Cell membrane biomimetic nanomedicines for cancer phototherapy

Phototherapy, mainly including photothermal therapy (PTT) and photodynamic therapy (PDT), can kill cancer cells by generating heat or reactive oxygen species, which has the advantages of being minimally invasive, high efficiency, and low toxicity. However, traditional phototherapeutic agents face challenges such as poor tumor targeting, susceptibility to passive immune clearance, and suboptimal biocompatibility, which limit their clinical application. Recently, cell membrane biomimetic technology endows phototherapeutic agents with unique biological functions, such as promoted immune escape, prolonged in vivo circulation time, improved biocompatibility, and enhanced anti‐tumor efficacy. In addition, phototherapy mediated by cancer cell membrane (CCM) or immune cell membrane‐modified phototherapeutic agents can promote broad anti‐tumor immunity. In this review, we deeply analyze the mechanisms of PTT and PDT, systematically discuss the synthesis strategies and biological functions of cell membrane biomimetic nanomaterials, and focus on the progress of phototherapy based on biomimetic nanomaterials and synergistic therapies such as chemotherapy, radiotherapy, immunotherapy, and sonodynamic therapy. Finally, we address the opportunities and future prospects of biomimetic nanomaterials in the field of cancer phototherapy. This comprehensive review is expected to provide insights into promoting the clinical translation of biomimetic phototherapeutic agents.


| INTRODUCTION
For centuries, cancer has been regarded as one of the major public health problems in the world, threatening human life and health. 1 At present, conventional clinical treatments mainly include surgery, chemotherapy (CT), and radiotherapy (RT). They suffer from low selectivity, severe toxic side effects, and insufficient tumor specificity, limiting the probability of tumor eradication. 2 Therefore, there is an urgent need to develop safe and effective therapeutic anti-tumor therapies. Phototherapy, mainly including photothermal therapy (PTT) and photodynamic therapy (PDT), has attracted widespread attention due to its advantages of minimal invasiveness, high efficiency, and spatiotemporal controllability. 3 Moreover, phototherapy selectively kills cancer cells by regulating the region, time, and dose of light irradiation, preferably nearinfrared (NIR) light with great tissue penetration ability. 4 Generally, PTT relies on photothermal agents (PTAs) to convert light energy into heat to ablate cancer cells. 5 PDT is based on photosensitizers (PS) that generate reactive oxygen species (ROS) induced by light to cause oxidative damage to cancer cells. 6 Currently, a variety of nanomaterials including noble metals, 7 carbon-based nanomaterials, 8 conjugated polymers, 9 two-dimensional nanomaterials, 10 Fe 3 O 4 nanomaterials, 11 TiO 2 nanomaterials, 12 and upconversion nanomaterials 13 have been developed for phototherapy. However, most phototherapeutic agents based on nanomaterials are difficult to reach clinical applications. The main reasons include the following aspects: (1) Nanoparticles-based preferential accumulation in tumors through enhanced permeability and retention (EPR) effect is controversial due to limited efficacy. 14 (2) The host mononuclear phagocyte system (MPS) captures and clears away the nanoparticles administrated systemically, resulting in short half-life and poor tumor accumulation. 15 (3) Phototherapeutic agents are not tumor-selective, causing uncontrollable biodistribution in normal tissues such as skin, blood vessels, and liver, further inevitably leading to the unnecessary side effects. 16 To improve biodistribution and extend halflife, the surface of phototherapeutic agents had been modified with biomaterials including liposomes, 17 polyethylene glycol, 18 hyaluronic acid, 19 and peptides. 20 However, there still exist challenges such as immune recognition and removal, insufficient tumor accumulation, and potential biological toxicity. 21 In recent years, cell membrane biomimetic technology has developed rapidly to address the above issues, breaking through the limitations of traditional surface modification in biomedical applications. Cell membrane biomimetic nanomaterials inherit the surface properties of source cells, endowing phototherapeutic agents with unique biological functions (e.g., homologous targeting, prolonged blood circulation, immune escape, and improved biocompatibility). 22 Currently, many types of cell membranes have been used to camouflage phototherapeutic agents, including red blood cell (RBC) membranes, 23 platelet (PLT) membranes, 24 macrophage membranes, 25 cancer cell membranes (CCM), 26 mixed cell membranes, 27 and engineered cell membranes. 28 The cell membrane biomimetic phototherapeutic agents integrate the photo-responsive performance of nanomaterials and the antigenic diversity of the source cells. 29 In addition, cell membrane biomimetic strategy can not only improve the biodistribution of nanomedicine to enhance phototherapeutic performance but also induce anti-tumor immune responses to enhance anti-tumor efficacy. 30 Furthermore, synergistic anti-tumor treatments such as CT, 25 immunotherapy (IMT), 31 sonodynamic therapy (SDT), 32 and RT 33 have been designed to maximize the advantages of phototherapy by overcoming the heterogeneity of tumors.
In this review, we comprehensively summarized the research progress of cell membrane biomimetic nanomaterials for cancer phototherapy. Firstly, we discussed the related mechanisms of PTT and PDT. Subsequently, we provided an in-depth analysis of the synthesis strategies and biological functions of cell membrane biomimetic nanomaterials. Next, we also outlined the recent progress of cell membrane biomimetic nanomaterialmediated phototherapy and other synergistic therapies (Scheme 1). Finally, we addressed the potential challenges and future prospects of cell membrane biomimetic nanomaterial-mediated phototherapy. We believe that this review is expected to provide theoretical guidance for the development of biomimetic materials and promote the clinical translation of phototherapy.

