A blood cell repelling and tumor cell capturing surface for high-purity enrichment of circulating tumor cells

Tong Li ab, Nan Li ab, Yao Ma ab, Yun-Jie Bai ab, Cheng-Mei Xing ab and Yong-Kuan Gong *ab
aKey Laboratory of Synthetic and Natural Functional Molecule Chemistry of Ministry of Education, College of Chemistry and Materials Science, Northwest University, Xi'an 710127, Shaanxi, P. R. China. E-mail: gongyk@nwu.edu.cn
bInstitute of Materials Science and New Technology, Northwest University, Xi'an 710127, Shaanxi, P. R. China

Received 5th August 2019 , Accepted 4th September 2019

First published on 6th September 2019

The detection of circulating tumor cells (CTCs), an approach considered to be “liquid biopsy”, is crucial in cancer diagnosis, monitoring and prognosis. However, the extremely large number of blood cells challenges the rare CTC isolation and enrichment. In this report, a red blood cell membrane mimetic surface (CMMS) is fabricated on material-independent substrates to repel blood cell adhesion. Meanwhile, tumor cell targeting ligands, folic acid (FA) and an arginine-glycine-aspartic acid (RGD) peptide, are tethered on the CMMS to give the decorated surface (CMMS–FA–RGD) tumor cell capture ability. The CMMS is composed of a mussel-inspired self-adhesive polydopamine layer and a covalently anchored non-fouling or anti-cell-adhesion layer of a phosphorylcholine zwitterion polymer and poly(ethylene glycol) (PEG). The protruding ends of the PEG chains of the anchored CMMS are further coupled with FA and RGD ligands to endow the tumor cells with specific binding. Furthermore, all the components of the step-by-step constructed surfaces are quantitatively controllable for optimizing the non-specific cell repellence and tumor cell binding performances. Thus, the delicately engineered CTC capture surface enhances the HeLa cell enrichment factor to 19[thin space (1/6-em)]000-fold by repelling the adhesion of >99.999% blood cells, resulting in high capture efficiency (91%) and capture purity (89%) from the spiked whole blood samples. This substrate independent tumor cell capture and blood cell repellent surface modification strategy may provide a facile, versatile and cost-effective technology solution for more efficient cancer diagnosis and targeted therapy.

1. Introduction

Circulating tumor cells (CTCs) are cancer cells circulating in the blood stream after escaping from original or metastatic tumors. They appear in the peripheral blood of cancer patients even when the cancers are in the early stage and play critical roles in cancer metastasis, which contributes to 90% of the cancer-related deaths.1–3 CTCs have been investigated as an important class of cancer biomarkers for better understanding of cancer metastasis and easier monitoring of cancer progression and treatment response.4–8 Therefore, various strategies have been developed to selectively capture CTCs from blood samples.8–11 The most prevalent detection and isolation techniques employ immunoaffinity targeting of an epithelial cell adhesion molecule (EpCAM), which is a membrane protein expressed by CTCs but not by blood cells.12,13 For example, some excellent approaches employ anti-EpCAM antibody-based capture at microchannel walls of microfluidic devices,14 or on magnetic beads (CellSearch, Veridex, Warren, PA, USA).15,16 However, the extremely low abundance of CTCs (with a range of 5–1281 CTCs amongst 5 billion erythrocytes and 10 million leukocytes per mL) in the peripheral blood of cancer patients is technically challenging for CTC isolation and analysis for clinical applications.5–8 For instance, although the CellSearch system is the only CTC detection equipment on the market approved by the Food and Drug Administration, it achieves a limited CTC capture efficiency of ∼40% with less than 0.5% purity from the spiked blood samples, originating from the nonspecific capture of 1000–3000 white blood cells (WBCs) per milliliter of the samples.17,18 Obviously, the nonspecific binding of leukocytes via physisorption on the CTC capture surfaces is a major factor of the low purity and the capture efficiency is strongly dependent on the effective antibody immobilization. Therefore, the relevant surface modifications should focus on both improving affinity-based CTC capture and preventing nonspecific cell adhesion.

