Dan Wu ab, Jiaqi Lei a, Zhankui Zhang b, Feihe Huang *cd, Marija Buljan e and Guocan Yu *af aKey Laboratory of Bioorganic Phosphorus Chemistry & Chemical Biology, Department of Chemistry, Tsinghua University, Beijing 100084, P. R. China. E-mail: [email protected] bCollege of Materials Science and Engineering, Zhejiang University of Technology Hangzhou, 310014, P. R. China cStoddart Institute of Molecular Science, Department of Chemistry, Zhejiang University, Hangzhou 310027, P. R. China. E-mail: [email protected] dZJU-Hangzhou Global Scientific and Technological Innovation Center, Hangzhou, 311215, P. R. China eEmpa, Swiss Federal Laboratories for Materials Science and Technology, 9014 St. Gallen, Switzerland fSchool of Medicine, Tsinghua University, Beijing 100084, P. R. China

First published on 29th March 2023

Vital biomacromolecules, such as RNA, DNA, polysaccharides and proteins, are synthesized inside cells via the polymerization of small biomolecules to support and multiply life. The study of polymerization reactions in living organisms is an emerging field in which the high diversity and efficiency of chemistry as well as the flexibility and ingeniousness of physiological environment are incisively and vividly embodied. Efforts have been made to design and develop in situ intra/extracellular polymerization reactions. Many important research areas, including cell surface engineering, biocompatible polymerization, cell behavior regulation, living cell imaging, targeted bacteriostasis and precise tumor therapy, have witnessed the elegant demeanour of polymerization reactions in living organisms. In this review, recent advances in polymerization in living organisms are summarized and presented according to different polymerization methods. The inspiration from biomacromolecule synthesis in nature highlights the feasibility and uniqueness of triggering living polymerization for cell-based biological applications. A series of examples of polymerization reactions in living organisms are discussed, along with their designs, mechanisms of action, and corresponding applications. The current challenges and prospects in this lifeful field are also proposed.

Dan Wu

Dan Wu was born in Henan, China, in 1988. She received her PhD degree from Zhejiang University in 2016 under the supervision of Prof. Guping Tang. She is currently a lecturer at the college of materials science and engineering at Zhejiang University of Technology. Her current research interests are focused on the construction of functional biomaterials.

Jiaqi Lei

Jiaqi Lei is a PhD candidate at the Department of Chemistry, Tsinghua University, under the supervision of Dr Guocan Yu. He received his BS degree from Northwestern Polytechnical University majoring in polymer science and engineering. Now, he is interested in understanding and exploring the pathway of immunology (especially the role of saccharides), the mode of cancer growing, and the protein interaction with small molecules or ligands. He is devoted to making potency drugs using supramolecular chemistry to contribute to human health.

Zhankui Zhang

Zhankui Zhang received his BS degree from Henan University of Technology in 2016. Now he is a postgraduate at the College of Materials Science and Engineering, Zhejiang University of Technology under the supervision of Prof. Wu Dan. His research interests focus on functional biomaterials.

Feihe Huang

Feihe Huang is Changjiang Scholar Chair Professor at Zhejiang University. His current research is focused on supramolecular chemistry. Awards and honours he has received include Chinese Chemical Society AkzoNobel Chemical Sciences Award, The Cram Lehn Pedersen Prize in Supramolecular Chemistry, the Royal Society of Chemistry Polymer Chemistry Lectureship award, Changjiang Scholar, The Royal Society Newton Advanced Fellowship award, and the Germany Bruno Werdelmann Lectureship Award. He sits/sat on the Advisory Boards of J Am Chem Soc, Chem Soc Rev, Chem Commun, Acta Chim Sinica, Macromolecules, ACS Macro Lett, and Polym Chem. He is an Editorial Board Member of Mater Chem Front.

Marija Buljan

Marija Buljan is a group leader at the Swiss Federal Laboratories for Materials Science and Technology and an associated group leader at the Swiss Institute of Bioinformatics. She obtained her PhD degree in computational biology from the University of Cambridge and after that she worked as a researcher at the MRC Laboratory of Molecular Biology in Cambridge, UK, German Cancer Research Institute in Heidelberg, Germany and ETH Zurich in Switzerland. Her research interests include computational and systems biology, proteomics and immunoengineering.

Guocan Yu

Guocan Yu received his PhD degree from Zhejiang University in 2015 under the supervision of Prof. Feihe Huang. He spent five years as a postdoctoral fellow in Dr Xiaoyuan (Shawn) Chen's research group at National Institutes of Health (NIH) from 2015 to 2020. He started his independent research career at the Department of Chemistry at Tsinghua University in the late 2020. His research interests are focused on the development of smart delivery systems, supramolecular theranostics and supramolecular strategies for immune modulation.

Recently, ELMs which make use of the dynamic metabolic profile of living organisms to modulate and assemble complicated synthetic structures, have shown promising potential in cell-based biological applications, such as cell-on-a-chips, cell-based sensors, cell therapy, biocatalysis and biomotors.24–30 Two polymerization strategies, namely, covalent and noncovalent (supramolecular) polymerization, are generally used for constructing ELMs inside and on the surface of living cells. For cell surface engineering, polymer-based cell encapsulation and direct polymerization on the cell surface are two very active directions. To date, numerous polymer-based strategies, such as layer-by-layer, cell-in-microgel and cell-in-shell, have been explored to encapsulate cells.31–34 However, these strategies are barely biocompatible and the formed polymer shells are always thick and stiff, thus impairing cell capabilities, to a certain extent.35 Although some strategies, such as biomolecular assembly, fast kinetic gelation and mussel-inspired chemistry, have been exploited to enrich the sorts of cell envelopes; however the problems of uncontrolled polymerization and low encapsulation efficiency remain.36–38

Currently, direct and in situ polymerization approaches have been developed to reduce polymer thickness and improve the coating efficiency with the aid of new chemical polymeric reactions, such as biosynthetic reactions and visible light-triggered graft polymerization.39,40 Chemical or physical polymer grafting reactions (also known as the “grafting-to” strategy) are preferred methods for cell surface polymerization.41–46 However, these approaches are often limited by low grafting ratios, inefficient cellular mass transport, signal transduction, and the excessive use of reactive polymers.47–50 The “grafting-from” strategy, where polymers directly grow from the surfaces of cells, can enhance the polymer grafting rate, control polymer chain length and acquire functional block copolymers, thus greatly improving the efficiency of cell surface engineering without affecting cell functions.51 Radical polymerization reactions are the most popular reactions for in situ polymerization on the cell surface. For example, photo-induced electron transfer-reversible addition-fragmentation chain-transfer (PET-RAFT) polymerization and atom-transfer radical polymerization (ATRP) have been extensively used to regulate the metabolic activity of cells. In addition, Pd-mediated cross-coupling reactions and cascade bioorthogonal reactions have attracted considerable scholarly attention in recent years.

For intracellular engineering, although some inorganic nanomaterials,52–55 such as fluorescent quantum dots,56,57 have been biosynthesized in cells, the explorations of in situ intracellular polymerization are relatively few due to some difficulties: (1) Monomers are hardly accumulated in living cells. (2) Polymerization is slow, which requires a long exposure time for cells to poison polymerization conditions. (3) Intracellular metabolic biomolecules, such as amino acids and dissolved oxygen, easily interfere with the polymerization reactions. (4) External stimuli-initiated polymerization reactions cannot be efficiently conducted in vivo due to the poor tissue penetration capability of external stimuli, such as light. To date, a couple of polymerization reactions have been realized in living cells; however, much work needs to be carried out to fully utilize bio-polymerization.