| Mechanism of action for PTT
One of the important factors of PTT is PTAs that convert light energy into heat energy to ablate cancer cells. Since tumor cells are less tolerant to high temperature than normal cells, PTT can be used as a selective therapy for tumor therapy. 34 Part of PTAs can emit fluorescence (FL) while generating heat after absorbing light energy, which can be used for FL imaging, or generate acoustic waves through thermal expansion, which can be used for photoacoustic (PA) imaging. 35 The photothermal conversion mechanisms can be divided into three categories according to the difference in the response of different types of PTAs to light, namely the localized surface plasmon resonance (LSPR) effect, the generation and relaxation of electron-hole pairs, and molecular thermal vibration. Some PTAs with multiple photothermal components can realize a combination of multiple photothermal conversion mechanisms. 36 The photothermal mechanism mediated by noble metal nanoparticles is usually considered to be as the LSPR effect. The LSPR effect means that when light irradiates the nanoparticles, if the frequency of the incident photon is consistent with the LSPR frequency of the noble metal, the nanoparticles strongly absorb the photon energy and generate an amplified local electric field on the nanoparticles surface. 37 As shown in Figure 1A, the absorption of photon energy by electrons lead to non-equilibrium heating and generate hot electrons. Subsequent relaxation occurs through electron-electron scattering processes and the metal surface temperature increases. Further cooling to equilibrium by electronphonon coupling by energy exchange. Finally, energy is transferred from the lattice into the surrounding medium via phonon-phonon coupling, increasing the surrounding temperature. 35b,38 It is worth noting that some metal oxide and metal sulfide nanomaterials, such as MoO 3-x , 39 In 2 O 3 , 40 W 18 O 49 , 41 CuS, 42 and MoS 2 43 also exhibit LSPR photothermal conversion mechanism. Another photothermal mechanism is the generation and relaxation of electron-hole pairs generally represented by semiconductor materials (e.g., SnS, 44 TiO 2 , 45 and Co 3 O 4 46 ) with intrinsic absorption band gaps. As shown in Figure 1B, when the semiconductor materials are excited by incident light energy higher than their band gap energy, electrons in the valence band (VB) absorb the photon energy and then transition to the conduction band (CB), leaving holes in the VB and forming electron-hole pairs. Electrons in the excited state are unstable. The electron-hole pairs relax to the band edge through non-radiative paths and release energy in the form of phonons, which causes lattice vibrations to generate heat. 47 Importantly, the photothermal conversion efficiency is related to the recombination of electronhole pairs influenced by band gap.
Most carbon-based materials and organic materials have π-conjugated structures and the photothermal mechanism follows molecular thermal vibrations. Compared to σ bonds, π bonds have a lower energy gap easily excited through π-π* orbitals. When the energy of the incident light matches the energy of the electron transitions in the molecule, the electrons are excited from the ground state π orbital (highest occupied molecular orbital [HOMO]) to a higher excited π* orbital (lowest unoccupied molecular orbital [LUMO]) to generate excitation electronic ( Figure 1C). Excited electrons transfer absorbed light energy from the excited electrons to the atomic lattice through electron-phonon coupling relaxation, generating heat through lattice vibrations. In addition, the π bonds will red-shift the absorption peak of the material and the maximum absorption wavelength of the material from the visible light region to the NIR region can be adjusted by changing the number of π bonds. It is also possible to reduce the energy gap between LUMO and F I G U R E 1 (A) LSPR effect. (B) The generation and relaxation of electron-hole pairs. (C) Molecular thermal vibration. (D) PTTmediated cell death pathways. HOMO, highest occupied molecular orbital; ICD, immunogenic cell death; LSPR, localized surface plasmon resonance; LUMO, lowest unoccupied molecular orbital; PTT, photothermal therapy; TME, tumor microenvironment HOMO in the molecule by increasing the number of π bonds in the molecule, thereby improving the photothermal conversion efficiency. 48 The temperature range of PTT in clinical application is controlled between 41 and 48°C. When the cellular temperature exceeds 39°C, the proteins begin to aggregate and deform. When elevated to 41-42°C, cells temporarily inactivate and produce heat shock proteins to reduce thermal damage. When the temperature is further increased, irreversible tissue damage occurs, eventually leading to cell necrosis and apoptosis. 49 As shown in Figure 1D, the mechanism of PTT-mediated tumor treatment includes the following three aspects: (1) High temperature induces protein denaturation; (2) PTT induces immunogenic cell death (ICD) and activates the inflammatory response; (3) PTT can also remodel the tumor microenvironment (TME). These include destroying extracellular matrix, improving oxygenation, increasing blood flow, causing tumor vascular damage, and lowering tumor pH, thereby improving the anticancer efficacy of other therapies such as CT, RT, and IMT. 49a,50 So far, clinical studies have been carried out on PTAs including indocyanine green (ICG) and gold-silica nanoshells. 51

| Mechanism of action for PDT
PDT is an important clinical technology for the treatment of cancer, especially superficial tumors. It mainly induces cancer cell apoptosis or necrosis through the dynamic interaction of laser with specific wavelength, PS, and molecular oxygen. 52 Under laser irradiation, PS-mediated photodynamic reactions include photophysical and photochemical processes, and finally generate ROS to act on target tissues. The photodynamic reactions in PDT can be divided into two categories according to the different pathways and types of ROS production ( Figure 2A). Illuminated PS absorbs photons and transforms from the ground state (singlet state, 1 PS) to the short-lived excited singlet state 1 PS*, which can decay back to the ground state by FL decay. ISC is performed to generate long-lived excited triplet states 3 PS*. 3 PS* generates ROS through type I and type II photodynamic reactions. In the type I pathway, triplet 3 PS* reacts directly with substrates such as cell membranes or molecules, transferring protons or electrons to form anionic or cationic radicals such as superoxide anion (O 2 •-), hydroperoxide radical (HOO•), peroxides (H 2 O 2 , ROOH), and hydroxyl radicals (•OH). Type I responses can achieve potent PDT effects in hypoxic TME. In the type II pathway, the triplet state 3 PS* transfers energy to the triplet ground state molecular oxygen ( 3 O 2 ), and 3 O 2 transitions from the triplet ground state to the excited singlet state to form singlet oxygen ( 1 O 2 ). 53 1 O 2 has a long half-life (~0.04 μs) and an action radius (~0.02 μm), and is highly toxic, which can effectively induce oxidative stress damage in cancer cells. 54 Both type I and type II reactions can occur simultaneously, and the ratio between these processes depends on the type of PS, the concentrations of substrate and oxygen, as well as the binding affinity of the sensitizer for the substrate. 55 As shown in Figure 2B, PDT-mediated tumor therapy typically involves three mechanisms: (1) Direct killing of tumor cells through oxidative stress; (2) Disruption of tumor-associated vasculature, resulting in disruption of O 2 and nutrient supply; (3) Triggers an inflammatory response and activates immune cells to enhance host immunity. 56 Currently, various PSs such as porfimer sodium (Photofrin), 5-aminolevulinic acid (5-ALA), and ICG have been approved for clinical PDT. 57

