Lipid-insertion enables targeting functionalization of erythrocyte membrane-cloaked nanoparticles †

RBC membrane-cloaked polymeric nanoparticles represent an emerging nanocarrier platform with extended circulation in vivo . A lipid-insertion method is employed to functionalize these nanoparticles without the need for direct chemical conjugation. Insertion of both folate and the nucleolin-targeting aptamer AS1411 shows receptor-speci ﬁ c targeting against model cancer cell lines.

A major goal in engineering nanocarriers for systemic drug delivery is to achieve long circulation half-lives, 1 as increased residence time in the bloodstream can improve the pharmacokinetic prole of therapeutic cargoes and allow a greater chance for the nanocarriers to reach the desired location through either passive 2,3 or active 4-6 targeting mechanisms. Ongoing search for new and effective ways to construct longcirculating nanoparticles has introduced numerous stealth functionalization strategies. While the use of synthetic polymers represents the current golden standard for prolonging particle residence time in circulation, recent efforts have drawn inspiration from nature to bestow immune-evasive properties on nanoparticles. 7,8 Bridging the complex surface biochemistry of nature's carriers, RBCs, with the versatile cargo-carrying capacity of polymeric nanoparticles, an RBC membrane-cloaked nanoparticle (RBC-NP) platform represents a new class of bioinspired nanocarriers with long-circulating capability. 9,10 In this system, a top-down approach is used to coat polymeric nanoparticles with natural RBC membranes, which possess a litany of immunomodulatory proteins responsible for RBCs' extraordinarily long survival in circulation. [11][12][13][14] This membrane cloak acts to camouage the nanoparticle in vivo to evade immune attacks. 9 Consisting entirely of biocompatible and biodegradable materials, the RBC-NPs have signicant potential for drug delivery applications.
In applying RBC-NPs for disease treatments, particularly against cancers, target-selectivity is a desirable feature that promises minimization of off-target side effects. [15][16][17][18][19] Cancer targeting has been made possible by carrier functionalization with ligands that target overexpressed tumor antigens, and numerous chemical conjugation techniques employing carboxyl-, amine-, or sulydryl-based chemistry have been used to decorate synthetic nanocarriers with targeting ligands. [20][21][22][23][24][25][26] In the case of biologically derived carriers such as RBC-NPs, however, the presence of biological components on the particle surfaces demands a non-disruptive functionalization strategy since the immune evasion capabilities of cellular membranes are predicated upon having fully functional proteins. In order to address the issue of chemical denaturation, we report a lipidinsertion approach for functionalization of RBC-NPs that exploits the uidity of bilayered RBC lipid membranes and precludes RBC membrane exposure to chemical reactions. As shown in Fig. 1, targeting moieties are incorporated onto RBC membranes through the aid of lipid tethers. The physical insertion of ligand-linker-lipid conjugates into the RBC membranes produces functionalized RBC membranes without damaging the existing surface proteins. Using two differently sized ligands, a small molecule folate (MW $ 441 Da) and a nucleolin-targeting aptamer AS1411 (MW $ 9000 Da), we demonstrate that the lipid-insertion technique can be applied to targeting ligands of different length scales. Following preparation of RBC-NPs with the functionalized RBC membranes, their receptor-specic targeting ability is veried in model cancer cell lines in vitro.
To demonstrate that lipid-tethered ligands can be spontaneously incorporated onto RBC membranes, a FITC-linker-lipid (excitation/emission ¼ 495/519 nm) conjugate was used for a proof-of-concept test, where FITC was linked to a lipid molecule through a short polyethylene glycol chain (MW $ 2000 Da). In the study, 40 mg of FITC-PEG-lipid was rst incubated with emptied red blood cells (RBC ghosts) collected from 1 mL of whole mouse blood for 30 min. The RBC ghosts were then centrifuged, washed with PBS, reconstituted, and examined using ow cytometric analysis ( Fig. 2A). Compared to unmodied RBC ghosts, modi-ed membrane ghosts had a signicantly higher signal under the FITC channel. Visualization by uorescence microscopy further conrmed the localization of the lipid-tethered FITC on the RBC membranes, as the microscopy image displayed strong FITC signals outlining the exterior of the RBC ghosts (Fig. 2B). To characterize the lipid-insertion efficiency and saturation level, varying amounts of FITC-PEG-lipid were incubated with the membrane ghosts followed by membrane purication through centrifugation. Fluorescence quantication showed that the retained FITC uorescence increased with the initial FITC-PEGlipid input (Fig. 2C). Saturation was observed as the retained FITC-PEG-lipid approached a plateau at approximately 40 mg per mL of RBCs. Based on the RBC concentration and the molecular weight of the ligand, it is estimated that each RBC ghost contains about 800 000 FITC-PEG-lipid. Fig. 2C also demonstrates the adjustability of ligand density on the membranes by controlling the lipid-tethered ligand input.
