Delivering living medicines with biomaterials.

In this Review, we cover three main types of living medicines: mammalian cells, microbial cells, and viruses. Table 1 provides representative examples of approved living medicines in the clinic and those in late-stage clinical trials. Mammalian cellular products are composed of live human cells, typically in the length scale of 10 – 30μm, and have been used in clinical settings for several decades, ranging from regenerative medicine to cancer immunotherapy. Early efforts focused on non-engineered native cells that were expanded ex vivo and implanted into diseased tissues, such as fibroblast grafts to regenerate damaged skin tissues 24 . In the last decade, breakthroughs in cancer immunotherapy have driven rapid clinical progress of genetically engineered cell therapies, highlighted by the first FDA approval of chimeric antigen receptor (CAR) T cells in 2017 25 . Microbes, including bacteria and yeast, are typically 0.5 – 5 μm and 3 – 10 μm, respectively, rapidly replicate, and are often resilient in harsh environments. These properties enable them to access physiological niches that are often difficult to reach with mammalian cell therapy, including mucosal, dermal, and cancer sites. Although the clinical progress of microbial therapy has lagged behind that of mammalian cells, a few products have been approved, such as Bacillus Calmette-Guérin therapy for bladder cancer 26 . Many probiotics are generally recognized as safe and adapted as over-the-counter supplements 27 . Viruses are the smallest of the three living medicine types, with dimensions down to the nanometer scale, and they rely on host cells for replication, rendering them metabolically inactive on their own. Owing to their ability to transfer genetic materials, an increasing number of viral gene therapies have been approved over the last 10 years 28 . Their selective cytotoxicity has also been leveraged as oncolytic viral therapy, with the first product approved in 2015 (T-VEC) 29 . Although no phage therapies have yet been approved by the FDA, it has historically been used in parts of Eastern Europe 30 . Given the focus of this Review on the delivery of living medicines, readers seeking a more comprehensive overview of living medicines are encouraged to refer to other reviews on the subject 2 , 3 , 31 – 34 .

Here, we highlight four emerging directions that we anticipate will address ongoing challenges and open new opportunities in living medicine research. New biomaterials innovations will likely be required to carry these opportunities forward. Whereas the impact of molecular signals, such as cytokines, on the phenotype of cells has been well-studied, physical environment also impacts the efficacy of living therapeutic cells. For instance, a 2023 study showed long-term imprinting of T-cell phenotype based on the viscoelasticity of the culture matrix 228 . To this end, one emerging direction to control the effectiveness of living medicines is to culture cells in matrices with specific mechanical properties that could instruct the cell’s phenotype ( Fig. 6a ). Furthermore, using such biomaterials as delivery vectors could stabilize the cell’s function once inside the body. 229 – 231 . As genetically engineered cells are increasingly deployed as living medicines, it is becoming clear that maintaining the functions of synthetic gene circuits is a real challenge, owing to mutations and loss of activity. In the hematopoietic stem cell therapy Zynteglo for beta thalassemia, for example, the long-term oncogenic potential from genetic insertion remains a concern 232 , 233 . In bacteria, one study found that 70% of tumor-homing bacteria lose plasmid-encoded therapeutic functions after four weeks, likely owning to the metabolic burden from additional expressions 234 . A potential approach to mitigate the effects of genetic instability is to contain engineered living medicines within biomaterials that prevent their leakage and propagation even in the event of genetic mutations. Many living medicines are motile and might disseminate to off-target sites. This Review has highlighted technologies that prevent dissemination by locally placing and controlling the release of cells. However, since many areas of the human body are mechanically active, such as muscles, a new focus is to develop materials that can withstand dynamic tissue environments to prevent the leakage of encapsulated cells. ( Fig. 6b ). For instance, bacteria encapsulated in tough hydrogels were more successfully contained under mechanical loading conditions than those in brittle hydrogels 235 , 236 . Moreover, tough adhesives have been developed that not only bond robustly to diverse biological tissues, but even remain adhered during tissue movements, such as the beating of a heart 237 . Such innovations in materials may improve local retention of living medicine in the various dynamic mechanical environments