MEMBRANE BIOMIMETIC NANOMATERIALS
The preparation method of cell membrane biomimetic nanomaterials mainly includes three parts: nanomaterials selection, cell membrane extraction, and fusion of membranes and nanomaterials. 58 The fusion of the cell membranes and nanomaterials is the key to preparing cell membrane biomimetic nanomaterials and many strategies have been proposed to achieve it. Among them, coextrusion and sonication are common techniques, and microfluidic electroporation is regarded as an advanced strategy. 59 The co-extrusion method works by repeatedly extruding nanomaterials and membrane vesicles through a mini extruder with porous polycarbonate membrane. The mechanical force generated in this process destroys the membrane integrity and then reconstructs around the inner core nanomaterials. After repeating extrusion several times, the unbound cell membrane vesicles are separated and cell membrane biomimetic nanomaterials are collected by centrifugation. This method can obtain resultant nanoplatform with a uniform size distribution, but it is not conducive to large-scale production due to low yield and time-consuming process. 60 Co-extrusion is the most widely used method for fusing cell membranes and nanomaterials. For example, Sun et al. 61 used mesenchymal stem cells (MSCs) membranes to wrap the perfluorocarbon (PFC) and sound sensitized agent verteporfin to prepare a biomimetic sound sensitized agent (M/LPV/O 2 ) that could effectively induce tumor inhibition ( Figure 3A). Sonication method means that the ultrasonic energy of a specific frequency can disperse the nanomaterials and provide destructive power to destroy the membrane structure in ice-cold condition, resulting in the spontaneous formation of cell membrane biomimetic nanomaterials with yolk shell structure. 64 The product is collected by centrifugation CHEN ET AL. and removing the left cell membrane. Importantly, the fusion efficiency and cell membrane integrity can be optimized by controlling the parameters of ultrasound such as frequency, power, and duration. In 2021, Liu et al. 62 prepared cell membrane biomimetic SiO 2 nanomaterials using sonication method, and proposed to use fluorescence quenching experiments to evaluate the integrity of cell membrane coatings ( Figure 3B). Compared with co-extrusion, the sonication method has the advantages of low material loss, simple operation, and timesaving. 29 It is microfluidic electroporation that has been considered as an advanced strategy for synthesis cell membrane biomimetic nanomaterials. The work principle of microfluidic electroporation is similar to sonication method. 65 In the microfluidic system, cell membrane and nanomaterials are mixed in an S-type channel and fusion of cell membrane and nanomaterials in a Y-type microchannel through the electroporation zone. During electroporation, the membrane pores allow nanomaterial enter into the cell membrane. 66 This strategy provides outstanding advantages, including high throughput, quantitative determination, scalability, and storing capacity despite the slightly high cost, which promotes its feasibility for industry translation. 29 In addition to the above methods, some unique strategies including spontaneous formation by electrostatic attractions and in situ polymerization also have been proposed. 67 For example, Li et al. 63 prepared platelet membrane-mimetic SiO 2 nanocarriers for drug delivery at tumor sites. The negatively charged platelet membrane was fused with the positively charged APTES-SiO 2 nanomaterial through electrostatic attractions ( Figure 3C). However, the principle of these approaches is different from traditional methods, their feasibility and value are yet to be verified. It is worth noting that the electronegativity and stabilizing effect caused by the stealth coating of hydrophilic polysaccharides of nanomaterials play an important role in the assembly process, so the interaction between nanomaterials and cell membrane should be considered substantially for the fabrication of cell membrane biomimetic nanomaterials. 68

| Affecting the in vivo fate of nanomaterials
To date, most nanoparticles have been passively targeted to tumor tissues mainly relying on the EPR effect. 69 Generally, only a small amount of nanoparticles can target the interested tissue through EPR effect, and most of them are cleared from the blood by the MPS, which shortens the blood circulation time, reduces the therapeutic effect, and results in side effects to non-target sites. 70 To overcome the above obstacles, biomimetic nanoengineering have been developed, providing opportunity for changing in vivo fate of nanomaterials. Cell membrane biomimetic technology is a new class of biomimetic nanoengineering, which integrates the advantages of both natural cells and artificial nanomaterials such as specific targeting ability, prolonged blood circulation, and promoted immune escape. Typically, Zhang's group developed nanoemulsions containing perfluorocarbon coated by RBC membrane (RBC-PFC). RBC membrane could reduce PFC uptake by cells and promote immune escape. RBC-PFC sowed the potential to replace RBC for blood transfusion. 71 Encouraged by the natural property of RBC, Liu et al. 72 designed an aggressive man-made RBC (AmmRBC) for oxygen-selfsupplied PDT to tumors ( Figure 4A). AmmRBC avoided immunogenic clearance and had long blood circulation benefiting from the ample natural markers of RBC. In addition, various cell membranes such as PLTs, cancer cells, and stem cells have the ability of tumor targeting based on the chemotaxis and surface interaction. For example, PLTs can recognize each other and affect the metastasis of tumor cells. Nanoparticles coated with PLTs membrane can prolong the circulation time in vivo and increase the accumulation in tumor sites. 74 Based on this property, Rao et al. 24 designed PLT-coated Fe 3 O 4 nanoparticles (PLT-MNs) for PTT and magnetic resonance imaging (MRI) ( Figure 4B). PLT-MNs possessed both the long-circulation, tumor-targeting ability, and decreased immune system response of PLTs and the magnetic and light-absorbing properties of Fe 3 O 4 . And it was found that the targeting of PLT-MNs and the effect of PTT can be mutually enhanced. They also developed cancer cell membrane-coated upconversion nanoparticles (CC-UCNPs) by exploiting the immune evasion and homologous targeting capabilities of cancer cell surface membrane proteins ( Figure 4C). 73 To assess the immune evasion ability of nanoparticles, UCNPs, RBC-UCNPs, and CC-UCNPs were incubated with macrophages for different times, and uptake was quantified by measuring Yttrium ion (Y 3+ ) content. The nanoparticles that were not coated by the cell membrane had the highest uptake rate by macrophages, which proved that the modification of the cell membrane can improve the immune evasion ability of the nanoparticles. The blood circulation time and targeting ability of CC-UCNPs were evaluated and CC-UCNPs exhibited better results compared to UCNPs and RBC-UCNPs, which proved that CC-UCNPs can accumulate the tumor region through homologous targeting. The combination of cell membrane biomimetic strategy and passive targeting improves the efficiency of nanoparticle targeting to tumor sites, which is expected to achieve effective anti-tumor effects simultaneously with the use of low-dose nanomedicines.