Biomimetic polymer synthesis has been rapidly developed in the past five years, with remarkable advancements in methods to establish sophisticated sequence-controlled materials. These novel synthetic polymers provide unprecedented possibilities in different fields ranging from biomedical sensing to therapeutic delivery. This review comprehensively summarizes the progress in intra/extracellular polymerization for overcoming the present medical problems. By taking advantage of different biological or cellular characteristics, versatile polymerization processes, such as free radical polymerization, bioorthogonal polymerization, oxidative polymerization and supramolecular polymerization, are ingeniously designed. The sophisticated designs of polymerization reactions and the therapeutic mechanisms are discussed in detail. We also reveal the current limitations of polymerization in living organisms, and propose possible solutions and prospective future development directions, providing new thoughts for the research community.

Although some radical polymerization reactions have been driven in open containers under ambient condition, the reactions usually prefer anaerobic environment and are restricted to the limited catalysts and monomers.63–66 Inspired by the oxygen consumption capability of glucose oxidase (GOx) in recent living polymerization reactions, Keitz et al. utilized microbial aerobic respiration to initiate metal-catalyzed radical polymerization under aerobic conditions. The facultative Shewanella oneidensis (S. oneidensis, MR-1) regulated radical polymerization reactions under aerobic conditions through a two-step reaction: (1) depleting the dissolved oxygen by aerobic respiration and (2) conducting extracellular electron transfer (EET) to the metal catalysts to initiate the ATRP (Fig. 2b, I).67 Under anaerobic and aerobic conditions, S. oneidensis was able to initiate free radical polymerization without pre-removal of oxygen. Of note, the polymerization efficiency was closely connected with the extracellular electron transfer protein-mediated anaerobic metabolism of S. oneidensis (Fig. 2b, II), and the polymerization rate remained high when a series of metal catalysts and monomers even at a low concentration were utilized. Importantly, polymerization reactions withstood the repeated oxygen perfusions (Fig. 2b, III) and could be activated by the lyophilized and recycled cells (Fig. 2b, IV). This novel hybrid strategy of fermentative metabolism, aerobic respiration and the facultative properties of S. oneidensis, will surely provide new ideas for artificial polymerization reactions and improve the biosynthesis abilities of organisms. Although this work promotes the success of bacteria-based radical polymerization reactions in aerobic environments, the cytotoxic Cu-based catalyst is a concern. Obviously, radical polymerization reactions with a high biosafety under aerobic conditions are the priority for future investigations.

Although hybrid structures composed of synthetic polymers and living cells have exhibited brilliant cell-based applications, it is still highly challenging to perform polymerization on the surface of individual living cells as the polymerization conditions are always fatal to cells. Taking surface-initiated ATRP (SI-ATRP), for example, transition-metal catalysts, anaerobic conditions and organic solvents are culprits for the cytotoxicity of SI-ATRP. However, these cytotoxic threats can be circumvented, to a certain degree, by adopting ATRP using activators regenerated by electron transfer (ARGET ATRP). Reducing agents (e.g., ascorbic acid) used for ARGET ATRP can reduce Cu(II) to Cu(I), and thus a litter metal catalyst is needed. Moreover, ARGET ATRP is conducted in water under atmospheric conditions, making ARGET ATRP more compatible than traditional ATRP. Nevertheless, cytotoxic radical species generated by ARGET ATRP will attack living cells, possibly leading to a low viability during polymerization reactions. Choi et al. utilized the radical-depleting ability of polydopamine (PD) to fabricate cell–polymer hybrid materials based on SI-ARGET ATRP. PD-based layers not only blocked radicals generated during SI-ARGET ATRP, but also acted as ATRP macroinitiators (Pdi) for SI-ATRP (Fig. 3a, I).68 After SI-ARGET ATRP on Pdi-primed yeast cells, negligible decrease in the viability of the yeast was monitored. However, an obvious cytotoxicity was observed in the control group (Fig. 3a, II and III), suggesting that the Pdi layer effectively reduced the peroxidation damage of the cell membrane. Scanning electron microscopy (SEM) revealed that the surface of the polymer-coated yeasts was rougher than that of native or Pdi-primed yeast cells (Fig. 3a, IV), confirming the generation of polymers on the cell surfaces. With the cooperation of fluorescent co-monomers and bioorthogonal chemistry, a polymer layer on the surface of yeast was obviously visualized (Fig. 3a, V). E. coli has abundant α-D-mannose-binding protein on the cilium, which can trigger the formation of cell aggregates when mixed with yeasts. However, no cell clusters formed in the group with co-culturing yeast and E. coli (Fig. 3a, VI), indicating that the compact polymer layer of yeasts efficiently blocked their agglutination. In addition, cell-division activity was suppressed and the lag phase of yeasts was prolonged after polymer grafting (Fig. 3a, VII), indicating that the cytocompatible and highly dense polymers were formed on the surface of individual living cells. Considering the diversity of synthetic polymers, it is believed that grafting polymers on cell surfaces will generate a number of delicate cellular hybrids for various biomedical and biotechnological applications.

Traditional antibiotics have been highly effective in combating microbiological contamination, but the appearance of resistant strains is a growing concern. New methods that do not arouse drug-resistant strains are required.69 The method of isolating bacteria strains away from the infective sites is attractive from a theranostic perspective.70,71 Nevertheless, the specific binding of targeted bacterial strains is difficult, and the present practice needs costly “cold-chain” biological reagents, such as aptamers and antibodies, severely limiting their popularization in large areas. Inspired by the role of ECM containing complex macromolecules synthesized by bacteria to support cell communities, synthesis of ECM mimics through naturally metabolic pathways would be greatly advantageous for targeted binding bacteria. Alexander et al. exploited the redox systems of bacteria for the polymerization of acrylic monomers via a Cu(I)-mediated ATRP on cell surfaces (Fig. 3b, I).49 In such a manner, the bacteria selected the matched binding monomers with them, and in situ grew disparate polymers on cell surfaces. The monomer composition of bacteria-instructed polymers was affected by the bacterial species, and the resulting polymers could specifically bind to the bacteria which provided the template. Bacterial redox chemistry was also engaged to “click” fluorescent markers onto bacteria-directed polymers on the surface of various clinically isolated strains (Fig. 3b, II), permitting rapid and facile binding and imaging of pathogens (Fig. 3b, III). This cell-mediated chemistry utilizes ready-made materials that do not need cold-chain storage, which is more adaptable for routine laboratory settings. Meanwhile, many possible monomers have been applied in bacteria-supported ATRP reactions, indicating that further refining of binding specificity would be possible via judicious choice and diversification of polymer components. Therefore, it is believed that this method would, if developed well, create new platforms for bacteria-specific theranostic applications.

Being able to realize spatio-temporal controllability and maintain the chain-end fidelity and the dispersity of radical polymerizations, PET-RAFT polymerization reactions have received great attention since 2014. Due to their great potential in biomedicine and “green” chemistry fields, photosensitizers with wideband and near-infrared (NIR) absorption are superior for PET-RAFT polymerization.80–82 Although some advanced photosensitizers have been developed to initiate RAFT polymerization reactions under a series of light sources, including blue, green, red light and natural sunlight, the requirement of non-aqueous solvent systems has not been violated.83 Qiao et al. presented the first example using the self-assembly of carboxylated porphyrin (SA-TCPP) as a photosensitizer for NIR-mediated controlled radical polymerization reactions in an aqueous solution (Fig. 5a).84 SA-TCPP showed a broad absorbance spectrum in the range of 300–950 nm, and the micro-fibre-shaped morphology extended the lifetime of the excited electron-hole pairs, thus significantly improving the efficiency of polymerization. The polymerization rate ranked from high to low in the order of white, blue, green, red and NIR light, which was in accordance with the absorbance of SA-TCPP. Importantly, PET-RAFT polymerization also successfully proceeded in mammalian fibroblast cells at the cost of cell viability, suggesting that in future, programmable multi-block polymerization reactions may be conducted in mechanized microliter injection systems after the problem of cell viability is solved. These biocompatible NIR-mediated controlled radical polymerizations are good pioneers for future in vivo biomedical polymerizations and “green” polymerizations.