of the body. Although the dissemination of living medicines could lead to toxicity, their active motility can also be harnessed for therapeutic delivery. Living medicines themselves can be used to carry other therapeutic agents, enhancing the specificity, efficacy, and safety of conventional medicines. For example, phage-guided systems delivered anticancer payloads to colorectal tumors by targeting intratumoral bacteria 238 . In a separate study, tumor-homing bacteria were used to deliver gold nanoparticles for photothermal therapy 239 . Cells are metabolically active and produce various biomolecules and signals. Emerging studies have shown that biomaterials can act as mediators to perturb cellular metabolism 240 ( Fig. 6c ). Molecules polymerized inside living cells were found to slow down the cells’ metabolic activities 241 , and hydrogels formed bioorthogonally within mammalian cells induced reversible metabolic dormancy in the cells 242 . One potential application of using biomaterials to modify metabolic activity is to finely control cellular activity in vivo . For example, the metabolism of a cell could be made dormant during delivery to maintain cellular viability, and reactivated upon reaching the disease site. Another application is to control cellular replication. A synthetic polymer network assembled inside bacterial cells rendered them incapable of division while preserving essential functions such as protein synthesis and motility 240 . This strategy prevents unwanted bacterial replication, thereby improving the safety profile of living therapeutics. The cellular metabolism of a living medicine can also be used therapeutically by coupling metabolic activity to that of a surface-attached material. For example, nanoparticles attached to tumor-homing bacteria utilized the bacteria’s metabolism to generate antitumoral reactive oxygen species 243 . Further, biomaterials have been designed to convert metabolic signals into therapeutic or diagnostic outputs. A device encapsulating blood-responsive bioluminescent bacteria was developed to convert the luminescent signal to wireless signals 244 , 245 . In another study, encapsulation of Staphylococcus epidermidis within conductive hydrogels enabled control over bacterial activity via electrical signals 246 . These examples illustrate the potential of new biomaterial–living medicine feedback systems. However, an often-overlooked aspect is the coupling of living medicines with surface-attached materials over the long term. For instance, as cells replicate, the ratio of surface-attached materials and cells decreases, potentially weakening the synergistic interactions between cellular metabolism and the surrounding materials. However, this dilution effect could also impose a natural limit on therapeutic output, serving as a built-in safety mechanism. Advances in synthetic biology have allowed living medicines to synthesize a wide range of biological materials, some of which have not been accessed by other means ( Fig. 6d ). For example, mammalian cells and bacteria have been engineered to produce gas vesicles, protein-bound nanostructures that can be harnessed for biomedical imaging and therapy 247 , 248 . Another study engineered mammalian cells to secrete RNA within nanoparticles, which can be used to transfect target cellular populations 249 . Bacteria have been engineered to produce adhesive mussel foot proteins to repair blood leaks 250 , and fibrous matrices to promote gut mucosal healing 137 . Using living cells to synthesize biological materials inside the body could offer a new way to deliver materials in a sustained manner. Another potential benefit of cellular production of materials is dynamic control over the material that enhances the function of living medicines. For example, bacteria have been engineered to produce polysaccharides on-demand 251 – 253 , including capsular polysaccharide, a surface coating that protects microbes from immune factors. By dynamically regulating the expression of this polysaccharide on the bacteria, the immune response could be dampened during systemic delivery 251 . Because the glycocalyx on cell surfaces alters their interaction with immune systems 254 , a similar approach based on glycocalyx production could be utilized for cellular therapy. In the future, temporal control over biological materials synthesis could allow living therapeutics to adapt to dynamic physiological environments. Beyond the biosynthesis of materials, emerging work has engineered living medicines to produce other new living medicines. A 2023 study engineered bacteria to produce oncolytic viruses 255 . This design allowed the efficient delivery of oncolytic viruses to the core of solid tumors and exemplifies a new way to combine multiple living medicines synergistically. The expanding design landscape of cellular materials could enable the next generation of delivery technologies for living medicine.