The affinity-based CTC capture surfaces are mainly nanostructured to provide a larger surface area for antibody immobilization and present nanotopography that promotes cell–surface interaction. For example, surfaces with nanowire array structures enhanced the CTC capture efficiency by several times compared to flat surfaces.19,20 Similarly, graphene oxide nanosheet, polymer nanofiber, and nanodot modifications were all reported to enhance CTC capture.21–24 Although nanostructured surfaces significantly improve the capture efficiency, they encounter a critical challenge of low capture purity. The results have shown that nanoroughened surfaces promote the nonspecific adhesion of blood cells and decrease the CTC capture purity from 84 to 14%.25,26 Moreover, the anti-EpCAM antibody immobilization on a surface via covalent chemistry enables limited yield, modest uniformity, and poor reproducibility.27 Furthermore, the poor stability and high cost of antibodies also constrain the dissemination and translation of CTC-based diagnostics, especially in a low-resource environment.8,28

To address these limitations, we propose a novel strategy including repelling blood cell adhesion and capturing CTCs by their receptor protein ligands, which can be highly effective, much cheaper and more stable than the anti-EpCAM antibody. In order to repel or prevent the nonspecific adhesion of blood cells, a blood compatible and anti-fouling/anti-adhesion surface is preferred, since leukocyte adhesion is a natural process of the body dealing with inflammatory response, and the adhesion is enhanced when WBCs are activated by various stimuli, including the physical chemistry of the surface.29 For example, biomimetic cell-membrane-camouflaged immunomagnetic beads reduced homologous WBC interaction and enhanced the isolation purity of CTCs to 96.9% from 66.5%,30 and a cell-membrane mimetic lipid coating on the nanopillar-based surface improved the CTC capture purity to 71.3% from 3.2%.31 Recently, red blood cells (RBCs) coated with a tumor-targeting folic acid (FA) ligand and magnetic nanoparticles were used to capture CTCs.32 Although the separation process was complicated, the RBC mediated CTC enriching method significantly enhanced the capture purity to 75% from 20%. The excellent effect is attributed to the anti-adhesion or anti-fouling nature of the RBC membrane, since RBCs circulate for long half-life (120 days) without any adhesion with other blood cells and vascular endothelial cells. Moreover, it has been confirmed that the zwitterionic phosphorylcholine outer surface of the cell membrane lipid bilayers plays a crucial role in resisting bioadhesion and biofouling.33–37

Inspired by the non-fouling nature of the RBC outer membrane, we have synthesized amphophilic zwitterion copolymers and prepared the copolymer micelles with a cell outer membrane zwitterion interface in aqueous medium.37–39 Excitingly, a cell membrane mimetic micelle has showed an extremely long circulating half-life (90.5 h) in the bloodstream of New Zealand Rabbits,37 demonstrating the excellent anti-bioadhesion or anti-capture performance in the body. On the other hand, tumor cell targeting ligands, folic acid (FA) and an arginine-glycine-aspartic acid (RGD) peptide, are more stable and much cheaper small molecules than the anti-EpCAM antibody, and worthy of being explored for capturing folate receptor (FR) overexpressed tumor cells,40–43 and integrin αvβ3 overexpressed tumor cells.44 Therefore, the combination of RBC membrane mimetic anti-adhesion with tumor cell specific binding properties on one surface will be promising in preventing blood cell adhesion and enhancing the selective capture of tumor cells. Furthermore, the strategy of directly and selectively capturing CTCs from blood may provide the mostly simplified and lowest cost platform for CTC detection, as well as for developing cancer targeting therapeutic and diagnostic reagents.