Upon conrming the incorporation of FITC onto the RBC ghosts, the membrane materials were then used to prepare FITC-modied RBC-NPs. Following a previously reported protocol, 9 the FITC-modied RBC membrane ghosts were extruded to form $100 nm vesicles, which were then mixed and extruded with 70 nm PLGA particles to generate FITC-modied RBC-NPs. For comparison, RBC-NPs coated with unmodied RBC membranes were also prepared. Characterization of the two particles by dynamic light scattering (DLS) showed similar physicochemical properties between the FITC-modied and unmodied RBC-NPs (Fig. 2D). Both particles were approximately 80 nm in mean diameter and with a zeta potential below À25 mV. Scanning electron microscopy (SEM) further demonstrated the similarity between the two particle types (Fig. 2E), both of which were spherical in morphology and exhibited monodisperse population distributions. To conrm the colocalization of lipid-tethered ligands with the polymeric cores, an in vitro uorescence colocalization study was conducted by loading DiD dye (excitation/emission ¼ 644/663 nm) into the polymeric cores. Following cellular uptake, signicant overlap was observed between the DiD-specic red punctates and the FITC-specic green punctates (Fig. 2F). The colocalization pattern conrms the presence of lipid-tethered FITC on the surface of the polymeric cores, demonstrating successful preparation of ligand-modied RBC-NPs.
Aer validating the lipid-insertion method for RBC-NP functionalization using lipid-tethered uorescent probes, particle modication with cancer-targeting ligands was explored. A small molecule ligand, folate (MW $ 441 Da), which has a similar molecular weight to FITC, was rst examined. Folate-functionalized nanocarriers have broad applicability as folate receptors are overexpressed on several types of cancers. 27 Upon receptor-mediated binding, folate-functionalized nanocarriers can deliver their cargoes intracellularly through an endocytic uptake pathway. The benet of folate-induced cancer targeting has been demonstrated on several nanocarrier platforms, 28,29 and thus its incorporation onto RBC-NPs would improve the particles' utility in cancer drug delivery.
To prepare folate-functionalized RBC-NPs, a commercially available folate-PEG-lipid conjugate was used (Fig. 3A). Folatefunctionalized RBC-NPs were prepared using RBC ghosts inserted with folate-PEG-lipid. Since the targeting ability of folate-functionalized nanoparticles has already been wellestablished, the KB cell line, a model cancer cell line overexpressing the folate receptor that is commonly used to evaluate folate targeting, 30-32 was used to conrm successful functionalization of the RBC-NPs. To assess for folate-mediated differential uptake, the cells were cultured in folate-free media and incubated with folate-functionalized RBC-NPs, non-targeted RBC-NPs, or folate-functionalized RBC-NPs together with 1 mM of free folate. The cells from each sample were then detached, washed, and analyzed using ow cytometry ( Fig. 3B and C). Compared to the non-targeted RBC-NPs, incubation of folate-functionalized RBC-NPs resulted in an 8-fold increase in cellular uptake. Conjoint incubation with folate-functionalized RBC-NPs and 1 mM of free folate yielded a similar level of cellular uptake as compared to non-targeted RBC-NPs, which indicated that the increased uptake of folate-functionalized RBC-NPs was receptor-specic. Fluorescence microscopy visualization of particle uptake further conrmed the results observed from ow cytometry. As shown in Fig. 3D, uorescence from the DiD dye encapsulated inside the particles was only observed in cells incubated with folate-functionalized RBC-NPs in the absence of free folate molecules. To demonstrate that the targeting effects were exclusive to cells overexpressing the folate receptor, a negative cancer cell line, A549, 33 that does not overexpress the folate receptor was incubated with either unmodied or folate-functionalized RBC-NPs. No increased uptake was observed for the targeted nanoparticles compared to the unmodied nanoparticles using both ow cytometry and uorescence imaging (ESI Fig. S1 †). It is also important to note that no cytotoxicity was observed for the RBC-NPs when incubated with human umbilical vein endothelial cells (HUVECs), a normal cell line, at the concentrations used in these studies (ESI Fig. S2 †). Overall, the results conrm the receptor-specic targeting capability of folate-functionalized RBC-NPs.