Living medicines face multiple hurdles during delivery ( Fig. 2a – c ). In this section, we summarize the key challenges in their delivery: targeting, or the medicine’s spatial and therapeutic specificity to cells or tissues of interest; effective dosage, the release and activity of the delivered medicine over time; and stability, the susceptibility of the medicine to physiological clearance and degradation mechanisms during delivery. We highlight commonalities and distinctions in these challenges for different types of living medicines. Mammalian cell therapies are typically delivered via local or systemic routes. A major challenge of both routes is the limited infiltration of these large cells into target disease sites. Intravenous delivery has been a commonly explored mechanism to administer mesenchymal stem cells in clinical trials, but the vast majority of these transferred cells become entrapped in the lung vasculature 35 . CAR-T cells are also delivered via intravenous infusion following ex vivo genetic engineering and expansion. Although intravenous CAR-T therapy has been successful for liquid tumor treatments 36 , efficacy against solid tumors remains low, likely in part owing to the lack of infiltration through the dense extracellular matrix of tumors 37 – 39 . Local delivery of cells, such as in skin grafts, can overcome the need for systemic administration by providing direct access for therapeutic cells to the disease sites. However, many disease sites are not suitable for local delivery owing to inaccessibility or the potential leakage of cells from the applied region. For example, local implantation is used for cartilage graft products via surgical procedures 40 . However, ensuring that the cell graft then remains in place can be a challenge, as these cells can leak from the implantation site if not properly secured. Another challenge of delivering mammalian cells is the specificity of therapy. In the case of CAR-T cells, the targeting epitope can be shared with healthy cells, leading to on-target off-tumor toxicity. For instance, CD19-targeted CAR-T cells, such as Kymriah and Yescarta, can attack both malignant B cells and normal B cells, resulting in B cell aplasia 41 – 44 . Tumors can also downregulate the expression of the targeted antigen, leading to escape and insensitivity to CAR-T cells 45 . This antigen loss has been observed in patients treated with CD19-targeted CAR-T cells, where the cancer cells eventually become resistant to therapy 46 – 49 . Microbial cells, especially bacteria, are small and highly motile, enabling them to infiltrate many target areas. Various delivery routes have been adopted for these cells, including oral, topical, intravenous, and local administration. Many microbial applications are based on the microbes’ inherent ability to colonize specific sites. For example, members of the oral, gut, and vaginal microbiota have been utilized to supplant their natural niches. Although these inherent targeting mechanisms make it possible to deliver certain microbes through systemic administration, there is still a risk that microbes will disseminate from the target site. For example, several types of bacteria have been demonstrated to colonize the immunosuppressive solid tumor environment, but potential bacteria dissemination leading to secondary infections remains a major concern 50 – 52 . Clinical trials involving Salmonella typhimurium VNP20009 and Listeria monocytogenes cancer vaccines were terminated owing to cases of septicemia 53 – 55 . To overcome systemic exposure of bacteria, microbial cells can be locally administered instead. A notable example is Bacillus Calmette-Guérin cancer therapy, in which the bacteria are applied directly to the bladder region through intravesical instillation. The route of delivery for viral vectors is highly dependent on the application. Oncolytic viruses, such as T-VEC, are often directly injected intratumorally to achieve high local concentrations. Although the small size of virus is advantageous to tissue infiltration, challenges still remain owing to their uptake by the mammalian cells. For instance, adenovirus therapies have shown limited efficacy in penetrating the dense mass of solid tumors, restricting their spread within the tumor 56 . Intravenous administration is common for gene delivery treating systemic diseases, allowing vectors to reach multiple body sites, as in the case of Zolgensma for spinal muscular atrophy 57 . Though viruses generally have high specificity owing to their cell entry mechanisms, off-target gene editing may occur in the case of shared viral binding domains 58 . For phage therapy, topical and oral delivery routes are commonly used for wound and enteric infections, respectively. Clinical reports on the targeting challenges of phages are limited, but experimental studies have documented issues with phage retention at the target site. The viral particles in general are also prone to being washed out in dermal environments owing to perspiration, and in mucosal environments owing to inherent mucus clearance mechanisms 59 . Even when therapeutic cells are targeted