Based on the above advantages, the red blood cell membrane mimetic surface (CMMS) is constructed for effectively preventing/repelling blood cell adhesion (Fig. 1). Meanwhile, FA and cyclic-RGD ligands are coupled on the CMMS of anchored HOOC-PEG-COOH chains as antenna-like linkers to form the CMMS–FA–RGD surface for an enhanced CTC capture surface. Furthermore, the blood cell repellence and tumor cell specific capture performances of CMMS–FA–RGD are all optimized quantitatively by tuning the surface densities.

image file: c9tb01649j-f1.tif
Fig. 1 Conceptual schematics of the fabrication, blood cell repellence and tumor cell capture of a ligand decorated cell membrane mimetic surface (CMMS–FA–RGD).

2. Experimental

2.1. Materials

3-Hydroxytyramine (dopamine) hydrochloride, bovine serum albumin (BSA), and bovine plasma fibrinogen (Fg) were purchased from Sigma-Aldrich. Dicarboxyl poly(ethylene glycol) (HOOC-PEG-COOH) with Mw of 2000 and 5000 was purchased from Suzhou Nords Parson's Pharmaceutical Technology Co., Ltd. 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) were purchased from Thermo Fisher Scientific Inc. 2-Methacryloyloxyethyl phosphorylcholine (MPC, 99% purity) was purchased from Nanjing Joy-Nature Science & Technology Development Institute, China. p-Nitrophenoxycarbonyl-oxyethyl methacrylate (NPCEMA) was synthesized according to the method reported by Konno.45 The cyclic (Arg-Gly-Asp-D-Phe-Lys) peptide (c-RGD) was purchased from China Peptides Co., Ltd. Mouse connective tissue fibroblast line (L929), mouse embryonic fibroblast line (3T3), human breast cancer cell line (MCF-7) and human cervical carcinoma cell line (HeLa) cells were purchased from Procell Life Science & Technology Co., Ltd (Wuhan, China). Folic acid bearing a reactive amino group (FA-NH2) was synthesized according to a previous report.38 Deionized water used in experiments was purified using a Millipore water purification system with a resistivity of 18.2 MΩ cm−1 at 298 K. All other chemicals and reagents were of analytical grade and were used without further purification unless otherwise indicated.

A random copolymer bearing cell outer membrane phosphorylcholine zwitterions and active ester groups (PMEN) was synthesized using monomers of MPC and NPCEMA according to a reported method.46 The copolymer was purified by dialysis and the number average molar mass determined by GPC was 19 kDa. The chemical structure and 1H-NMR spectrum of PMEN are shown in Fig. S1 (ESI).

2.2. Surface plasmon resonance (SPR) measurements

The SPR technique was chosen to quantitatively monitor the surface fabrication and protein adsorption processes. All the measurements were conducted on a SR7500DC dual channel SPR system (Reichert, USA). After the SPR sensor chip was attached at the base of the prism, a stable baseline was established by injecting a degassed buffer solution (10 mM K2HPO4/KH2PO4, 150 mM NaCl, 0.005% (v/v) Tween 20, pH 7.4) over the chip for at least 20 min. Sample solutions preloaded into a 500 μL sample loop were then injected into the flow cell using a syringe pump for a specified time. Then the deposited chip surface was rinsed with the buffer for about 30 min to wash off the loosely associated molecules and obtain a baseline of the coated layer surface. The surface density or thickness of the deposited layer was calculated from the SPR signal change between the baselines by simply assuming 1 μRIU = 0.1 ng cm−2 and a density of 1.0 g cm−3 for the deposited polymers.47

By repeating the coating deposition procedure with different sample solutions, single and mixed layer coatings with designed thicknesses were prepared. Similarly, the protein adsorption process and the adsorption amount (1 μRIU = 0.1 ng cm−2) on the coated chip surfaces were also measured in real-time using the SPR system.