| Improving stability and biocompatibility of nanomaterials
Good physiological stability and biocompatibility are the basic requirements for the in vivo application of CHEN ET AL.
-7 of 24 nanoparticles. 75 Most nanoparticles have limited their clinical translation due to low physiological stability, poor biocompatibility, and insufficient tumor targeting. 76 Hydrophilic polysaccharides on the surface of cell membranes can reduce the adsorption to proteins in vivo and improve the stability and biocompatibility of internal nanoparticles. 77 For example, Gao et al. 78 used stem cell membrane-coated mesoporous silica (SiO 2 )-encapsulated UCNPs to obtain SUCNPs@SiO 2 . In phosphate buffered saline (PBS) and fetal bovine serum (FBS), SUCNPs@mSiO 2 remained relatively stable for 2 weeks ( Figure 5A). To evaluate the biocompatibility of the materials, SUCNPs@mSiO 2 were treated with mouse blood.
The results showed that SUCNPs@mSiO 2 did not exhibit obvious hemolysis, proving the good biocompatibility of SUCNPs@SiO 2 ( Figure 5B). Mature red blood cells are beneficial to membrane extraction and purification due to their lack of organelles and their large numbers. 81 Therefore, RBC membrane is the first natural cell membrane used to coat nanomaterials, and it is also one of the most studied cell membranes. 82 Curcumin (Cur) is a phenolic pigment extracted from the rhizomes of turmeric with anti-tumor, wound healing, and immunomodulatory effects. 83 However, Cur exhibits low stability, poor water solubility, low bioavailability, and fast metabolism in vivo, which limits its clinical application. To  Figure 5C). RBCM-p-PLGA@Cur NPs could be uniformly dispersed in H 2 O or PBS for a long time, and had almost no toxic side effects on normal tissues in mice with good stability and biocompatibility. It has been reported that the metal-organic framework (MOF) zeolite imidazolate framework 8 (ZIF-8) is often used as a drug carrier, but it has a certain potential toxicity in in vivo applications because it releases Zn 2+ to accumulate in cells and eventually induces cell necrosis. 84 Qiao et al. 80 showed low cytotoxicity in cell experiments after coating ZIF-8 and doxorubicin (DOX) with RBCM (ZIF-8-DOX-LY-RM NPs), which improved the biocompatibility of the carrier ( Figure 5D). In conclusion, the stability and biocompatibility of the nanoparticles can be improved to a certain extent after being coated by the cell membrane, which is beneficial for in vivo applications.

| Regulating anti-tumor immunity
The immune system plays an important role in cancer therapy. However, cancer immunosuppression induces cancer cell proliferation and rapid recurrences or metastases, undermining the effectiveness of common cancer treatments. 85 Coincidentally, the immunotherapies based on nanomedicine and biomimetic engineering have the potential for reducing cancer immunosuppression and  86 The current reported cancer immunotherapies mainly target the stage of tumor immune escape and aim to modulate immune cells in lymphoid tissue and TME to suppress tumors. Anti-tumor immunity triggered by IMT can not only fight against metastatic tumors but also establish longterm immune memory to prevent tumor recurrence. 87 The rapid development of biomimetic technology and nanotechnology has promoted the application of cancer IMT. Specific targeting with biomimetic nanoparticles can effectively reduce off-target toxicity and immune side effects in IMT. 86 Cell membrane biomimetic nanomaterials play essential roles in improving IMT efficacy due to the specific immune-associated functions, including macrophage polarization capability, antigenicity, and immune checkpoint inhibition. Particularly, activated M1 macrophages is an important factor in weakening protumor ability and boosting antitumor response. Thus, using biomimetic nanomedicine such as natural killer (NK) cell membrane-coated nanomedicine to polarize M2 macrophages to M1 macrophages is a promising approach. NK cell membranes can trigger M1 macrophage polarization to stimulate the immune system to induce anti-tumor immunity. Zou et al. 88 designed artificial NK cells (aNK) to recruit immune cells and polarized macrophages to kill tumor cells ( Figure 6A). Deng et al. 89 sed NK cell membranes to coat the photosensitizer Tetra(4-carboxyphenyl) porphyrin (TCPP) nanoparticles (NK-NPs), which could induce strong anti-tumor immunity in vivo and effectively inhibit primary tumors and distal tumors growth. TCPP could damage cancer cells through ROS generated by PDT and induce ICD to improve the anti-tumor immunity efficiency of NK cell membranes ( Figure 6B). Differences in mRNA levels of M1 and M2 macrophage markers were analyzed by real-time quantitative polymerase chain reaction (RT-PCR) and results showed that macrophages treated with NK-NPs and NK cell membranes significantly increased the expression of the M1 macrophage marker iNOS/CD86 and decreased the expression of the M2 macrophage marker CD206 compared with the control group. Therefore, NK cell membrane biomimetic nanoparticles are expected to become a new strategy for cell membrane IMT. Because tumors have inherent evasion of immunological surveillance, which leads to immunosuppressive responses through multiple pathways including downregulating tumor-associated antigens expression and upregulating immune checkpoint molecules expression. To better prevent local and distant tumor recurrence, the immune checkpoint blockade (ICB) strategy exposes a notable potential in enhancing IMT. Zhang et al. 90 genetically engineered platelets to express the PD-1 and the derived microparticles, which could target to tumor as well as revert exhausted CD8 + T cells, eventually directly preventing tumor relapse ( Figure 6C). For example, Yin et al. 28a designed a biomimetic nanodrug delivery of engineered macrophage membrane-coated polylactic-coglycolic acid (PLGA) and rapamycin (RAPA) composite spheres with programmed cell death-1 (PD-1) expression system (PD-1-MM@PLGA/RAPAs) for glioblastoma therapy. PD-1 antibodies can effectively block the programmed cell death ligand protein 1 (PD-L1) immune checkpoint and upregulate adaptive immune responses. Due to the complexity of the blood-brain barrier (BBB), which hinders the delivery of PD-1 antibodies to tumor sites and further reduces the efficacy. It has been reported that macrophages can cross the BBB and target tumor tissue. 91 PD-1-MM@PLGA/RAPAs could not only reduce damage to the central nervous system across the BBB, but also combine ICB and CT to obtain the greatest antiglial in vivo compared to other groups.
