encapsulate an engineered reporter capable of detecting secreted autoinducers. They synthesized a hydrogel by combining 1,4-bi(phenylalanine-diglycol)-benzene (PDB) with an aqueous alginate (alg) solution to form a PDB gelator. Next, CaCl₂ is allowed to diffuse into the hydrogel, resulting in the creation of a PDB-alg-Ca hydrogel with alginate-based cross-links. Encapsulation of a live E. coli reporter strain expressing green fluorescent protein (GFP) under the control of the pLuxR promoter was achieved by adding the bacteria to the alginate prior to Ca²⁺-mediated cross-linking (Figure 1A). The hydrogel porosity allowed trapping of the bacterial reporter strain while permitting diffusion of Pseudomonas aeruginosaN-(3-oxododecanoyl)-l-homoserine lactone autoinducers, which induced GFP expression. The reporter bacteria remained viable inside the hydrogel and even proliferated with a doubling time of 32 ± 7 min.²⁵

Figure 1.

Various hydrogel designs for bacterial encapsulation and therapeutic delivery. (A) Encapsulated reporter bacteria in a hydrogel matrix can sense diffused autoinducers and report via GFP expression.²⁵ (B) The shell–core hydrogel bead design ensures bacterial encapsulation while allowing diffusion of nutrients.³² (C) Smart hydrogel device used in sampling the contents of a particular region in the GI tract for further analyses. The PDMS membrane prevents sample contamination from more distal GI tract locations and seals the capsule after hydrogel swelling.³⁴ (D) Structural composition of a drug delivery hydrogel consisting of a mesoporous silica core with a multilayer polymer (TRP-CS and Azo-HA cross-linked with CB[8]) coating consisting of 9 or 19 layers. This naturally derived hydrogel is synthesized from chitosan, hyaluronic acid, and silica.³³ (E) Engineered bacteria are incorporated into hydrogel ink and 3D printed into a matrix on an adhesive elastomer.³⁵ This figure was created with BioRender.com.

In a separate study by the same group,³⁰ the E. coli reporter strain was loaded into alginate-methacrylate (alg-MA) hydrogel beads. These alg-MA hydrogel microbeads were produced by an electrostatic extrusion method resulting in a controllable alg-MA bead diameter between 100 and 300 μm. To increase the permeability and decrease the swelling ratio properties of these alg-MA beads, the authors employed a combination of ultraviolet photo-cross-linking and ionic cross-linking. The alg-MA hydrogel bead technology was integrated into a dipstick biosensor for P. aeruginosa detection. Upon inclusion of a different engineered E. coli reporter strain, this could be more widely applied to detection of other bacteria or small molecules.³⁰ This alg-MA hydrogel was not tested for leakage of the encapsulated bacteria, which would be the logical next step prior to validating this methodology for in vivo experiments.

Courbet et al. developed hydrogel-based whole cell biosensors (bactosensors) to detect pathological biomarkers in human samples.³¹E. coli and Bacillus subtilis are used as chassis and genetically engineered by incorporating Boolean integrase logic gates and genetic switches. To generate the eventual bactosensors, these engineered bacteria are encapsulated in hydrogel beads consisting of polyvinyl alcohol and sodium alginate, cross-linked with a boric acid solution and calcium chloride. This type of cross-linking increased the robustness of the hydrogel beads, enhancing their susceptibility for clinical use. As a proof of concept experiment, the bactosensor was used to detect glycosuria in the urine of diabetic patients, demonstrating its potential for diagnosis of renal complications.³¹ The concepts of hydrogel-based living biosensors were further developed by Tang et al. for detecting heavy metals in water samples.³² Their core–shell hydrogel bead design consists of a multilayer encapsulation technique that ensures physical and chemical containment. An alginate-base hydrogel core provides chemical containment of an engineered E. coli reporter (Figure 1B). Cores are formed as bead-like droplets and subsequently solidified by immersion in CaCl₂. Physical containment of the core is provided by a tough shell composed of acrylamide, alginate, ammonium persulfate, and N,N-methylenebisacrylamide. Various engineered E. coli reporter strains were introduced into the biosensor core to evaluate responses to external stimuli. In a proof of concept application, acyl-homoserine lactone (AHL) sender and receiver strains were each encapsulated in separate hydrogel beads and co-incubated in growth medium. The bacteria in the “sender beads” produce AHLs in response to an external stimulus with anhydrotetracycline. The AHLs in turn are sensed by the reporters encapsulated in the “receiving beads”, resulting in an increased level of GFP production that is directly proportional to the amount of anhydrotetracycline provided to the senders. In an additional application, biosensors containing engineered E. coli reporter strains were placed in a tea bag and immersed in river water to detect the presence of heavy metals. These core–shell hydrogel beads were capable of detecting toxic levels of cadmium and other heavy metals and demonstrated robust physical containment of bacteria within the hydrogel core.³² Core–shell hydrogel bead technology shows great potential for biocontainment of microbial biosensors because it has an added benefit over previously discussed hydrogels by providing both physical and chemical containment. In future research, the robustness of core–shell hydrogels throughout the GI tract will require further evaluation. Low-pH conditions in the stomach might cause degradation, and the gels should maintain their molecular exchange capabilities without excessive swelling. In general, stomach acidity can negatively affect various types of hydrogels, depending on the properties of the polymers (e.g., hydrophobicity and hydrophilicity), fabrication techniques, and added stabilizers (e.g., chitosan and alginate). Hence, hydrogel properties need to be tailored to prevent degradation and leaking of the contents and to avoid excessive hydrogel swelling, which causes losses of its molecular exchange capabilities. Hsu et al. provide an excellent example of tailoring the stability of polymeric coatings to withstand the stomach’s acidic conditions. The goal of their study was to use engineered phage λ for targeted modification of gut bacteria. To enable oral delivery, these phages are encapsulated in alginate beads and the beads’ resistance to acidic conditions was achieved by coating in pectin and polyethylenimine in a layer-by-layer methodology. The authors demonstrated a direct relationship between the number of polymer layers and the encapsulation’s resistance to the stomach environment.¹²