Protein interaction parameters on the SPR chip surfaces were also determined from the SPR sensorgrams with different protein concentrations. Each of the sensorgrams was fit according to a global curve fitting analysis based on a Langmuir model (1[thin space (1/6-em)]:[thin space (1/6-em)]1 binding) to determine the kinetic adsorption rate constant ka and dissociation rate constant kd.48 The equilibrium constant KA = ka/kd and equilibrium dissociation constant KD of the protein adsorption were calculated from the two rate constants.

2.3. Cell membrane mimetic surface (CMMS) fabrication and ligand decoration

The CMMS was firstly constructed on the SPR sensor chip on-line by a step by step deposition procedure49 with some modifications. Step 1, 2.0 mg mL−1 dopamine in pH 8.5 Tris–HCl solution was injected over the chip surface at 25 °C for 20–40 min to form a 1–2 nm thick polydopamine (PDA) layer for anchoring the next layers. Step 2, 2 mg mL−1 HOOC-PEG-COOH solution activated by EDC (0.1 mg mL−1) and NHS (0.05 mg mL−1) in pH 6.5 phosphate buffered saline (PBS buffer) was injected at 50 °C for 20–40 min to anchor the PEG chains (PDA/PEG) with a surface density of 200–600 ng cm−2. Step 3, 5 mg mL−1 PMEN in pH 8.5 PBS buffer was injected at 60 °C for 6–8 hours until the PMEN layer (PDA/PEG/PMEN, simplified as the CMMS) reached a maximum thickness. Step 4, activation solution of EDC (12 mg mL−1) and NHS (16 mg mL−1) in pH 6.5 PBS buffer was injected at 50 °C for 20 min, and then 0.2 mg mL−1 FA-NH2 dissolved in DMSO[thin space (1/6-em)]:[thin space (1/6-em)]PBS = 1[thin space (1/6-em)]:[thin space (1/6-em)]4 (v/v) mixed solvents was injected at 50 °C for 20–60 min to form an FA ligand tethered surface (CMMS–FA) with an FA surface density of 8–50 ng cm−2. Alternatively, 0.1 mg mL−1 c-RGD in DMSO[thin space (1/6-em)]:[thin space (1/6-em)]PBS = 1[thin space (1/6-em)]:[thin space (1/6-em)]4 (v/v) mixed solvents was injected at 50 °C for 20–60 min to form c-RGD ligand tethered surface (CMMS–RGD). Step 5, the second ligand (RGD or FA) was further coupled at the remaining PEG ends on the CMMS–FA or CMMS–RGD surface to form dual ligands targeting surface (CMMS–FA–RGD) for enhanced tumor cell binding. All the flow rate of the injected coating solutions was 10 μL min−1, and stable baselines were run before and after each of the injections.

The CMMS was also constructed and decorated off-line on other substrates, including glass, silicon wafers, polydimethylsiloxane and polystyrene cell culture plates, according to the fabrication conditions optimized by the SPR on-line experiments.

2.4. Water contact angle measurements

Water contact angles (WCA) of the coated sample surfaces were recorded with a video-based contact angle measuring system (DSA25, KRÜSS, Germany). 2 μL pure water was dropped on the surfaces and the WCA values were calculated by averaging at least five data at different locations for each sample. The WCA values were expressed as means ± SD.

2.5. XPS analysis

The elemental composition of CMMS coated and ligand decorated surfaces was measured by X-ray photoelectron spectroscopy (XPS) with a spectrometer using monochromatic Al Kα radiation (200 W, 12 kV, 1486.68 eV) (PHI5000 Versaprobe III). All the spectra were collected at an electron take-off angle of 90° from the surface under vacuum. The binding energy scale was calibrated relative to the C1s peak (284.8 eV) from hydrocarbons on the coatings. The high-resolution spectra of N1s, P2p and Si2p were fitted using a Shirley background subtraction and a series of Gaussian peaks using the XPSPEAK software.