Grafting synthetic polymers to cell surfaces offers potential to regulate cellular functions and expand the structural repertoire of cells. However, the conventional “grafting-to” strategies involving chemical or physical polymer conjugation are often challenged by low grafting efficiency and the abuse of polymers. In contrast, “grafting-from” strategies in which the designed polymers directly grow from cell surfaces, display multiple advantages, such as enhanced grafting efficacy, controlled chain length and high adaptivity and specificity. Hawker et al. reported a cytocompatible “grafting-from” strategy to engineer living yeast and mammalian cells via PET-RAFT (Fig. 5b).40 For yeast cells, a two-step method was first developed to conjugate chain-transfer agents (CTAs) onto cell surfaces. After surface-initiated polymerization, a strong purple fluorescence (Alexa Fluor 647) was observed around the cell surface of yeast cells Y2, suggesting that PET-RAFT successfully proceeded on the cell surface. Furthermore, there was distinct fluorescence arising from fluorescein diacetate in the cytoplasm, indicating the high viability of polymer-modified yeast cells. Without addition of tannic acid (TA) which can cross-link poly(ethylene glycol) (PEG) polymers via hydrogen bonding, few cell aggregates were observed, whereas considerable aggregated cells were monitored in the presence of TA, clearly demonstrating that cytocompatible polymer coatings exhibited feasible capability to control cell assembly and aggregation. Similar to yeast cells, an obvious Alexa Fluor 647 fluorescence emerged from the Jurkat cell surface after surface polymerization, confirming that polymers also efficiently grew from mammalian cell surfaces. In addition, high viability was also acquired for the polymer-anchored Jurkat cells, indicating that the active metabolism and intact cell membranes were maintained after polymerization. In comparison with conventional “grafting-to” strategies, these “grafting-from” approaches not only greatly improve the grafting efficiency but also actively manipulate cellular phenotypes. Based on the diversity of functional polymers in this “grafting-from” strategy, a wide range of applications may be possible. For example, grafting high density glycopolymers to rewire signalling pathways and spatiotemporally controlling polymer distributions on cell surface to manipulate cell–cell interactions, etc.

Recently, construction of ELMs by harnessing the reducing power of bacteria has been developed, but the specific metal bio-reduction pathways and the cytotoxicity of metal catalysts are always involved.85 Qiao et al. reported a metal-free biosynthesis strategy to initiate a “living” radical polymerization reaction, in which bacterial electron flux was captured to generate abiotic radical species (Fig. 5c).86 Two bacterial species including S. Typhimurium and E. coli were utilized to reduce aryl diazonium salt (4-bromobenzenediazonium tetrafluoroborate, 4-BT) into aryl radicals near microbial membranes. The generated aryl radicals initiated “living” RAFT polymerization to synthesize vinyl polymers using methacrylate monomers. Multiple excellent polymerization features including high fidelity of chain-transfer agents, low dispersity and first-order chain growth were observed in the regulation of polymerization conditions, suggesting that this bacteria-facilitated polymerization was well controlled. In addition, it was demonstrated that the reduced contents in bacteria, such as metal ions with a low redox state and glutathione, facilitated transmembrane electron transfer and speeded up the reduction of 4-BT. Meanwhile, active metabolism was necessary for the sustained generation of radicals, guaranteeing the following “living” RAFT polymerization. The universal mechanism of redox equilibrium and the validity of the reducing environment of active bacterial cultures provides significant potential to extend this “living” radical polymerization to other microorganisms and replaceable redox-activated chemistry. Based on the performance of this “living” radical polymerization, hijacking microbial redox force to synthesize artificial polymers instead of reliance on the conventional metal-catalyzed polymerization reactions provides a new method for constructing ELMs at the abiotic/biotic interface.

Various artificial coating techniques have been engaged to endow organisms with additional functionalities and protection from hostile conditions.93–95 However, a full coating or uniform coverage may inevitably block functional receptors on cell membranes, impairing their communication with surrounding cells, therefore partial or asymmetric coating is a more feasible surface modification approach.96–98 Yang et al. developed a visible light-initiated graft polymerization for symmetric or asymmetric surface coating on individual yeast cells (Fig. 6b, I).99 The degree of coating could be tuned by simply regulating irradiation time and density and high viability of yeasts was invariably retained (Fig. 6b, II). As expected, the yeast cells with polymer shells showed a delayed division and enhanced lysis resistance. In addition, obvious incomplete ellipse emerged on the PEGDA-decorated yeast cells (Janus-poly(PEGDA)-cells) after adding poly(sodium acrylate) (PAAS), suggesting PAAS was selectively introduced in the PEGDA-patched area of cell surfaces (Janus-PAAS-cells) (Fig. 6b, III). It was worth noting that the asymmetrically decorated yeast cells exhibited a strong self-propelling capability in the urea-containing fermentation broth by anchoring urease on the surface of Janus-PAAS-cells (Fig. 6b, IV and V). Considering the structural diversity of living graft polymerization and the loading functionality of the hydrogel layer, this flexible cell coating strategy provides great potential in multiple fields, such as cell therapy, cell-based advanced drug delivery, tissue engineering and biocatalysis.

Although fluorescent quantum dots100 and inorganic nanomaterials have been successfully synthesized in human cells and microorganisms, radical polymerization in living organisms has rarely been reported and we still lack a deep understanding that how these endogenous polymers modulate cellular functions or track cells. Bradley et al. reported a light-mediated radical polymerization reaction for in situ polymer synthesis in cells using a series of monomers and a biocompatible initiator (Fig. 7a, I).101 The monomer conversion of these polymerization reactions remained high in different mediua, such as ∼50% in PBS and cell lysate, and ∼68% inside cells, suggesting that the complex cellular microenvironment rarely interfered with this radical polymerization. Furthermore, high viability was acquired for different cell lines even on the 7th day post-polymerization, indicating the good biocompatibility of these polymerization reactions. There was no obvious difference in the G0/G1 and M phase for the cells, but a delayed S phase was observed (Fig. 7a, II), demonstrating that this intracellular polymerization affected the cell cycle. In addition, cell motility (Fig. 7a, III) and actin polymerization (Fig. 7a, IV) were also regulated by the cellular polymerization. Direct evidence for intracellular polymerization was obtained by using 4-styrenesulfonate (NaSS) as a monomer. Even after five passages, strong fluorescence signals were still observed in the polymerized cells (Fig. 7a, V), providing a possibility for long-term tracking of the implanted cells. After incubating HeLa cells with ferrocenylmethyl methacrylate (FMMA) monomers, spherical nanoparticles with a diameter of 50–70 nm were observed both in the cytoplasm and nucleus (Fig. 7a, VI), suggesting that FMMA not only crossed the cellular membrane, but also penetrated into the nucleus. Using light-mediated polymerization reactions to synthesize new-to-nature polymers in living cells paves a new way for intracellular engineering where the in situ formed functional polymers can be applied in controlling cytoskeleton functions, fluorescence imaging and cellular motility. The applicability and potential power of this approach permits us to add intracellular synthetic macromolecules into a toolkit of cytocompatible chemistries and further intracellular polymerization exploration promises to be a great adventure.

Peptide amphiphile (PA)-constructed biomaterials have been broadly used in diverse fields of biomedicine, such as biosensors, drug delivery and tissue engineering.102–107 Importantly, the biofunctionalities of PAs are closely related with their supramolecular assembling behaviors, and therefore a deep understanding of the assembly mechanism of supramolecular architectures is crucial for the applications of peptide biomaterials.108,109 Jiang et al. reported two kinds of diacetylene-containing peptide amphiphiles (DPAs) (zwitterionic PA-OH and cationic PA-NH2), which exhibited entirely different polymerization behaviors in living cells and aqueous solution (Fig. 7b, I).110 For example, PA-NH2 did not undergo polymerization in a series of aqueous solutions owing to the irregular arrangement of diacetylene chains but finished in situ polymerization in response to the microenvironment of tumor cells. In contrast, PA-OH accomplished the polymerization in aqueous solution, but not in cells (Fig. 7b, II). Because of the higher positive charge and intracellualr polymerization of PA-NH2, PA-NH2 achieved a lower half inhibitory concentration (IC50) than that of PA-OH (Fig. 7b, III), serving as a potential anticancer agent. This work not only provides indications on the relationship between the charge characters of PAs and their polymerization behaviors, but also demonstrates that diacetylene-containing PAs with the polymerizing capability in cells can serve as a highly effective anticancer agent.