to disease sites, their dosage and functionality must be tightly controlled. Adoptive cell therapies, including CAR-T cells and tumor-infiltrating lymphocytes, can rapidly proliferate and kill large amounts of tumor cells, leading to cytokine release syndrome and neurotoxicity. The aforementioned Kymriah and Yescarta have been associated with this syndrome, necessitating interventions such as tocilizumab to mitigate these effects 60 – 62 . Another challenge is the sustainability of cell phenotype. For example, maintaining the immunosuppressive phenotype of therapeutic T-reg cells has been difficult owing to their plasticity in response to the surrounding cytokine environment, such as IL-2, and the potential loss of FOXP3 expression in an inflammatory disease environment 63 – 65 . CAR-T cells can also become exhausted after continuous stimulation, a phenotype characterized by reduced cytokine production and cytolytic activity. Additionally, many tumor microenvironments are immunosuppressive and can deactivate the immunogenic phenotypes of adoptive cell therapies 66 , 67 . Microbial cells replicate rapidly compared to mammalian cells, complicating dosage control. Cases of oral probiotic overdosing have been reported, highlighting the challenge of maintaining appropriate bacterial levels. For example, overdosing of Culturelle, which contains Lactobacillus rhamnosus , resulted in sepsis 68 . Uncontrolled microbial growth can also elicit strong immune responses. The inherent immunogenicity of bacteria can limit their use for injectable applications owing to the high risk of cytokine storms. In clinical trials, Clostridium novyi -NT therapy showed promising efficacy, but some patients experienced severe inflammatory reactions 69 . The dependence of viral replication on host cells makes dosage control difficult. Rapid and overwhelming replication can lead to cytokine release syndrome, exacerbated by host cell death and innate viral immunogenicity. Gendicine, an adenoviral vector used in China for head and neck squamous cell carcinoma, has been associated with notable immune-mediated adverse effects 70 . These adverse effects highlight the challenge of balancing effective viral replication and oncolysis with managing host immune responses to prevent severe toxicities. The rapid replication of viral vectors can also complicate the prediction and control of viral spread within the body, necessitating close monitoring and potential adjustments to dosing regimens 58 . Maintaining the viability of delivered mammalian cells has been a long-standing challenge. A particular issue is immune rejection of cellular products, which depends on the source of cells. Although autologous cell sources derived from the patient’s own cells pose minimal risk of immune rejection, they are typically slow and costly to manufacture 71 , 72 . Allogeneic cells can be readily sourced from donors but are typically rejected by host immune systems, requiring the use of immunosuppressants 73 , 74 . For example, a pancreatic islet cell implant approved in 2023, Lantidra, has been shown to be effective against type 1 diabetes. However, because the host immune system can prevent efficient grafting and long-term performance, immunosuppressive drugs are needed for an extended period 75 . Transplanted cells require a few criteria to remain viable. New blood vessels must typically be formed to supply oxygen and nutrients to the transplanted cells; inadequate angiogenesis can lead to hypoxia, then cell death and clearance of the cells. A proper local microenvironment that provides necessary cues for survival, such as adhesion molecules and cytokines, is also important. This challenge is exemplified in regenerative medicine applications that use transplanted cells to repair damaged tissues. For example, Carticel is an implanted chondrocyte product used to repair cartilage defects, but ensuring that the chondrocytes remain viable in the inflammatory environment of osteoarthritic joints remains a hurdle 76 . In skin grafts, transferred cells have been well-documented to rapidly disappear after delivery 77 , 78 . Although microbial cells are typically more tolerant to harsh environments compared to mammalian cells, several physiological conditions can still compromise their viability. In gastrointestinal applications, many microbial products struggle to survive gastric acid during passage and compete with existing microbiota, leading to low colonization in the intestines 79 . A Phase III trial involving intestinal colonization of SYNB1618, an engineered probiotic therapy for managing phenylketonuria 80 , 81 , reported significant variability between patients and was terminated in 2024. Additionally, competition from the native microbiome for space and nutrients can hinder the grafting and colonization of exogenously added mircobes 82 , as shown by the probiotic yeast Saccharomyces boulardii , which does not permanently colonize the gut 83 . Other anatomical locations also present distinct challenges. For example, the native microbiome in the vaginal tract, dominated by acid-resistant Lactobacillus species, creates a low