We have summarized the coating method of cell membrane biomimetic nanomaterials and biological functions of cell membrane in Table 1 for direct comparison.

| Single-mode phototherapy
Phototherapy has developed rapidly in recent years due to its minimally invasive, highly effective, and low drug resistance in cancer treatment. 92 Nanomaterial-based PTAs are usually cleared by MPS in the liver and spleen, resulting in less dose reaching the tumor area and lower phototherapy efficiency. Therefore, it is very necessary to design a nanomedicine that can circulate for a long time in vivo and specifically target tumor tissue. Cell membrane biomimetic strategy can achieve these effects. For example, ICG is widely used in clinical research because of its good photothermal performance and biological safety. However, free ICG has the disadvantages of poor tumor targeting specificity and easy to be quickly cleared in vivo, which limits its application. 93 To overcome these shortcomings, Chen et al. 26 coated PEGylated MCF-7 CCM with ICG/polylactic-glycolic acid nanoparticles (ICNPs) for FL/PA imaging and PTT ( Figure 7A). The CCM on ICNPs could reduce the interception of nanoparticles by liver and kidney, and achieve high accumulation at tumor sites through homologous targeting. In addition, it was proved in animal experiments that ICNPs accumulate the most in tumors after intravenous (i.v.) injection compared with other nanomaterials without cell membrane modification, which enhanced the heating effect inside the tumor during PTT and increased the tumor inhibition rate to 100%.
In addition to the use of homologous targeting of CCM to increase the accumulation in the tumor area, other sources of cell membranes such as RBC, PLT, and myeloidderived suppressor cells (MDSC) can also achieve immune  Figure 7C). In animal experiments, the accumulation of MNP@RBC and MNP@MDSC in the liver and spleen were significantly lower than MNP, proving that cell membrane biomimetic strategy can effectively reduce the uptake of MNPs by MPS. In the B16/F10 melanoma mouse model, the MNP@MDSC + Laser group showed the best tumor suppressive effect ( Figure 7D). Subsequently, the expression levels of high-mobility group protein B1 and calreticulin were detected with the tumor sections of each group after treatment, which proved that the MNP@MDSC + laser group could significantly induce ICD. Meanwhile, immunofluorescence analysis of tumor sections showed that the synergistic treatment of MNP@MDSC and PTT was beneficial to M1 macrophage polarization and enhanced CD8 + T cell infiltration.
The blood circulation time is prolonged and tumor targeting ability is significantly enhanced for nanomaterials benefiting from cell membrane biomimetic strategy and some external energy fields such as magnetic fields. In 2020, Lin et al. 23 used RBCM to coat Fe 3 O 4 @Cu 2 −x S nanoparticles (SCS@RBCM). Under the protection of RBCM, the clearance of SCS by macrophages was reduced. Moreover, under the guidance of an external magnetic field, the tumor accumulation of SCS@RBCM was significantly increased, and synergistic PTT exhibited a good anti-tumor effect.
It can be seen that the therapeutic effect of cell membrane biomimetic nanomaterials in vivo is better than the similar non-encapsulated ones. Although phototherapy has shown good therapeutic effects on solid tumors in animal models, laser can only penetrate  Abbreviations: NK, natural killer; PLGA, polylactic-glycolic acid; RBC, red blood cell.

of 24 -
small and shallow tumors because the scope of phototherapy is limited by the penetration depth of the laser. And phototherapy cannot solve the problem of tumor recurrence. In addition, the heat generated during PTT inevitably damages the normal tissue surrounding the tumor. These problems limit its clinical application. To address these issues, researchers combine phototherapy with CT, IMT, and other therapies to achieve synergistic effects on tumors. 4, 96 We summarize the cancer phototherapy and synergistic treatment strategies for different cell membrane biomimetic nanomaterials in Table 2.

| Combination of phototherapy with CT
CT has been widely used in cancer treatment. At present, commonly used anti-cancer drugs include cisplatin, paclitaxel (PTX), and DOX. 103 However, CT efficacy is affected by tumor heterogeneity and drug resistance. In addition, due to the lack of selectivity of anti-cancer drugs, CT can produce toxic side effects on normal tissues. 104 To reduce the side effects, nanomaterials can be used as carriers for loading chemotherapeutic drugs and passively target to the tumor site through the EPR effect. Cell membrane biomimetic nanomaterials can further improve targeting efficiency and reduce side effects. The high temperature generated by PTT can increase blood flow and improve oxygenation within the tumor, thereby increasing the rate of drug delivery and enhancing efficacy. Therefore, phototherapy synergistic with CT is a promising therapeutic strategy. After the cell membrane biomimetic nanomaterials are loaded with drugs, the properties of TME (e.g., acidity, high concentration of H 2 O 2 , and reduced glutathione) and light-triggered methods can be used to control drug release. 105  combining CT with CuS-based PTT ( Figure 8B,C). Lei et al. 25   signaling pathway and eventually led to cell apoptosis.