2.6. Morphology characterization

The CMMS and ligand decorated surfaces fabricated on glass substrates were observed using a Multi Mode 8 atomic force microscope (AFM) (Bruker Corporation, USA). 3D images were scanned by the tapping mode in air using a silicon tip on a nitride cantilever with a nominal spring constant of 0.40 N m−1. The surface roughness was assessed by analyzing the roughness parameter, Rq (root mean square roughness). The morphology of the cells adhered on each of the surfaces was observed using a cold field emission scanning electron microscope (SU8010, Hitachi). The adhered cells on a surface were firstly immobilized in 4% paraformaldehyde solution for 30 min. Then, the cells were dehydrated successively in each of the 50%, 70%, 80%, 90% and 100% ethanol solutions for 20 min. SEM images were collected after freeze drying.

2.7. Protein adsorption measurements

BSA and Fg as nonspecific protein models and FRα as a tumor cell specific protein were selected to evaluate the nonspecific protein adsorption resistance and specific protein binding ability on CMMSs. All the protein adsorption measurements were performed on the modified sensor chips using the SR7500DC SPR instrument. Single protein solution in PBS buffer (1.0 mg mL−1 BSA or Fg, or 0.5 μg mL−1 FRα) was injected over the chips for 20 min at a flow rate of 20 μL min−1 at 25 °C.

2.8. Cell adhesion and capture

Single cell adhesion and capture. HeLa, MCF-7, L929, 3T3 and WBC cell suspensions (1 × 105 cells per mL) were seeded respectively on the surfaces of differently modified samples placed in the wells of CMMS coated 24-well cell culture plates after coming into contact with PBS buffer for 30 min. The samples were cultured in a CO2 cell incubator (SERIES II, Thermo Scientific) at 37 °C for 20 min, 40 min and 60 min, respectively. The adhered cells were stained in 2 μL mL−1 Syto 9 solution for 20 min in the dark. After removing the dye solution and washing with PBS, the reserved cells were defined as captured cells and were observed/recorded with an inverted fluorescence microscope (Ti-U, Nikon). In order to compare the cell adhesion/capture ability on the different surfaces, all the rinsing conditions were fixed at a 20 μL min−1 rate using a microfluidic injection pump. The adhered cells on a sample were counted at least at 3-positions on each of the 6-parallel surfaces.
Capture of tumor cells from mixed cell samples. In order to evaluate the capability of enriching rare tumor cells by our simple surface capture strategy, a mixed cell suspension was prepared with Hochest 33342-prestained normal cells (blue fluorescence) and Syto 9 prestained tumor cells (green fluorescence). The mixed cell suspension (e.g., L929[thin space (1/6-em)]:[thin space (1/6-em)]HeLa = 100[thin space (1/6-em)]:[thin space (1/6-em)]1) was then dropped on the CMMS–FA–RGD surface placed in the wells of CMMS coated 24-well cell culture plates. After being incubated at 37 °C in a CO2 incubator (SERIES II, Thermo Scientific) for 20 min, 40 min and 60 min, respectively, the adhered cells after rinsing with PBS were recorded using an inverted fluorescence microscope (Ti-U, Nikon). The tumor cell capture efficiency (E) and capture purity (P) were calculated according to the equations below.
image file: c9tb01649j-t1.tif(1)
image file: c9tb01649j-t2.tif(2)
Capture of tumor cells from mimetic clinical samples. 20–9400 HeLa and MCF-7 tumor cells were spiked into 1000 μL whole blood and captured on the CMMS–FA–RGD surface by the same procedure used in the mixed cell samples. Whole blood was drawn from healthy volunteers after obtaining informed consent for research testing that complied with the ethical regulations under an Institutional Review Board approved protocol at the Northwest University. For immunocyto-chemistry identification, the adhered cells were incubated with 4% paraformaldehyde and 0.1% Triton-X 100 for 10 min respectively for fixation and permeabilization. Then the cells were stained with FITC-labeled anti-CK19 mAb, Hoechst 33342 and PE-labeled anti-CD45 mAb at 37 °C for 30 min, respectively. The immobilized cells were observed using a Ti-U fluorescence microscope. The captured tumor cells were confirmed to be positive for Hoechst 33342 and CK19, and negative for CD45.
Cell viability test. The adhered cells were incubated with 4.5 μM PI and 2 μM calcein AM at 37 °C in a CO2 incubator (SERIES II, Thermo Scientific) for 30 min. After washing, the cells were observed and counted with an inverted fluorescence microscope (Ti-U, Nikon) to quantitatively analyze the viability.