pH environment challenging for exogenous bacteria 84 . The oral environment is spatially diverse with well-defined local microbiome communities 85 , 86 , which could compete with exogenously delivered microbes. The respiratory tract is characterized by unique immune components such as secretory IgA, which neutralizes pathogens 87 . For delivering engineered bacteria to skin, a significant challenge is maintaining their stability and preventing rapid desiccation in the external environment 88 . In systemic administration of microbial cells, such as in cancer therapy, bacteria are rapidly cleared by innate immune responses, including phagocytes and complement systems 52 . Viral therapies are prone to rapid clearance from the body through filtration and neutralization. Given their small size, viruses are susceptible to being swiftly filtered from the bloodstream by the kidneys. Adeno-associated viruses (AAVs) used in gene therapies, such as Hemgenix for hemophilia B, have been shown to have short circulation half-lifes 89 . Some viral vectors, such as lentiviruses, can be cleared by the liver through the mononuclear phagocyte system, further reducing their bioavailability 90 . To counteract the rapid clearance of virus from the bloodstream, high viral doses are often administered, which can increase the risk of immune responses and toxicity. Importantly, pre-existing immunity can neutralize viral vectors and reduce therapeutic effectiveness. This last mechanism has been a long-standing challenge in the field, especially for commonly used vectors such as AAVs, to which a substantial portion of the population—approximately 30–60%—has pre-existing immunity from prior exposure to natural AAVs 91 . Even for uncommon viral vectors, patients can develop antiviral antibodies after the first dosage, reducing efficacy 92 .

Various physiological systems work to clear and degrade exogenously administered drugs from the body, and biomaterials have long been used to protect conventional drugs from these processes. Orally administered drugs can be degraded by gastric acids. Once drugs are in circulation, they are cleared by filtration, enzymatic degradation, and innate immune mechanisms such as phagocytosis. Drugs can also be neutralized via the adaptive immune systems (such as the binding of antibodies), which impacts long-term drug effectiveness. For living medicines that must remain both viable and stable to function properly, these clearance and neutralization processes pose additional considerations. The viability of microbial cells is typically compromised by gastric acids during oral administration and innate immune responses during systemic administration. Viral vectors are susceptible to neutralization upon repeated exposure, and mammalian cells can be rejected by the host immune system via recognition of non-self antigens. Specific biomaterial designs are required to protect cells from these harsh physiological environments and the immune system to ensure stability of the living medicine. The mechanism of circulatory clearance depends on the size of the drug. Typically, small molecule drugs (< ~60 kDa) are metabolized by enzymes such as cytochrome P450 and are rapidly cleared via renal filtration 10 , 179 . Conjugating drugs to larger biomaterials can increase their hydrodynamic volume, preventing small drugs from being excreted. Genexol-PM, which is paclitaxel formulated in polymeric micelles, employs this strategy: the micelles, owing to their size, prevent rapid renal clearance and enhance paclitaxel’s circulation time and accumulation in tumor tissues 180 . Large protein-based drugs are susceptible to multiple degradation and immune mechanisms. The reticuloendothelial system takes up circulating proteins via phagocytic cells. A popular strategy to circumvent reticuloendothelial system clearance is to attach drugs to components that are naturally recycled in the body, such as albumin and immunoglobulin 181 ( Fig. 5a ). Levemir, a long-acting insulin analog, binds to albumin in the bloodstream, prolonging its action by preventing rapid degradation and clearance 182 . Albiglutide, another albumin-fused drug, extends the half-life of GLP-1 to achieve once-weekly dosing for diabetes management 183 . Enbrel, a fusion protein, combines a therapeutic protein with the Fc region of an immunoglobulin. In addition to being cleared by the reticuloendothelial system, protein drugs are also prone to protease-mediated degradation 184 . Conjugating the drug to polymers that act as a steric barrier, such as PEG chains, can slow the drug’s degradation. Adagen, used to treat severe combined immunodeficiency disease associated with adenosine deaminase deficiency, became in 1990 the first PEGylated drug to be approved 185 . Since then, many PEGylated drugs have been developed, including Pegasys for hepatitis, which extends the circulatory half-life of interferon alpha-2a from several hours to 80 hours 186 , 187 . Beyond circulatory clearance mechanisms, drugs taken orally also face the harsh acidic environment of the stomach and the variable pH conditions throughout the gastrointestinal