To demonstrate that the lipid-insertion method can be applied to targeting ligands of different length scales, particle functionalization with a nucleolin-targeting oligonucleotide, AS1411 aptamer (MW $ 9000 Da), was also tested. Oligonucleotide-based targeting agents, or aptamers, are a versatile class of ligands that can be customized against specic receptors through affinity screening. 34 AS1411, a 26-mer DNA aptamer with the sequence GGT GGT GGT GGT TGT GGT GGT GGT GG, has shown targeting capability against several cancer cell types owing to frequent overexpression of surface nucleolin on cancerous cells. [35][36][37] The ligand has also been applied for the preparation of cancer-targeted nanoparticles, 38,39 thus its integration onto the RBC-NPs would greatly benet the utility of the platform.
To incorporate aptamers onto RBC-NPs via lipid-insertion, a lipid-tethered AS1411 was rst prepared. AS1411 aptamers containing a 3 0 thiol modier was reduced using tris- [2-carboxyethyl]phosphine (TCEP) and conjugated to lipid-PEG-maleimide via maleimide-sulydryl chemistry (Fig. 4A). Aer purication, the AS1411-PEG-lipid conjugates were used to prepare RBC-NPs following the aforementioned procedures. In a cellular uptake study using a surface nucleolin expressing breast cancer cell line, MCF-7, 35 differential targeting was observed. Flow cytometry analysis revealed that the targeted nanoparticles induced a two-fold increase in cellular uptake as compared to the non-targeted RBC-NPs ( Fig. 4B and C). The uptake enhancement by the targeted nanoparticles was also conrmed to be receptor-specic, as blocking by free AS1411  Fluorescence microscopy of MCF-7 cells incubated with AS1411-functionalized RBC-NPs, non-targeted RBC-NPs, and AS1411-functionalized RBC-NPs together with free AS1411. A fluorescent probe DiD was loaded inside the RBC-NPs for visualization (red) and cellular nuclei were stained with DAPI (blue). Scale bars ¼ 25 mm. reduced the particle uptake to the level of the non-targeted nanoparticles. Fluorescence imaging corroborated the ow cytometry results with the AS1411-fuctionalized RBC-NPs showing much greater uptake than the non-targeted and blocked samples (Fig. 4D). The results demonstrate that the lipid-insertion method can be applied to relatively large targeting ligands (e.g., MW $ 9000 Da) with a molecular weight larger than that of the lipid anchor (MW $ 748 Da).

Conclusions
In summary, by employing a lipid-insertion technique for the functionalization of biological membranes, targeted RBC-NPs were successfully prepared with two different types of targeting ligands. Through the aid of lipid tethers and the dynamic conformation of membrane bilayers, targeting ligands can be spontaneously incorporated onto the RBC-NP platform without exposing the biological membranes to chemical reactions. The robustness and simplicity of this functionalization scheme can enable a wide array of functionalized RBC-NPs for specic disease treatments. In addition, the technique can be generalized to help improve the applicability of emerging biologically inspired nanocarriers possessing complex surface chemistry. The capability to control and adjust ligand density through the lipid-insertion technique also provides versatility for platform optimization. Future studies are warranted to examine the in vivo implications of ligand functionalization on RBC-NPs. The targeted RBC-NPs reported in the present work possess significant potential for cancer treatments as they integrate nature's immune-evasive moieties with cancer-binding ligands.