For cell replacement therapy, exogenous cells are encapsulated in hydrogels to prevent the recognition of immune biomacromolecules but allow the free entry and exit of small molecular weight nutrients and water.212 Therefore, hydrogel coatings with size selectivity are the key for cell transplantation therapy. Berron et al. developed a free-radical polymerization reaction to coat ultrathin hydrogel films on mammalian cell surfaces, realizing the selective transport of small molecular nutrients (Fig. 7c, I).50 Photoinitiators with a low concentration were first grafted on the cell surface, and thus the subsequent restricted chain growth induced a thin polymer coating (Fig. 7c, II). These thin hydrogel films displayed similar size-selective behaviors to the corresponding bulk hydrogel materials, and the lower thickness led to a higher flux of species with a low molecular weight (Fig. 7c, III–V). Hence, in the post-transplantation cellular environment, it is expected that the delivery of vital small nutrients across the gelified cell membrane is slightly affected while host immune response is efficiently prevented. Because cells are engineered in the solution, the cell suspension can be injected into target tissues in a less invasive manner than the conventional cross-linked cell scaffolds with a large bulk, potentially shortening the recovery time and decreasing the frequency of cell therapy.

Owing to the similar mechanical properties with many substances in living systems, hydrogels have been applied to encapsulate and stabilize many biological entities, such as oligonucleotides, proteins and living cells.112–115 However, precisely controlled release of biological objects is a crucial challenge in these applications. Although some acid-degradable hydrogels have been recently reported by utilizing copper(I)-catalyzed azide–alkyne cycloaddition (CuAAC)116 and free radical cross-linking,117 sensitive biomolecules are hardly encapsulated by these hydrogels, attributed to the cytotoxicity of metal ions and free radicals.118 Combining the droplet-based microfluidics and bioorthogonal strain-promoted azide–alkyne cycloaddition (SPAAC), Haag et al. fabricated a series of pH-responsive microgels for living cells (Fig. 8b).119 Cytocompatible poly(ethylene glycol)-dicyclooctyne (PEG-DIC) and dendritic poly(glycerol azide) (dPG) were utilized as bioinert macromonomers. Azide groups were covalently connected to dPG via changing the substituted acid-sensitive benzacetal linkers, realizing accurate control of release kinetics in the pH range of 4.5–7.4. Integrating the bioorthogonal encapsulation with the pH-controlled release, an on-demand release of encapsulated cells was achieved without impact on cell activity and viability. Thus, these microgel particles can be applied for short-term cell encapsulation, permitting the cells to be manipulated and studied during encapsulation and eventually be isolated and collected via the decomposition of microgel scaffolds.

Seiffert et al. prepared hydrogel particles with micrometer-size and cell-laden ability through a combination of droplet-based microfluidics (technically) and bioorthogonal bioinert polymers (chemically) (Fig. 8c).120 The gelation of microgels was realized through the thiol-ene click reactions between dithiolated PEG and acrylated hyperbranched PG without any initiator. The microgel properties were systematically investigated through simply varying the molecular weights and concentrations of macromonomers, to deeply understand the impacts to the proliferation and viability of these encapsulated yeast and mammalian cells. Further optimization of cell-laden microgel structures could be achieved via integrating droplet-based microfluidic technologies for single cell encapsulation followed by the fluorescence-activated classification to separate living cell-containing microgels from dead cell-containing microgels.

Staphylococcus aureus (S. aureus) infection is one of the principal reasons for numerous severe epidemic diseases including bacteriaemia, osteomyelitis and pneumonia.121 Early and rapid detection of S. aureus infection can be treated directly, to reduce the abuse of antibiotics and harness long-term beneficial results.122 The diagnostic approaches of S. aureus on tissue biopsies or blood sample analyses are time-consuming, complicated and invasive procedures.123,124 Benefiting from the merits of low cost, easy operation, real-time, non-invasiveness and superb sensitivity, fluorescence imaging represents one of the most hopeful techniques for diagnosis of bacterial infection.125 The “turn-on” probes display a better sensitivity than the traditional “always-on” probes when considering signal-to-noise ratios,126 yet rare “turn-on” fluorescence probes have been exploited for real-time monitoring of S. aureus infection. Liang et al. developed an NIR fluorescence probe Cys(StBu)-EDA-thioketal-Lys(Cy5.5)-CBT (TK-CBT) that assembled into fluorescence-quenched TK-CBT-NPs through the supramolecular polymerization driven by noncovalent interactions, including hydrophobic interaction and π–π stacking (Fig. 9a).127 The disulfide bond of TK-CBT-NPs could be oxidated through the elevated reactive oxygen species (ROS) in S. aureus-tainted macrophages, leading to the disintegration of TK-CBT-NPs, thus turning on the NIR fluorescence for real-time detection of S. aureus infection in cells and animals. It is expected that TK-CBT-like “turn-on” NIR probes may be used to clarify the mechanism of biological pathology in ROS-related theranostics and to diagnose the clinical S. aureus-infected diseases in near future.

Because there is a big difference in the 5-year survival rate of cancer patients at early stages and late stages, early diagnosis of tumors plays an extremely important role in cancer therapy.132 Cathepsin B (CTSB) is overexpressed at early stages of numerous tumors; thus it is considered as a potential biomarker for the early cancer diagnosis.133 Precise monitoring of the CTSB expression will provide useful information for the early detection of tumors.134 With high spatial resolution and prominent tissue-penetrating depth, PA imaging has been regarded as a promising and noninvasive imaging modality to screen early stage tumors.135 Although the fluorescence of most imaging agents is often quenched after aggregation owing to the ACQ effect, their PA signal displays an amusing aggregation-induced enhancement,136 which can combine intracellular self-assembly to provide an “off-on” fluorescence method for detection of CTSB activity in vivo. Liang et al. developed a CTSB-activatable PA probe Val-CitCys(SEt)-Lys(cypate)-CBT (cypate-CBT), which polymerized into cypate nanoparticles (near infrared fluorophore) through a supramolecular way after GSH reduction and CTSB catalysis inside tumor cells (Fig. 9c, I).137 These nanoformulations showed unparalleled advantages for monitoring of CTSB activity with high specificity and sensitivity (Fig. 9c, II). Attributed to the enhanced intracellular accumulation, the cypate-CBT probe exhibited a nearly 5-fold enhancement of PA signals in CTSB-overexpressed tumor cells or solid tumors (Fig. 9c, III), showing a potential for clinical diagnosis of early tumors. It is expected that this intelligent cypate-CBT probe could be used as a PA agent to diagnose CTSB-overexpressing tumors at early stages with a high precision and sensitivity in clinics after reasonable refinement.

In situ coating of polypyrrole on the surface of microorganisms not only improves the efficiency of extracellular electron transfer, but also provides great promise to regulate their biology catalytic activity.141,142 Wang et al. endowed thermally-sensitive Aspergillus oxyzae (Asp cells) with photo-sensitivity via coating a layer of photothermal polypyrrole clothing on their surfaces.143 The production and catalytic activity of α-amylase from Asp cells were obviously enhanced under the irradiation of solar light (Fig. 10b). Attributing to the low cost, biocompatibility and non-contact modulation, this photothermal modification strategy offers a great promise for coating living cell–polymer hybrid structures on microorganism systems to improve their low-temperature environmental adaptation.