2.9. Statistical analysis

All of the data are expressed as means ± SD. The significance of differences between two samples was determined by an unpaired Student's t test using Microsoft Excel 2010 and P values less than 0.05 were considered statistically significant, and P values less than 0.01 were considered very significant.

3. Results and discussion

3.1. Fabrication and physicochemical characterization of CMMS–FA–RGD surfaces

The cell membrane mimetic surface (CMMS), FA and cyclic RGD (c-RGD) ligand decorated CMMS (CMMS–FA–RGD) were constructed via a step by step procedure illustrated in Fig. S1. The mussel inspired polydopamine (PDA) adhesive layer was firstly coated on a material-independent substrate to provide reactive amino groups for functionalization.50 Carboxyl group end caped poly(ethylene glycol) (HOOC-PEG-COOH) and phosphorylcholine zwitterion copolymer bearing nitrophenyloxycarbonyl (active ester) side chains (PMEN, Fig. S2, ESI) were then covalently anchored on the PAD coating to form PAD/PEG/PMEN (CMMS) according to a previous report.49 The surface grafted PEG antenna chain ends were further coupled with FA and c-RGD ligands by amidation to endow tumor cells with a specific binding ability.

The successful fabrication of CMMS–FA–RGD was qualitatively demonstrated by the rational changes in water contact angle (WCA), atomic force microscopy (AFM) 3D morphology and X-ray photoelectron spectroscopy (XPS) spectra on each of the coated layer surfaces. The WCA of a surface is very sensitive to the hydrophilicity/hydrophobicity of the materials. As shown in Fig. 2A, most of the WCAs increased after depositing a more hydrophobic molecular layer onto the previous surfaces. Exceptionally, the covalently anchored PMEN polymer layer renders the surface much hydrophilic than the parent PEG surface, since the zwitterionic PMEN is strongly hydrophilic while PEG is amphiphilic.51 The obvious changes in AFM 3D morphology on each of the coated layer surfaces (Fig. 2C) also support the step by step molecular layer deposition. Furthermore, the newly appeared peaks in both of the high resolution N1s and P2p XPS spectra (Fig. S3, ESI), together with the changes in elemental composition (Table 1) of the step by step coated glass surfaces provide strong evidence for the successful fabrication. By considering the nanometer thin layer formation and 10 nm detection depth of the XPS measurements, the changes in surface element compositions shown in Table 1 is reasonable. More importantly, all the fabrication steps were quantitatively investigated using a very sensitive surface plasmon resonance (SPR) instrument. Each of the coated layer thickness (Table 1) and surface density were derived from their SPR sensorgrams (Fig. S4, ESI). The layer thickness measured by SPR was consistent with that measured by ellipsometry.49 It is worth noting that the sensitive and quantitative SPR method enables us to control the fabrication and optimize the performance of all the CMMSs. Since all the fabrication steps were carried out in aqueous solutions, and the CMMSs could be immobilized on-line or off-line onto a variety of material surfaces with different shapes and sizes,49 this facile, versatile and cost-effective tumor cell specific binding surface construction strategy could be further extended to other ligands or tumor cell targeting molecules for capturing different kinds of tumor cells.