Hydrogels that do not contain previously encapsulated bacteria can also be used as a diagnostic tool for microbiome applications. Waimin et al. developed a smart hydrogel-based capsule for non-invasive, localized sampling of the GI microbiota.³⁴ A methacrylate housing is made using a stereolithographic 3D printing process. Next, this housing is filled with a desiccated, superabsorbent hydrogel made of acrylic acid and acrylamide that is topped off with a flexible PDMS plug that will seal the capsule postsampling. The upper part of the housing consists of a twistable cap with a biodegradable coating that can be adjusted to dissolve at a certain pH (Figure 1C). As the capsule is ingested, a biodegradable coating dissolves in the target location in the GI tract, leading to absorption of fluids along with microorganisms in the target location. The hydrogel’s hydrophilic polymer network can absorb large amounts of liquid while maintaining its structure. Due to postsampling hydrogel swelling, the PDMS plug is mechanically pushed toward the entrance of the housing, thereby sealing it off from further absorption and preventing sample contamination. The moist environment of the hydrogel postsampling, in combination with the gas permeable properties of the PDMS plug, maintains the viability of the sampled microbiota until the capsule is passed in the stool.³⁴ Current methods for sampling the GI tract are biased toward the large intestinal community found in feces or are invasive and have a significant chance of sample contamination. The smart capsule provides an inexpensive and non-invasive method for the sampling of GI microbiota that can also be targeted to analyze the small intestinal community. Localized sampling of the small intestine will enable more in-depth analyses of the role of the microbiota in nutrient metabolism and conditions like inflammatory bowel diseases and small intestinal bacterial overgrowth.

Cheng et al. designed a hydrogel-based drug delivery system for targeting the colonic gut microbiota consisting of mesoporous silica nanoparticles.³³ These nanoparticles are well-suited for drug delivery purposes because of their biocompatibility, adjustable morphology, large pore volume and surface area, and chemical stability. The particles consist of a mesoporous silica core with multilayer polymer coatings (Figure 1D). The monodispersity of the spherical nanoparticles was validated by transmission electron microscopy prior to polymer coating. The coatings consist of the macrocyclic molecule cucurbit[8]uril (CB[8]), which forms an enclosed cavity and acts as a host to neutral or cationic species. CB[8] is cross-linked with tryptophan-functionalized chitosan (TRP-CS) and azobenzene-functionalized hyaluronic acid (Azo-HA). Due to the opposite charges of TRP-CS and HA, a layer-by-layer fabrication technique can be used to produce the coating’s multilayer thin films. The strategy for targeted delivery to the large intestine relies on gut microbial azoreductases that will reduce the azobenzene in the multilayer polymers, thereby disassembling the coatings and releasing the drug preloaded from the mesoporous silica core. In addition to the delivery of the preloaded drug, the tryptophan incorporated into the coatings is intended to induce aryl hydrocarbon receptor signaling, resulting in alleviation of colitis and bowel inflammation.³³ The stability of the nanoparticles upon exposure to the GI tract’s various pH conditions and reducing agents was first validated in vitro. Silica spheres with either 9 or 19 polymer layers were tested (Figure 1D), and the thicker coatings resulted in slower drug release. The azo-reducing agent dithionite was used to mimic the azoreductase enzyme in vitro, and drug release rates were measured by using a Cy5 fluorescent dye loaded in the core. For in vivo testing, Cy5-loaded nanoparticles were administered to mice via oral gavage. The nanoparticles passed the upper GI tract with minimal release and subsequently caused accumulation of the payload in the colon. The efficacy of this delivery system was demonstrated in a mouse model of dextran sodium sulfate colitis, using nanoparticles preloaded with hydrocortisone.³³ In future applications in colonic delivery, probiotic or engineered bacteria could be encapsulated in the silica core. This will require prior experiments to ensure bacterial viability inside the core and might require adjustment to the polymer encapsulation procedure or the number of layers used.