Except for extracellular surface modification, formation of nanomaterials inside cells is also a novel approach to adjust cellular functions.144 Yang et al. reported an extra/intracellular dual-modified strategy based on the red blood cell (RBC) template to realize synergistic photothermal-chemotherapy.145 CaCO3 nanoparticles were first generated in RBCs via the mineralization of Ca2+ and CO32− followed by extracellular pyrrole oxidative polymerization (PPy) and folic acid (FA) modification, finally establishing a CaCO3@RBC@PPy-FA structure (Fig. 10c). Attributed to the coordination interactions between –CO/–OH groups of DOX and CaCO3, targeting ability of FA and photothermal effects of PPy, dual-modified RBCs possessed a high drug loading and a targeted and light-controlled drug release behavior, greatly inhibiting the growth of tumor cells combining photothermal therapy and chemotherapy. This intra/extracellular dual-modification of RBCs provides new insights for future clinical cancer therapy based on their synergetic photothermal-chemotherapy.

Living cell-based therapies, such as fecal microbiota transplantation, infusion of hemopoietic stem cells and T-cell therapy, have been successfully applied in the treatment of various intractable diseases, including inflammatory bowel disease, cancers, and congenital defects.146,147 Nevertheless, cells are fragile to hostile environmental stressors, transplantation-conducted metabolic microenvironments and host immune systems, resulting in a low therapeutic efficacy and undesired cell death.148 Therefore, methods simultaneously integrating therapeutic modalities with protective functions are highly desired to frame satisfied cells for cell-based therapy. Liu et al. reported a new cell engineering approach for coating individual living cells with multimodal motifs.149 Dopamine was first deposited on the surface of living cells by a cascade dopamine oxidation and polymerization reaction. Functional molecules such as amino-terminated polyethylene oxide (PEO-NH2), rhodamine B and fluorescein isothiocyanate (FITC) were then co-deposited with dopamine to construct multifunctional coatings on the surface of individual cells, including fungi, bacteria and mammalian cells. The main driving forces for this surface polymerization included π–π stacking, hydrogen bonding, Schiff base reaction and Michael addition (Fig. 11a, I–III). Importantly, no obvious change was detected in bioactivity and viability after coating. Owing to the strong tolerance against bile salts and gastric acid due to the presence of the coating, engineered cells presented an obviously improved retention in the stomach and gut compared with bare bacteria (Fig. 11a, IV). Benefiting from the targeting ability of chitosan to inflamed colonic mucosa, coated cells showed a higher accumulation in sick tissue (Fig. 11a, V) and significantly increased the treatment efficacy (Fig. 11a, VI) of the dextran sulfate sodium-lured murine colitis. Bandaging using a multifunctional and biocompatible coating opens a new window for engineering living cells for treatment and prevention. It is anticipated that this approach can be expanded to the decoration of other cells, especially immune cells, aiming to treat cancer and autoimmune disease.

The pharmaceutical efficacy of anticancer drugs is often hampered by their poor tumor invasiveness and limited accumulation in tumors.150,151 Although a number of excellent nanomedicines have been designed to detain active ingredients in tumor tissues with the aid of enhanced permeability and retention effects,152 the total enrichment is still disappointing. Therefore, methods that are able to disperse antitumor drugs lastingly and homogenously throughout the whole tumor tissue are desirable to maximize their therapeutic efficacy. Liu et al. reported a bacteria-directed spatio-temporally controllable allocation of synergistic therapeutics in tumor tissues to reverse the immunosuppression microenvironment for maximizing antitumor efficacy (Fig. 11b, I).153 By combining genetic engineering and polymerization reaction, bacteria were first remoulded from inside and outside synchronously to express photothermal melanin (BacMel). The immune checkpoint inhibitor (αPD-1) was then deposited onto the surface of BacMel which was pre-coated with polydopamine (B-αPD-1). Because bacteria tended to colonize in intratumoral hypoxia circumstances, both BacMel and αPD-1 could be lastingly and homogenously distributed in excised human (Fig. 11b, II) and intravital mouse tumors, leading to a uniform heating area of tumor tissue. Benefiting from the dual photothermal-stimulated and immune checkpoint inhibitor-primed immune activation effect (Fig. 11b, III), the immunosuppressive tumor microenvironment was synergistically reprogrammed (Fig. 11b, IV). As expected, tumor growth was significantly inhibited, and the survival of mice was prolonged in both subcutaneous and eutopic 4T1-tumor murine models with the aid of BacMel + NIR irradiation. This bacteria-guided collaborative treatment combining photothermal therapy and immunotherapy paves a new avenue for a lasting and homogenous distribution of antitumor therapeutics in solid tumors. Considering the future translation, some underlying challenges like the mode of administration, the frequency and dosage of modified bacteria, the mechanism of bacterial distribution and safety issue, should be evaluated systematically.

Inflammatory bowel diseases (IBDs) are closely connected with elevated ROS and highly disordered gut microbiota.154–156 Although antioxidants have been applied to eliminate ROS from the intestine,157–159 and probiotic-guided methods have been developed to help recover the normal gut microbes,160,161 great limitations are still present in the process of application. Nonspecific biodistribution, rapid clearance and weak ROS-scavenging ability determine the inconsistent efficacy of antioxidants for the treatment of inflammatory disease. Simultaneously, probiotics are greatly sensitive to the unfriendly circumstances in the gastrointestinal (GI) tract, significantly limiting their retention time and viability in the intestine. Hu et al. synthesized a amphiphilic hyaluronic acid-poly(propylene sulfide) (HA-PPS) polymer which self-assembled into HA-PPS nanoparticles (HPN) with ROS-scavenging ability (Fig. 11c, I).162 To improve the delivery of probiotics (E. coli Nissle 1917, EcN) to the colons for enhancing bacteriotherapy, they were encapsulated with norepinephrine (NE), which underwent auto-oxidization and self-polymerization to deposit a poly-NE coat on the surface of probiotics (NE-EcN) to buffer them from hostile environmental assaults and prolong their detention time in the intestinal tract owing to their strong mucoadhesive ability. Based on the colon tissue-targeting capability of probiotics, HPN was conjugated onto the modified probiotic surface (HPN-NE-EcN) and efficiently delivered to the inflamed colons to regulate ROS levels with minimum side effects (Fig. 11c, II). In the dextran sulfate sodium-lured murine colitis model, HPN-NE-EcN displayed sharply improved prophylactic and treatment efficacy. Moreover, the diversity and abundance of gut microbes were boosted after HPN-NE-EcN treatment (Fig. 11c, III), conducive to the remission of IBDs. Because some research studies disclosed that the EcN-included E. coli strain shows the probability of facilitating colon cancer, the biosafety of this EcN strain-based study requires to be carefully assessed before clinical application, and other bacterial strains with a higher safety factor may be considered.

It has been demonstrated that the polymerization-induced self-assembly (PISA) method in living systems is able to regulate cellular functions166,167 and even acts as superior therapeutics.168 Nevertheless, PISA proceeding in a specific organelle has not been implemented. Ryu et al. developed a mitochondria-targeting PISA method to induce the necroptosis of tumor cells (Fig. 12b, I).169 Under the effect of mitochondria targeting groups (triphenylphosphonium, TPP), monomers with two thiol units (Mito-1) highly accumulated in the mitochondria of tumor cells which contained a high level of ROS. The ROS in the mitochondria acted as a chemical fuel to catalyze the disulfide polymerization. Furthermore, this intramitochondrial polymerization induced oxidative stress that further elevated the ROS level, thus auto-catalyzing polymerization (Fig. 12b, II). The formed long polymer chains self-assembled into fibrous structures and penetrated mitochondrial membranes (Fig. 12b, III), disrupting the integrity of mitochondria (Fig. 12b, IV) and eventually inducing cellular necroptosis (Fig. 12b, V). Because mitochondria are crucial organelles responsible for the energetic metabolism, this energy-outage strategy may overcome multidrug resistance in tumor treatment. Based on the synergistic effect of Mito-1 and β-lapachone (Lapa) which generated abundant ROS after the catalysis by the intracellular NQO1 enzyme, the tumor growth of SCC7 (Fig. 12b, VI) and 4T1 tumor-bearing mice was significantly inhibited. Of note, negligible systemic toxicity was detected after the synergistic treatment (Fig. 12b, VII). This in situ polymerization in mitochondria shows huge potential for cancer treatment including multidrug-resistant cancers.