With the temperature rise caused by PTT effect, ABC decomposes into NH 3 and CO 2 , which further triggered the release of DOX to enhance the inhibitory effect on tumor cells ( Figure 8D). The authors verified that intratumoral (i.t.) injection of DIRA combined with PTT/PDT and CT could effectively inhibit the growth of breast cancer at the animal level, allowing 80% of mice to survive 100 days. Phototherapy based on cell membrane biomimetic nanomaterials combined with CT in the treatment of cancer has shown many therapeutic advantages, including: (1) Reducing the dose of chemotherapeutic drugs and reducing side effects; (2) The biomimetic function endowed by cell membranes and EPR effect of nanomaterials can enhance the delivery of chemotherapeutic drugs to the tumor site; (3) Cell membrane biomimetic strategy can prevent premature leakage of drugs in vivo 106 ; (4) The heat generated by PTT can promote the delivery of chemotherapeutic drugs inside the tumor; (5) The ROS generated during PDT can control the release of ROS-responsive drug 107 ; (6) Phototherapy combined with CT can effectively inhibit tumor metastasis. However, phototherapy combined with CT still has limitations of insufficient drug loading and difficulty in determining the drug dose during combined treatment. 4 Therefore, it is crucial to design an effective drug delivery system.

| Combination of phototherapy with IMT
Cancer IMT can activate or enhance the innate immune system to recognize and kill cancer cells, and plays an important role in cancer therapy. 75 Cancer IMT includes ICB therapy, cancer vaccines, cytokine therapy, and chimeric antigen receptor cell therapy. 76 Although IMT has developed rapidly in cancer treatment in recent years, there are still some challenges related to safety and efficacy of IMT. For cancer vaccines, the complex manufacturing process and uncertainty about optimal dosage make it difficult to achieve the expected clinical results. 77 In addition, off-target side effects and cytokine release syndrome are also the problems faced by other immunotherapies. 94 Interestingly, nanocarriers and cell membrane biomimetic strategy can improve drug delivery efficiency to tumor sites and reduce off-target effects. In addition, CCM or immune cell membrane biomimetic can enhance anti-tumor immune response. Phototherapy can also induce immune response and promote the transformation of cold tumors into hot tumors. Phototherapy combined with IMT has the effect of enhancing IMT and preventing tumor recurrence after phototherapy. 108 For example, Xiong et al. 99 coated ICGloaded Fe 3 O 4 nanoparticles with a mixed membrane composed of ID8 ovarian CCM and RBCM (Fe 3 O 4 -ICG@IRM) for ovarian cancer therapy ( Figure 9A). Compared with the monotypic cell membrane, the hybrid membrane could combine the functions of the original cell membrane. The combination of different types of cell membranes could prepare hybrid membranes with multiple functions and broaden biomedical applications. 109 Fe 3 O 4 -ICG@IRM simultaneously retained the immune escape ability of RBCM and the homologous targeting of ID8 ovarian CCM to prolong the in vivo circulation time and enhanced the accumulation of nanoparticles at tumor sites. In addition, tumor antigens were key factors in immune cells activation. The authors incubated different materials with macrophages to detect immune inflammatory cytokines, and proved that the tumor antigens on the surface of Fe 3 O 4 -ICG@IRM could activate the immune response for a long time and realize tumor-specific IMT. And the antigens released after killing tumor cells based on Fe 3 O 4 -ICG@IRM-induced PTT could also cause immune responses and effectively treat metastatic tumors ( Figure 9B).
Tumor cells aberrantly express PD-L1, which interacts with PD-1 and leading CD8 + T cell exhaustion. Administration of PD-1 antibody (aPD-1) rejuvenates CD8 + T cells from exhaustion. 110 To enhance the antitumor efficiency, Liang et al. 31 combined RBCM-coated black phosphorus quantum dots (BPQD-RMNVs) with aPD-1 to mobilize the immune system to eliminate metastatic tumors and prevent recurrence. BPQD-RMNVs exhibited long circulation and tumor-targeting properties in vivo. After BPQD-RMNVs-mediated PTT kills tumor cells, dendritic cells will be recruited to capture the released antigens and activate CD8 + T cell to inhibit tumor growth. Using aPD-1 could reduce tumor cell metastasis and restore CD8 + T cell viability. Therefore, BPQD-RMNVs-mediated PTT synergistic with IMT has great potential in the treatment of metastatic tumors.
The combination of PDT and CT is limited by tumor penetration and retention, drug release time, and the immunosuppressive environment at the tumor site. To reduce immunosuppression and effectively control drug release, Yu et al. 100 co-loaded the photosensitizer pheophytin A (PheoA), paclitaxel dimer prodrug (PXTK), and anti-PD-L1 peptide (dPPA) on gold nanoclusters (CAuNCs). Then, the mixture of BSA and HA is adsorbed on the surface, and finally RBCM is coated on the surface to form pPP-mCAuNCs@HA. pPP-mCAuNCs@HA showed prolonged circulation time in vivo and increased the amount of accumulation in tumor sites. The ROS generated by the PheoA during PDT could stimulate the release of PXTK. The produced cinnamaldehyde in the 16 of 24hydrolysis process of PXTK could also stimulate mitochondria to produce ROS, leading to apoptosis or necrosis of tumor cells. In addition, PDT could also induce ICD to enhance the immune response in vivo. However, the immune response induced by PDT alone will be limited by the immunosuppressive environment, reducing the efficacy. Therefore, combining PDT with immune check point inhibitor could reduce the limitation of the immunosuppressive environment and effectively inhibit tumor growth ( Figure 9C). The authors used dPPA to relieve immunosuppressive environment in tumors and enhance the function of cytotoxic T lymphocytes activated by PDT-induced ICD. Flow cytometry demonstrated that the combination of PDT and dPPA enhanced CD8 + T infiltration ( Figure 9D). In addition, the number of tumor-infiltrating CD4 + T cells and the levels of TNF-α and IL-12 in serum were significantly increased after synergistic treatment. In animal experiments, pPP-mCAuNCs@HA + L group had the best therapeutic effect on mice after treatment, and the tumor growth inhibition rate was as high as 84.2%. Therefore, the combined treatments of CT, PDT, and IMT could significantly enhance the anti-tumor efficacy.