3D Printing of Living Materials

3D printing of living materials is becoming more popular for biomedical applications. Liu et al. integrated engineered E. coli into hydrogel matrices termed “hydrogel ink”.³⁵ The hydrogel ink is then placed in a 3D printing machine with optimized conditions for pressure and concentration of the copolymer surfactant Pluronic F127 (Figure 1E). This technique can be applied to 3D print living tattoos, which can be adhered to the skin and contain genetically engineered bacteria to sense parameters like pH and temperature.³⁵ Engineered living materials have great potential for future applications considering the broad array of bacterial sensing capabilities for environmental stimuli, biomolecules, or touch.³⁶ These stick-on living materials could be developed as topical drug delivery adhesive bandages, which can deliver repeated doses before being peeled off or being biodegraded.

Electronic Devices

Mimee et al. engineered an electronic capsule-like system containing engineered bacteria to detect GI diseases.³⁷ As a proof of concept, this device was developed to detect the presence of blood in the GI tract. An E. coli Nissle 1917 chassis was equipped with a heme-sensing circuit and luciferase reporter and tested in a model in which indomethacin was used to induce intestinal bleeding in mice. The stool of animals injected with the blood-sensing bacteria showed a significant increase in luminescence in comparison to stool from control animals. This synthetic strain was next incorporated into an electronic capsule capable of wireless data transfer to a computer or smartphone. The electronic components are encapsulated with parylene and epoxy, and the outer shell of the capsule was made of PDMS. The capsule’s diagnostic ability can be customized by loading different synthetic E. coli sensor strains into its cell carrier reservoir. In addition to heme sensing, the capsule was adapted for detection of other biomarkers such as thiosulfate and AHL and tested in an in vivo porcine model. A limitation of this device is that longer-term exposure to acidic environments of the GI tract could cause degradation to the biosensor and thereby impair its functionality. Use of electronic devices in humans poses additional risks because of their nonbiodegradable nature and potential interference with pacemakers or other implantable electronics.³⁸ The electronic capsule developed by Mimee et al. is an excellent example of how the combination of engineered bacteria with electronics provides a non-invasive, real-time data collection procedure that can be customized to the detection of various molecules in difficult parts of the GI tract.³⁷

Conclusion and Future Perspectives

This Perspective discusses recent synthetic biology applications of bacteria as cancer diagnostics and therapeutics and their incorporation into engineered hydrogel-based or electronically enabled devices. While synthetic microbes show promise for systemic administration and tumor targeting,^20,21,24 this approach introduces additional complications such as the presence of endotoxin in the bloodstream. Bacteremia is not considered to be a healthy host status, and more effort is needed to better characterize and contain synthetic organisms in the bloodstream or incorporate built-in kill switches for removal of the synthetic chassis to avoid septic shock or other adverse effects. Even in naturally colonized body sites such as the GI tract, containment and safety are important concerns when using engineered bacteria. Encapsulation in hydrogel or electronic capsules has been demonstrated to retain bacterial viability and functionality throughout the GI tract. Hence, hydrogels can be used for in vivo sampling of gut microbiota and metabolites, sensing inflammation and various autoinducers in the environment, and targeted delivery of therapeutics. Future advances in this field are expected to involve the integration of various functions into the engineered bacteria and the optimization of hydrogel properties for localized activation and eventual disposal. For example, encapsulated bacteria may be engineered to sense colon tumors, triggering the release of a therapeutic molecule in the target area. After treatment, a built-in kill switch can remove the synthetic organism and activate biodegradation of the hydrogel.

Systemic delivery strategies used for MDSC interaction and device encapsulation with containment in the gut are at present not compatible. We foresee that both approaches could become more integrated in future applications, after additional mechanistic characterization of the involved molecular mechanisms. For example, instead of the systemic delivery of live bugs, only their immune reactive compounds (such as protein or lipid antigens or bacterial membrane vesicles) could be delivered from the contained devices in the GI tract.³⁹ There are currently no preclinical studies that have investigated cancer treatment using synthetic bacterial therapeutics encapsulated in hydrogels. Our Perspective highlights the possibilities for integrating both technologies with the aim of retaining the therapeutic efficacy of the engineered microbes while ensuring their safety and containment via encapsulation. Because several types of biocompatible hydrogels have been developed, the field would benefit from standardized analysis and in vivo tests to compare small molecule delivery and bacterial containment properties across different systems. As studies are emerging on the safety and efficiency of hydrogel devices in animal models, we envision that they will likely be introduced in clinical trials in the next five years.

Author Contributions

Writing of the original draft and figures: L.H.N. Writing, review, and interpretation: L.H.N. and J.C.

J.C. is supported by a National Institutes of Health grant (R01 AI153173), a Research Grant from the Prevent Cancer Foundation (PCF2019-JC), an American Cancer Society Institutional Research Grant (IRG-16-186-21), a Jump Start Award (CA043703) from the Case Comprehensive Cancer Center, and seed funding from the Cleveland Clinic Foundation.

The authors declare the following competing financial interest(s): J.C. is a Scientific Advisor for Seed Health, Inc.