Polymerization on cell surfaces has been widely explored, but polymerization in living cells remains challenging owing to the complex cellular environment. However, from another point of view, the cellular microenvironment is abundant in various biological molecules, full of opportunities to regulate cell activities. For example, the high concentration of ROS in tumor microenvironments, may be a favorable condition to initiate oxidative polymerization in cells. Xu et al. developed a ROS-initiated oxidative polymerization reaction to synthesize functional polymers in an ROS-abundant intracellular microenvironment (Fig. 13a).170 In order to elevate the local concentration of monomers for polymerization, a nanoreservoir containing plentiful organotellurides was constructed based on the coordination interaction between gold and organotelluride (Fig. 13b). Attributed to their ultrasensitive redox responsiveness and coordination properties, organotellurides were oxidized into Te–O polymers by the ROS in tumor cells. The obtained Te–O polymers perturbed the intracellular antioxidant system by inhibiting the expression of selenoproteins. As a consequence, a high concentration of ROS was induced in the tumor microenvironment, which continuously fueled the synthesis of Te–O polymers via a self-amplification mechanism (Fig. 13c). Because the ROS level in normal cells was not enough to activate oxidative polymerization of organotellurides, the self-amplification effect was prevented. Therefore, a selective anticancer activity was achieved by the Te nanoreservoir. With a continuous increase in ROS, the ROS-associated apoptosis pathway in tumor cells was activated; thus Te-O polymers exhibited a high anticancer activity and a low systemic toxicity (Fig. 13d). Compared with previous approaches, this oxidative polymerization was initiated without an external stimulus and highly concentrated monomers. The only initiator of the reaction was the intracellular microenvironment. Based on the advantages, such as easy preparation, strong targeting ability and high efficiency, this oxidative polymerization will create a new sensation for the selective regulation of cell behaviors in vivo.

It is well known that cellular interactions are mainly regulated by the membrane proteins, the conformations and expression of which are dynamically adjusted via intracellular programs.179,180 Therefore, development of membrane protein-simulated architectures that are able to recognize cell-responsive signals in response to environmental stimuli and consequently perform allosteric modulation to control cellular interactions will offer new opportunities for the study of cell-signaling networks. With the advantages of predictable structures, good biocompatibility and high programmability, DNA nanostructures are considered as one of the most prospective membrane protein-mimicking materials.181–183 Tan et al. engineered a cell surface-anchored nanoarchitecture that implemented molecular recognition-activated DNA polymerization to imitate the dynamical behaviors of membrane proteins, realizing the perception of cellular adaptive response to environmental changes (Fig. 14b, I).184 Amphiphilic DNA tetrahedrons (T-cho3) could be stably and efficiently anchored onto cell membranes with a low ratio of detachment and internalization (Fig. 14b, II). Upon sensing the cellular adaptive response to intracellular and extracellular stimulations, such as endogenous and exogenous adenosine triphosphate (ATP), these membrane-anchored DNA nanostructures were activated followed by the tandem polymerization of multiple functional nucleic acids via the hybridization chain reaction (HCR) (Fig. 14b, III and IV). The multivalent effect of HCR not only enhanced cell aggregation efficiency and facilitated the generation of the highly clustered cells (Fig. 14b, V), but also recognized the targeted cells for a precise killing effect (Fig. 14b, VI). Profiting from the fast dynamics of nucleic acid chemistry and the modular assembly, this approach could be broadened to manipulate and mimic diversiform biological events such as cellular communication and adaptation, providing a new prototype for intelligent synthetic biology and customized cell engineering.

Although some polymer-based cell encapsulation approaches have been developed, encapsulation of target living cells into mechanically and biotunable polymer shells is a major challenge in current research.185–187 Li et al. proposed a biocompatible and programmable DNA-oriented polymerization approach (isDOP) to in situ weave DNA oocysts on living cells (Fig. 15a, I).188 By the feat of DNA isothermal duplication and programmed DNA assembly, the surface chemistry, mechanical properties and polymer density can be tailored. Since DNA polymers are molecularly precise assemblies with high homogeneity, they can be precisely addressed by DNA base pairing (G–C and A–T) and DNA-modifying enzymes. Rolling cyclic replication (R1) and branching replication (R2) were employed in weaving DNA polymer networks, seeded by initiation primers (IP) and branch primers (BP), respectively. First, IP initiated R1, which generated the long and periodic DNA polymers (LonDNA) based on the replication template of single-stranded circular DNA (cirDNA). Then, BP initiated R2, which generated another single-stranded DNA polymer (LatDNA). LatDNA cross-linked with LonDNA to induce the assembly of connections based on the replication templates; thus in situ DNA cocoons on the cell surface were fabricated (Fig. 15a, I). When an increase in concentration of IP or BP, the density and thickness of DNA polymers also increased (Fig. 15a, II), whereas the pore size of DNA cocoons decreased, suggesting that the structure of DNA cocoons could be regulated by adjusting the IP or BP concentration. After DNA weaving, over 95.6% MCF-7 cells were encapsulated and remained alive (Fig. 15a, III), indicating that this isDOP was highly efficient and biocompatible. Different fluorescent color-labeled DNA cocoons were specifically catched in different regions of the DNA-stereotyped slide surface which was pretreated with the corresponding capture strands (CSs), indicating that LatDNA and LogDNA were automatically synthesized and assembled according to the sequence codes (Fig. 15a, IV). Most importantly, with the aid of DNA-modifying enzymes which can cleave the designed encoding sequences in DNA cocoons, the DNA-weaved cells were selectively released from the capture zones (Fig. 15a, V). This precise isDOP not only enables to overcome cell encapsulation challenges but also provides flexible mechanisms to regulate physiological and biophysical phenomena at cell interfaces. Aside from DNA weaving technique, other tunable natural materials for the control of rigidity, permeability and stimuli-responsiveness of the coating, and the ability to vest the cells with specific functions, including targeted cell delivery, specified assembly manners and designed cell–cell interactions, should be given more attentions.

The transmembrane redox potentials originating from prokaryote respiration, have been successfully applied for in situ construction of biomedical materials over recent years. Compared with traditional biomedical materials, functional materials fabricated in situ can omit the transporting process and targeting ability, thus greatly promoting their adaptivity and specificity.189–191 Meanwhile, supramolecular polymerization reactions are able to endow the final materials with excellent stimuli-responsiveness and biodegradability, attributed to the reversible and dynamic nature of the noncovalent interactions, which are excellent qualities for biomedical applications. Benefitting from the advantages of the in situ fabrication strategy and supramolecular polymerization reactions, it is expected that supramolecular polymers constructed in situ may become novel diagnostic and therapeutic tools. Xu et al. developed a cucurbit[8]uril (CB[8])-based supramolecular polymerization strategy, in which the bifunctional monomer VDV was equipped with a positively charged 1,4-diazabicyclo[2.2.2]octane unit and two viologen groups at the ends (Fig. 15b, I).192 Under anaerobic conditions, the redox potential of E. coli reduced VDV into cation radicals which then self-assembled into supramolecular polymers (VR-SP) driven by the 1:2 ternary host–guest complexation between CB[8] and viologen cation radicals (Fig. 15b, II and III). Free radicals with one or more unpaired single electrons are characterized by a special open-shell structure, and their band gaps are always narrow. It is expected that organic free radicals can show NIR absorption and are good choices for NIR photosensitizers for photothermal therapy. Upon NIR irradiation, the photothermal effect of VR-SP efficiently killed E. coli, displaying a high antibacterial efficiency (Fig. 15b, IV). Because the positively charged VR-SP owned a strong adhesion ability to the negatively charged bacteria (Fig. 15b, V), the photothermal antibacterial ability of supramolecular polymers was further enhanced by the local enrichment of VR-SP. After killing E. coli, VR-SP were easily oxidized by the dissolved oxygen, thus automatically turning off their antibacterial activity. Importantly, the building blocks were biocompatible and exhibited negligible cytotoxicity against mammalian cells (Fig. 15b, VI), suggesting that VR-SP was a safe photothermal agent in potential applications. Better selectivity to specific microbes and antimicrobial performance would be acquired by integrating bio-responsive units and photothermal motifs in this system. This in situ supramolecular polymerization provides a new idea for the fabrication of novel biomedical materials with high adaptivity and programmability.