tract. Various biomaterials have been developed to protect drugs during this transit. Glumetza, for example, employs a polymer-based matrix to protect its insulin payload from the acidic environment of the stomach. This approach enhances delivery of the insulin to the intestine, where it can be absorbed 188 . With repeated or sustained exposure to a drug, the body can produce anti-drug antibodies, which quickly neutralize the drug when it enters circulation. To prevent these antibodies from neutralizing the drug, materials have been conjugated to the drug to provide immune-evasive properties. The aforementioned PEGylation and albumin conjugation strategies provide a steric shield that can reduce immunogenicity 189 ( Fig. 5b ). However, more recent evidence has suggested that treatment with PEGylated drugs can increase anti-PEG antibodies 190 . Alternative materials have shown promise to provide immune-evasive properties to the drug. For example, CD47-conjugated nanoparticles, which signal to the immune system to avoid phagocytosis 191 , can be leveraged to evade immune clearance Zwitterionic materials, which possess both positive and negative charges, can also reduce protein adsorption and immune recognition 192 . Biomaterials have improved microbial viability upon systemic administration. As one example, cloaking of the yeast Saccharomyces cerevisiae with a zwitterionic material reduced phagocytosis and prolonged circulation times in mice 193 ( Fig. 5c ). Bacteria have also been placed inside of mammalian cellular materials to protect them during systemic administration. Tumor-homing bacteria were wrapped with mammalian cell membranes to protect them from phagocytosis, leading to more efficient accumulation of bacteria in breast tumors in a murine model 194 . Oncolytic Salmonella were loaded into macrophages by initial phagocytosis via coculture, followed by snap-freezing using liquid nitrogen, which reduced bacterial clearance in circulation and increased their accumulation in tumors after systemic administration 195 . Hydrogel encapsulation strategies have kept orally administered bacteria viable as they pass through the harsh gastric environment. For example, encapsulating probiotics layer-by-layer with alginate and chitosan improved the probiotics’ survival through acidic stomach environments and bile salts 178 . Other material coatings, such as lipids 196 and silk fibroin 197 , or formulations, such as ingestible macrogels, have also been utilized to improve oral delivery of bacteria 198 . Viral clearance in circulation is mediated by renal filtration and neutralization by antibodies. As shown for conventional drugs, conjugation of PEG and albumin to viral vectors also improves viral stability by reducing renal clearance and immunogenicity 199 ( Fig. 5d ). A conjugation method to modify AAV with PEG was successfully identified and applied to transgene delivery in the lungs 200 . Similarly, adenovirus was engineered to express albumin-binding domains, associating the virus to serum albumin upon injection 201 . This approach extended viral half-life in mice, even in the presence of pre-existing antibodies. An emerging strategy for viral delivery is hitchhiking, or attaching the virus to cells that naturally possess circulatory stability. Anchoring AAVs to the surface of red blood cells, for example, enhanced viral stability and delivery to the lungs to enable successful transgene expression upon re-administration, even following prior AAV exposure 202 . Oncolytic viruses have been placed inside inactivated cancer cells to protect virus from degradation. Oncolytic viruses have also been attached on the surface of T cells to improve their stability during systemic administration, thereby enhancing their delivery to tumors in mice 203 – 205 . For mammalian cells, immune rejection of allogeneic cells has been a long-standing challenge that necessitates the use of immunosuppressive agents in the clinic. A key innovation was the encapsulation of cells with a biomaterial that protects them from immune rejection while allowing the release of therapeutic agents secreted by the cells ( Fig. 5e ). An early design from 1980 showed that alginate hydrogels could act as a protective barrier for pancreatic islet cells while allowing the secretion of insulin 206 . Over the years, other hydrogel designs have been developed to encapsulate pancreatic cells and genetically engineered cytokine-secreting cells 207 – 210 , and this approach has been applied to various delivery scenarios 211 ranging from injectable hydrogels 212 – 214 to retrievable implants 215 . Microencapsulation approaches that allow more efficient nutrient and oxygen diffusion to the cells within have been another notable advance ( Fig. 5f ); encasing single mesenchymal stem cells in thin layers of alginate microgels enhanced cellular survival and function 216 , 217 . Other means to increase the viability of encapsulated cells have included the incorporation of cytokines, extracellular matrix components, and adhesion molecules within the hydrogel matrix. For instance, collagen matrices significantly increased the survival