Phototherapy based on cell membrane biomimetic nanomaterials can play a good synergistic effect during combined IMT. Cell membranes can enhance the immune escape ability of nanomedicines, improve the delivery efficiency of tumor sites, and some cell membranes can also activate immune responses. IMT can make up for the deficiency of phototherapy in the treatment of metastatic tumors and prevent tumor recurrence. Phototherapy can not only directly kill tumor cells through high temperature or ROS, but also improve the immune response rate of IMT. However, the efficacy of phototherapy combined with IMT is limited by the size of the tumor. Due to the shallow penetration depth of light, it is difficult to inhibit the growth of large-volume tumors. Therefore, it is necessary to develop better immunotherapeutic agents combined with phototherapy to enhance tumor suppression. 111 Enzyme-mediated elevation of ROS at tumor sites increases intracellular oxidative stress levels and damages cancer cells. Natural enzymes exhibit disadvantages such as poor stability, short half-life, immunogenicity, and systemic toxicity in vivo, which limit their biomedical applications. 112 The cell membrane biomimetic strategy can effectively overcome the above limitations, enhance the accumulation of enzymes at the target site, and improve the therapeutic effect. Due to insufficient H 2 O 2 concentration in TME and hypoxic conditions that limit the catalytic performance of the enzyme, a good tumor inhibition effect cannot be achieved. Combining enzymatic catalysis and phototherapy can enhance ROS generation and improve efficacy. For example, Zhang et al. 101 linked glucose oxidase (GOX), chloroperoxidase (CPO), and deoxygenation-activating drug tirapazamine (TPZ) to Ti 3 C 2 nanosheets and used high-expressing CD47 surface modification of CCM (m e TGCT) as a bionic cascaded-enzyme nanoreactor for breast cancer therapy ( Figure 10A). CD47 on the surface of m e TGCT could reduce nanoparticles phagocytosis by macrophages and enhance tumor site accumulation. Once in tumor cells, GOX catalyzed the production of H 2 O 2 and gluconic acid from glucose under O 2 , resulting in a decrease in pH and O 2 content. Then, CPO could catalyze H 2 O 2 and Cl − to HClO, which further generates 1 O 2 . Laser irradiation could speed up the enzyme-catalyzed reaction rate and increase the production of 1 O 2 . The cascade enzymecatalyzed reaction was verified by measuring O 2 consumption, pH value, and HClO generation. In addition, PDT exacerbated O 2 depletion in the TME and activated the chemotherapeutic drug TPZ to cause DNA fragmentation and ultimately apoptosis. Thus, m e TGCT exhibited an amplified synergistic therapeutic effect of PDT, enzyme dynamic therapy (EDT), and CT ( Figure 10B).
SDT refers to the fact that sonosensitizers can generate ROS under ultrasound (US) stimulation to cause oxidative damage to tumor cells. 113 Compared with NIR laser, US has stronger tissue penetration and is expected to break the limitation that traditional phototherapy can only effectively treat superficial tumors. The combination of SDT and phototherapy can generate more ROS at low doses and enhance cytotoxicity. 114 However, the insufficient targeting ability of sonosensitizers limits the application of SDT in the field of tumor therapy. 115 Cell membrane biomimetic strategy can enhance the tumor targeting of sonosensitizers. Thus, Shen et al. 32 synthesized CCM camouflaged iridium (III) complexfunctionalized black titanium nanoparticles (Ir-B-TiO 2 @CCM) for hierarchical targeting synergistic PTT and SDT in the treatment of cervical cancer. Compared with nanoparticles without cell membrane biomimetic, Ir-B-TiO 2 @CCM had stronger targeting to tumor cells and could selectively target mitochondria, showing a dual targeting effect to reduce damage to normal tissues. Ir-B-TiO 2 @CCM could effectively generate heat and ROS under the action of laser/US leading to cell death. Ir-B-TiO 2 @CCM could also be used as an imaging agent to identify tumor sites by performing photoacoustic imaging (PAI), photothermal imaging (PTI), and two-photon imaging (TPI) under laser/US radiation ( Figure 10C). During PTT, O 2 supply was increased, which could improve the efficiency of SDT. The authors demonstrated by flow cytometry that the combination of SDT and PTT can significantly enhance the production of ROS in tumor cells. The tumor cell death rate was 76.3% after SDT and 67.7% after PTT, while the tumor cell death rate increased to 83.3% when SDT was combined with PTT. These results further demonstrated that Ir-B-TiO 2 @CCM has a synergistic effect during SDT and PTT, which can enhance the anti-tumor effect.
RT is one of the main modalities of clinical cancer treatment. RT is based on the induction of DNA single/ double strand breaks by ionizing radiation (X-ray, γ-ray) that ultimately leads to apoptosis. 116 Due to the low selectivity, it is necessary to strictly control the dose of radiotherapy drugs in clinical practice to minimize the damage to normal tissues. Cell membrane biomimetic strategy can increase drug accumulation at tumor sites and reduce side effects. The hypoxic environment inside the tumor will increase the resistance of tumor cells to RT and weaken the therapeutic effect. 117 Photothermalinduced warming increases blood flow and improves oxygenation, thereby reducing tumor cell resistance to radiation. 118 Phototherapy is not effective in treating deep tumors and the high-energy rays used in RT can penetrate deep tissue and kill tumor cells. Therefore, the combined use of RT and phototherapy can enhance tumor suppression and reduce damage to normal tissues. 119 Recently, Yu et al. 102 utilized CCM to coat biodegradable nanoparticles (UiNPs) composed of hafnium (Hf) clusters and chloride e6 (Ce6) molecules through covalent coordination. CCM on the surface of UiNPs could enhance the immune evasion ability and tumor targeting of nanomaterials. Compared with nanomaterials without cell membrane modification, the blood circulation halflife was extended by 4 times, and the accumulation in tumor sites was increased by 2.5 times. Therefore, nanomaterials modified by CCM could improve drug delivery efficiency. Under the camouflage of CCM, UiNPs reduced the accumulation in the liver and spleen, and could be excreted through feces and urine, so it has good biological safety. UiNPs not only had dual-modality X-ray therapy with radio-radiodynamic effects (RT-RDT), but also realized dual-modality phototherapy (PDT-PTT), which improved the killing effect and penetration depth of tumor cells ( Figure 10D). In a mouse colon cancer model, the tumor growth inhibition rate after phototherapy combined with X-ray therapy was as high as 93.8%. Therefore, synergistic phototherapy with X-ray therapy is expected to achieve high-efficiency antitumor treatment under low-dose conditions. Therefore, phototherapy can also be combined with other therapies such as EDT, SDT, and RT to enhance the anti-tumor efficacy. Cell membrane biomimetic strategy can not only prevent premature leakage of internal drugs, but also improve the delivery efficiency to the tumor site and reduce side effects. Phototherapy combined with EDT can use the TME to enhance the production