Self-assembly is a useful tool to prepare supramolecular materials.199 However, efforts have been mainly concentrated on applying a single-step catalytic reaction or a molecule to initiate the self-assembly (also called supramolecular polymerization) process;200 thus the structures of self-assemblies are relatively plain. Recently, a two-reaction controlled supramolecular strategy has attracted increasing interest and has presented diversified self-assemblies with smart functions.201–203 Yang et al. reported a peptide derivative-based tandem molecular self-assembly which was hierarchically controlled by the synergistic effect of extracellular alkaline phosphatase (ALP) catalysis and intracellular GSH reduction (Fig. 16b, I).204 Because hepatoma cells possess higher concentrations of ALP and GSH than normal cells, the hierarchical supramolecular polymerization of the peptide derivative (18) occurred in HepG2 and QGY7703 cells. The peptide 18 first formed nanoparticles outside the cell membrane and then transformed into nanofibers around karyotheca or inside the cell membrane (Fig. 16b, II). Attributed to this tandem supramolecular mechanism, the peptide derivative 18 exhibited stronger cellular uptake and inhibition effects against liver cancer cells than against normal liver cells (Fig. 16b, III–V). This extracellular and intracellular tandem molecular self-assembly mechanism will push the further development of supramolecular self-assemblies with improved performances in cancer theranostics.

Drug resistance is a major obstacle for ovarian cancer treatment. Among various approaches reversing drug resistance, cisplatin-involved combination chemotherapies are most explored because of the high efficiency of cisplatin (CDDP) in clinical therapeutics.205 However, the 5-year survival rate of ovarian cancer patients has hardly increased over the past decade.206 Thus, innovative cisplatin-based combination therapies are urgently needed. Xu et al. used enzyme-instructed self-assembly to construct intracellular supramolecular assemblies which greatly enhanced the capability of CDDP to reverse the drug-resistance of ovarian cancer cells (Fig. 17a, I).207 The pre-installed ester groups on the small peptide precursors (L-22 and D-22) could be cleaved by carboxylesterase (CES), and they were then transformed into L-23 and D-23, respectively, which eventually self-assembled into nanofibers in cells through supramolecular polymerization. At appropriate concentrations, both precursors were innocuous to cancer cells, but significantly boosted the activity of CDDP against the resistant ovarian cancer cells (Fig. 17a, II). Intracellular nanofibers exhibited transient cytotoxicity which minimized the long-term system toxicity in combination therapy (Fig. 17a, III). This enzyme-instructed self-assembly strategy promises a new method for drug-resistant cancer treatment. Other antitumor platinum drugs such as carboplatin which is regarded as a preferred platinum-based drug may display a better anticancer effect in synergy with this enzyme-instructed supramolecular self-assembly strategy.

Molecular self-assembly has developed into a cogent strategy to construct versatile structures with suitable association and folding in living cells.208,209 Although infusive successes have been acquired in vitro, precise construction of molecular assemblies in living cells remains challenging, because the self-assembly precursors will inevitably experience hydrolysis by the non-specific hydrolytic enzymes with similar activities. Wang et al. reported an easily accessible and efficient multistage strategy to accurately control the formation of supramolecular assembly in live cells (Fig. 17b, I).210O-[bis(dimethylamino)phosphono]tyrosine motifs in the precursor NBD-DWpDY(NMe2)2DFDK(Ac)-NH2 (Pro-24P-NMe) acted as the Trojan horse, which resisted hydrolysis by moving phosphatases in and out of the cells as the enzymatic cleavage occurred selectively in lysosome-like acidic environments. After phagocytosis by Saos-2 cells, the phosphodiamidate groups of Pro-24P-NMe underwent acid-triggered hydrolysis to generate the substrate (24P) by acid phosphatase only inside the lysosome. Acid phosphatase cleaved the phosphate group of 24P and induced the multilevel self-assembly of Pro-24PNMe. The supramolecular polymerization from oligomers to nanofibrous networks was confirmed by CLSM (Fig. 17b, II) and biological electron microscopy (Fig. 17b, III). Replacing the acid-catalyzed unit and the type of enzyme by other functional groups, this precise controllable nanostructure formation strategy is expected to develop more functional higher-order structures.

Various polymerization reactions, including metal/enzyme/photo-catalyzed radical polymerization, reduction/oxidation polymerization, bioorthogonal polymerization, and the latest bacteria-powered radical polymerization, have been successfully implemented extra/intracellularly. Most of the numerous designed cell-mediated polymerization reactions require catalysts, toxic conditions (ultraviolet light, free radicals, and organic solvent), harsh conditions (hypoxic or anaerobic), and complicated pretreatment processes (genetic modification). Thus, developing new strategies depending on simple and biologically benign reaction conditions is extremely urgent.

As a result, diverse supramolecular polymers have been in situ established extra/intracellularly. The reversible and dynamic noncovalent interactions endow supramolecular polymers with stimuli-responsiveness and biodegradability properties, thus reducing toxic side effects on normal cells and avoiding long-term immunotoxicity. Unlike traditional supramolecular structures developed based on weak and single noncovalent interactions, supramolecular polymers are the products of multiple non-covalent forces, making their structure formation and maintenance under dynamic environments relatively feasible. Nevertheless, ELMs still face some challenges: (1) intracellular ELMs are rarely generated in selected cells or a specific organelle; (2) extracellular fabrication of polymeric materials seems easy because most ELMs are initiated on or near the cell membrane, but intracellular biomimetic self-assemblies, which must survive in the complex intracellular microenvironment are rare, thus requiring more efforts to be concentrated on the development of intracellular or even in vivo ELMs; (3) the chemical structures of ELMs are usually systematically characterized in vitro, but their theranostic mechanisms are not well investigated. A rational combination of high therapeutic efficiency and the thorough treatment mechanism can push the rapid expansion of the in situ biomimetic ELMs.

We propose the five most critical issues hindering the development of ELMs and figure out some possible solutions, which will enrich readers' knowledge and facilitate in-depth research.

(1) Improving the polymerization activity of monomer molecules. The stability of free radicals of monomer molecules determines the regulation of monomer activity, which can be considered from the following aspects: resonance effects, charge effects (push–pull effects), and steric effects. For example, methacrylate has more stable free radicals than acrylate due to the steric hindrance effect of methacrylate, and the reaction rate is much faster. Thus, the monomers can be rationally modified with different groups according to the specific circumstances and cells for polymerization.

(2) Enhancing the initiation efficiency. Conventional initiation methods, such as light-initiating or thermal-initiating methods, are not applicable due to the poor penetration of light and heat in vivo (initiation energy). Therefore, to enhance the initiation efficiency of polymerization, the current mainstream method uses peroxides as initiators in the body or photoinitiators with longer absorption wavelengths. Notably, newly proposed electrochemical-regulated radical polymerization and ultrasonically regulated controlled radical polymerization reactions are promising, because the high penetration depth of ultrasound and electrical field is remarkably favorable to initiate the polymerization in vitro and in vivo.