of transplanted cardiomyoblasts by promoting vessel ingrowth into the porous graft 218 . In another example, a silica layer provided an adherent surface for mesenchymal stem cells, resulting in improved survival in suspension. Biomaterials can also be conjugated with peptides that promote the survival of encapsulated cells, as in the attachment of pro-survival peptides to a collagen matrix to enhance stem cell viability post-transplantation 219 . Although cell encapsulation enables immune protection in the short term, foreign body reactions often occur after prolonged implantation. Historically, fibrotic encapsulation—where fibrous tissue formed around a graft restricts the exchange of oxygen and nutrients, thereby limiting graft survival and functionality—has been a considerable barrier to the clinical translation of cell-based therapies. In particular, islet transplantation for diabetes treatment has been held back by the adverse effects of this process. In clinical trials involving alginate-encapsulated islets, such fibrotic overgrowth progressively compromised the diffusion of essential molecules, leading patients to experience a loss of glycemic control 220 , 221 . This example highlights the ongoing need for biomaterials that can more effectively modulate the host response. Along these lines, several biomaterials have been developed to mitigate foreign body reactions. A combinatorial chemical modification approach identified three promising triazole-containing alginate variants that significantly inhibit foreign body reactions by preventing macrophage recognition and fibrous deposition 222 ( Fig. 5g ). This high-throughput approach was expanded to in vivo screening, where cellularly barcoded hydrogels were used to identify lead candidates 223 . Various physical parameters such as charge, pore size, and particle geometry have also been shown to impact the foreign body response 224 – 226 . For instance, the size of pores in scaffolds were found to differentially promote a pro-inflammatory phenotype of macrophages and alter the extent of the response 227 .

Living medicines, including mammalian cells, microbial cells, and viruses, are emerging as a novel class of therapeutic agents ( Fig. 1a ). Sophisticated genetic engineering tools can program these cells to perform complex tasks for therapeutic applications 1 – 3 ( Fig. 1b ). For instance, cells can dynamically respond to environmental stimuli 4 , possess metabolic capabilities for controlled production of drugs and biomolecules 5 , and be engineered for active motility and translocation to target specific sites within the body 6 . Replicating cells can adapt to changing physiological conditions and evolve over time 7 . These unique properties have fueled clinical translations of living medicines, resulting in an increasing number of approved products 8 – 10 . Despite the rapidly expanding scope of living medicines, it remains difficult to deliver them specifically to desired tissues and organs. The inherent complexity of a living cell presents new challenges for delivery. The cells’ viability and activity must be maintained to ensure therapeutic effect. Furthermore, because cells can alter their state over time in response to environmental signals, their behavior and function are difficult to control. The body’s immune responses also pose obstacles, as many exogenously manipulated cells are recognized as foreign by the immune system 11 . Improved strategies for delivering living medicines are expected to potentiate and expand their clinical applications. Biomaterials have revolutionized the delivery of various therapeutic modalities ranging from small molecules to biologics 12 – 14 ( Fig. 1c ). By engineering the chemical and physical properties of biomaterials, one can modulate pharmacokinetic and pharmacodynamic properties of the payloads. For instance, conjugation of drugs to carrier materials has been widely utilized to enhance the drugs’ solubility, stability, and half-life, resulting in more effective treatments with reduced dosing frequency 13 , 15 . The development of controlled-release systems has improved therapeutic effectiveness and patient compliance 16 , 17 . This Review discusses the promise of biomaterials in advancing the delivery of living medicines. We first introduce the main types of living medicines developed to date and identify their delivery challenges. Next, we introduce notable biomaterials strategies used for conventional medicines and then highlight progress in adapting these strategies to living medicines. In addition to drawing parallels between the two medicines, we describe specific considerations for living medicines that necessitate different designs of their biomaterial delivery systems. Lastly, we chart remaining challenges in living medicine delivery and highlight converging innovations in material science and synthetic biology that have the potential to create the next generation of living medicine. Although this Review does not focus on the material design, Box 1 outlines basic design principles of biomaterials; other reviews cover these principles in more depth 18 – 23 .