of ROS, which is beneficial to tumor treatment. However, there are still some problems with this synergistic therapy. As mentioned above, both GOX and PDT need to consume O 2 during treatment, which is not conducive to the continuous production of ROS. Therefore, it is very important to design a bionic cascadedenzyme nanoreactor with sustainable generation of ROS. Since US has a deeper penetration depth than NIR light, the combination of phototherapy and SDT can break through the limitation that phototherapy can only treat superficial tumors and improve the therapeutic effect. However, both PDT and SDT need to consume O 2 during treatment, and insufficient endogenous O 2 content will greatly limit the combined efficacy of PDT and SDT. 120 Similarly, when phototherapy is combined with RT, high-energy X-ray/γ-ray can also solve the problem of insufficient penetration depth during phototherapy. And during PTT, the O 2 content in tumor blood vessels can be enhanced to improve the curative effect of RT. However, there is still a major problem that high-energy rays can damage surrounding normal cells when phototherapy is combined with RT. In order to improve the effect of local precise treatment at a relatively safe dose, it is necessary to design a radiosensitizer with strong radiation absorption and transformation capabilities. Furthermore, guidance by bioimaging can enable precise treatment. Therefore, the design of nanomedicine for image-guided low-dose radiation therapy is very important. 4

| CONCLUSION AND PERSPECTIVE
In the past 2 decades, nanomaterials-mediated phototherapy has made encouraging progress in the field of tumor treatment, but its anti-tumor efficiency and clinical translation are still limited by the defects of phototherapeutic agents, such as poor targeting, easy clearance by the immune system, and poor stability. Cell membrane biomimetic strategy based on the properties of cellenvironment interactions endows nanomaterials with enhanced biological properties. This strategy does not affect the optical properties of the phototherapeutic agents and, at the same time, improves tumor homologous targeting, immune evasion ability, and biocompatibility of the nanomaterials and significantly improves the phototherapy rate. Phototherapy has become a new modality of clinical treatment of solid tumors, which can kill tumor cells by generating heat by PTT or induce cancer cell apoptosis and necrosis by generating ROS through PDT. Considering that monotherapy may lead to incomplete tumor ablation, combination therapy is the first choice for current clinical treatment. The heat generated in PTT can improve tumor tissue blood flow and O 2 content, further promoting the tumor penetration of phototherapeutic agents to enhance PDT. In turn, ROS generated by PDT can inhibit heat shock protein production, which benefits PTT. 92 Killing tumor cells and altering tumor tissue vascular permeability by phototherapy is crucial for the efficiency improvement of synergistic therapy. For example, PTT-induced changes in tumor cell permeability greatly enhance the sensitivity of RT, making it possible to effectively eradicate residual cancer cells and metastases. Taking advantage of the precise spatiotemporal localization of phototherapy, the on-demand release of chemotherapeutic drugs at the tumor can be controlled. Furthermore, phototherapy induces ICD, which triggers the release of tumor-specific antigens and cytokines to activate immune responses, creating room for synergistic IMT. 121 The tumor antigens on the surface of the CCM can enhance anti-cancer immunity. Cell membrane biomimetic nanomaterials can also be combined with PA, FI, magnetic resonance imaging, and other imaging methods in the process of phototherapy to precisely guide cancer treatment. 23,122 Although a large number of encouraging results have been reported in this field, the further clinical application of cell membrane biomimetic nanomaterials in phototherapy against tumors still faces many challenges.
(1) The long-term in vivo safety of phototherapeutic agents, especially inorganic nanomaterials, still needs to be systematically evaluated. 123 Although the cell membrane biomimetic strategy has significantly improved the targeting and biocompatibility of phototherapeutic agents, the non-degradability of nanomaterials and the effects of ablated cell membranes on the body are unknown, limiting the clinical translation of this therapy. Therefore, the development of biodegradable phototherapeutic agents for biomimetic phototherapy may have a higher clinical value.
(2) The cell membrane coating technology is complicated in operation and low in yield, which brings challenges to the scale-up production of biomimetic phototherapeutic agents. 29 Some cells such as RBC, PLT, and white blood cell can be obtained from blood products, but others require large-scale in vitro culture. 124 In addition, cell membrane biomimetic nanomaterials require strict aseptic storage to ensure that they are not affected by chemical and biological contaminants, which greatly increases the resource cost of biomimetic nanomaterial preparation. To expand production and reduce cost, it is necessary to optimize the synthetic conditions, improve reproducibility, and establish relevant standards. Due to the existence of many functional proteins on the surface of cell membranes, they should be stored stably for a long time under suitable conditions in industrial production to prevent protein denaturation and inactivation. It is worth noting that the membrane source should be carefully selected in the development of biomimetic nanomaterials to ensure patient compliance and maximize the therapeutic effect. The use of patient CCM in personalized nanomedicine can enhance immune tolerance and effectively suppress tumor growth. 86 (3) The efficacy of phototherapy mediated by cell membrane biomimetic nanomaterials is still insufficient. On the one hand, the light penetration depth is limited. Even with the NIR laser responsive system, the penetration depth does not exceed 1 cm. 125 For superficial tumors such as skin cancer or oral cancer, localized phototherapy can be achieved with the aid of imaging equipment. For deep tumors, suitable medical equipment such as optical fiber is required to complete the interventional treatment. 126 However, this strategy may cause postoperative infection, causing additional distress to the patient. On the other hand, due to the dense tumor extracellular matrix and elevated interstitial fluid pressure, it is difficult for cell membrane biomimetic phototherapeutic agents to enter the interior of solid tumors, resulting in failure to eradicate tumor lesions and prevent tumor cell metastasis. There is no doubt that synergistic therapies, including CT, RT, and gene therapy, 127 will shine in future cancer treatment strategies. The development of multifunctional cell membrane biomimetic nanomaterials for personalized anti-cancer therapy has important clinical value and broad application prospects. Cell membrane biomimetic nanomaterials are not only widely used in the field of cancer treatment but also have great potential in the treatment of ischemic stroke, myocardial infarction, and articular cartilage damage. 128