(3) Choosing suitable catalysts. Most catalyst used in polymerization are pristine ions, metal–ligand complexes, or bulky clusters, biosafety of which is always considered. In addition, the catalytic capability can be easy diminished by the intracellular components, especially for the sulfur-containing compounds. The incorporation of active catalysts into nanoparticles might resolve the issue. For example, the chelator ligands can be grafted or physically encapsulated by polymeric or inorganic nanoformulations to protect the active catalytic sites during delivery. The catalytic capability is activated at the sites of interest by de-shielding the matrix materials in response to the intracellular microenvironment, thus optimizing the polymerization and preventing unwanted cytotoxicity.

(1) The first order: organ targeting. It can be achieved by manipulating the physical and chemical properties of delivery systems, including size, charge, components, amphiphilicity and stability. For example, selective organ-targeting nanoparticles can be fabricated by varying the zeta potential of lipid nanoparticles. The nanoparticles are delivered specifically to the lungs, liver, spleen, and lymph nodes. In addition, nanoformulations can be enshrouded by a layer of biomolecules in the bloodstream to form a protein corona, which plays a pivotal role in organ-targeted delivery. The protein compositions in the corona vary according to the amphiphilicity and stability of nanoformulations. The interactions between the surface-bound proteins and cognate receptors highly expressed in specific tissues lead to organ-targeting effects.

(2) The second order: cell targeting. It can be achieved by modifying the nanoformulations with targeting ligands, such as antibodies, peptides and other small molecules that can interact with the receptors on the cell membrane. Antibodies exhibit extremely high binding affinity, and chemical modifications possibly denature these ligands. Moreover, the introduction of antibodies on the surface might increase the size of the resulting delivery systems. Peptides and other small molecules show outstanding tolerance to chemical reactions, while their stability and targeting effect in vivo are sometimes unsatisfactory. Thus, it is necessary to establish a balance between their advantages and disadvantages in biomedical applications.

(3) The third order: organelle targeting. It can be achieved by conjugating ligands to monomers. Morpholine, TPP, and methyl sulphonamide moieties have demonstrated excellent targeting ability toward lyso/endosomes, mitochondria, and endoplasmic reticulum, respectively. These groups can be easily conjugated to the monomers, thus facilitating the polymerization in targeting organelles.

To enrich the theranostic functions, diagnostic and therapeutic monomers are urgently desired for in situ polymerization. Fluorescence dyes, radiolabeling chelators, chemotherapeutics, and immune-regulating drugs can be modified with polymerizable groups, such as acrylate and methacrylate. Owing to the multivalent effect of polymerization, it is expected to improve the sensitivity and resolution of imaging signals. For therapeutic applications, the efflux phenomenon responsible for drug resistance can be solved through the intracellular polymerization of prodrug monomers. The use of cleavable linkers facilitates the release of potent drugs from the formed polymers or assemblies to activate their therapeutic functions. Moreover, the polymerization modifications should be scrutinized during design. There are also ideas to screen small molecules with both polymerizable sites and pharmacodynamic activity from the molecule database through high-throughput screening using artificial intelligence with powerful prediction ability.

In addition to the pharmacophoric moiety that exerts therapeutic effects and polymerizable functional groups, domains that can regulate the solubility of the entire drug, activate pharmacophoric moieties, or have an enrichment effect on specific organs or cells, can also be integrated. For example, the introduction of sugar compounds and peptides consistently improves solubility, favoring their pharmacokinetic behaviors and final therapeutic performances.

(1) Cascade reactions for polymerization. After delivery of non-toxic and inactive molecular precursors to targeting cells, specific enzymes in the cells endow monomers with pharmacological activity and polymerization activity, thereby realizing cascade polymerization reactions. At present, there are very few reported cases of this line of thought. However, the polymerization cascade reaction has a broad application, because it can solve the issue of toxic and side effects of monomer molecules due to off-targeting, that is, to exert drug efficacy only at the active sites.

(2) Bioorthogonal polymerization. Several cases of bioorthogonal polymerization in living organisms have been reported. However, various polymerization modes can be developed combined with the high-efficiency characteristics of bioorthogonal reactions. For example, the modification of monomers with orthogonal groups (Fig. 19f) endows the monomers with the ability to polymerize in complex microenvironments.

(3) Biological enzyme-catalyzed polymerization. Biological enzymes have specific and highly efficient catalytic activities. As a result, the monomers can be directly polymerized into polymer chains using enzymes in living organisms. For example, horseradish peroxidase and soybean peroxidase can be used to catalyze the polymerization of phenolic compounds. In addition, some highly expressed enzymes in the tumor have the ability of targeted enrichment. Tyrosinase in malignant cancer cells can catalyze the polymerization of short peptides to promote antigen cross-presentation, which plays a pivotal role in cancer immunotherapy.

(4) Supramolecular polymerization. It is important to take full advantage of the reversible and stimuli-responsive ability of supramolecular polymeric substances. Noncovalent interactions occur when building blocks are incorporated into monomers, initiating supramolecular polymerization in living organisms. Considering the specific microenvironment in tumors, it is possible to construct supramolecular polymers that can self-assemble only in tumors triggered by abnormal conditions.

Naturally, the assembly of tubulins, and the double helical structure of DNA or DNA origami are identical to hybridization covalent polymerization and supramolecular polymerization. Specifically, the first step can be realized in situ via covalent polymerization. The second part of supramolecular polymerization can be recognized by influential intermolecular forces, such as multiple hydrogen bonds. This direction has the potential to realize polymer-based molecular machines in living organisms.

(2) Virus manipulation. Virus activity can be influenced by chemical reactions. For example, amine groups on the surface of the virus can react with epoxy compounds or formaldehyde, which is the basic principle for manufacturing inactivated or attenuated virus vaccines. Therefore, the spike proteins on the surface of the virus can possibly be engineered as the sites for polymerization, inhibiting the ability of the virus to infect and multiply. Thus, the virus can be converted into a potential vaccine in one step.

(3) Cell manipulation. Intracellular polymerization can significantly change the cell cycle, cytoskeleton structure, cell motility, and other properties. Liver fibrosis, thrombosis and other vital diseases are expected to be regulated through intracellular polymerization. For example, liver fibrosis is often caused by excessive accumulation of extracellular matrix. The metabolism of the superfluous extracellular matrix plays a vital role in chronic inflammatory diseases. Long-term regulation of the microenvironment in the pathological area can be achieved through in situ polymerization of inflammatory regulatory molecules in the extracellular matrix.

Unlike the regulation of other cellular functions, direct manipulation of immune cells is of a higher value, especially in cancer therapy and autoimmune diseases. T cells, responsible for the elimination of cancer cells, are often tricked by immune checkpoints on cancer cells, which lead to premature exhaustion of T cells. One of the important reasons is that the glycosylation signal recognition on the surface of cancer cells can manipulate T cells. As a result, it has become a promising alternative to traditional antibodies with glycan molecules through polymerization reactions, especially sugar polymerization reactions. However, manipulating sugar synthesis in vivo remains blank or unknown. As the most critical and powerful antigen-presenting cells of acquired immunity, the development of dendritic cells (DCs) is one of the most effective ways to stimulate the body's immune system to resist cancer and virus invasion. It will be a milestone if the activation and antigen-presentation of DCs can be regulated by polymerizing unique monomers inside the cells. In addition, the repolarization of macrophages in the tumor microenvironment and the disruption of interactions between tumors and T cells need to be further explored using the tool of in situ polymerization by researchers.

The Nobel Prize of 2022 was awarded for the scientists who found the click chemistry and bioorthogonal chemistry.215 Bioorthogonal chemistry is particularly attractive for its ability to conduct reactions within living organisms without interfering with intrinsic chemical reactions. On the other hand, in situ polymerization involves creating multifunctional polymers within living organisms, and polymers play a central role in all stages of life activities. In vivo polymerization that can be achieved through bioorthogonal chemistry and free radicals has emerged as a direct regulator of metabolism, immunity, and signal transduction. Therefore, it is of great significance to explore the in situ polymerization reaction in vivo. By combining advanced nanotechnology with polymer chemistry,216–218 more breakthroughs can be expected in the field of polymerization in living organisms, which will ultimately benefit humanity in the near future.