image file: c9tb01649j-f2.tif
Fig. 2 (A) Water contact angles of the surfaces coated with PDA, PEG, PMEN, FA, RGD, and FA + RGD as the outer layers during the CMMS fabrication. Bars represent mean ± SD (n = 5). (B) Molecular structures of the coating materials PMEN, c-RGD and FA-NH2. (C) AFM 3D morphology of (a) glass substrate, (b) PDA, (c) PDA/PEG, (d) PDA/PEG/PMEN (CMMS), (e) CMMS–FA, (f) CMMS–RGD, and (g) CMMS–FA–RGD. The red colored numbers show the surface roughness (Rq), and all the SD data are less than 0.04 nm. The PEG is 5 kDa HOOC-PEG-COOH. All the images were scanned on the surfaces of 1.0 μm × 1.0 μm.
Table 1 The outer later thickness and elemental composition of the step by step modified surfaces
Surface Layer thicknessa (nm) Relative molar ratio of surface elementb (mol%)
C N O P Si
a Measured by SPR. b Measured by XPS analysis.
Bare glass 14.21 0 61.55 0 24.24
PDA 1.51 52.87 5.26 34.43 0 7.45
PDA/PEG 0.52 60.28 4.69 29.90 0 5.13
PDA/PEG/PMEN (CMMS) 3.02 58.50 4.97 31.12 1.29 4.12
CMMS–FA 0.25 60.26 5.82 29.02 1.28 3.62
CMMS–RGD 0.18 62.86 5.59 27.44 1.06 3.05
CMMS–FA–RGD 0.33 60.01 5.70 29.63 1.03 3.63

3.2. Blood cell repelling performance of the CMMS

Blood cell adhesion is a complicated process, which is effected by plasma protein adsorption. It is well known that the amount of plasma proteins adsorbed on a surface has a dominant effect on the blood compatibility of the materials, since protein absorption is the first event that triggers later bioresponses, including cell activation, adhesion and blood coagulation.52,53 In this study, the most abundant blood serum albumin (BSA) and adhesive fibrinogen (Fg) were selected as representative proteins for evaluating the performance of blood protein adhesion resistance on CMMSs. As shown in Fig. 3A, the protein adsorption amount could be reduced to a few ng cm−2 with decreased PEG content on the fabricated CMMSs. Moreover, after the –COOH end group was reacted or changed to –OH, the protein adsorption resistance of the grafted PEG chains could be further enhanced significantly.49 The excellent protein adsorption resistance of this easily coated nanometer thick CMMS is comparable with that of difficultly prepared zwitterionic polymer brushes by a grafting-from approach.54,55
image file: c9tb01649j-f3.tif
Fig. 3 (A) BSA and Fg protein adsorption amounts on the CMMSs containing different weight percentages of PEG measured by SPR. The mixed polymer coatings were fabricated on-line by covalently anchoring different amounts of activated HOOC-PEG-COOH (5 kDa), and then anchoring PMEN on the PDA/PEG surface for 6 h to approach saturated deposition. The bars represent mean ± SD (n = 3). (B and C) Microscopic images of whole blood cells adhered for 40 min on a polystyrene (PS) cell culture plate and CMMS, respectively. (D) Adhered blood cell density changes with adhesion time on the PS and CMMS under the same conditions.

The success of this zwitterionic polymer anchoring approach depends on the ability to achieve high surface packing densities, which is important for a zwitterionic surface to achieve super-low fouling properties.49,56 The surface density of the phosphorylcholine zwitterion groups was calculated to be 2.0 zwitterions per nm2 from the PMEN polymer composition and deposited mass. By considering the very thin thickness (3.02 nm, shown in Table 1) of the PMEN layer and large size of the hydrated phosphorylcholine group, the phosphorylcholine zwitterions formed a densely packed structure on the CMMS, very similar to that of phosphorylcholine zwitterions in the cell outer membrane.

The blood cell repelling performance of the CMMS was demonstrated by immersing the surface in whole blood for a designed period of time and then rinsing with PBS buffer. By comparing the results of adhered cells on a polystyrene (PS) cell culture plate surface under the same conditions (Fig. 3B), there are almost no adhered blood cells on the CMMS (Fig. 3C). Further detailed quantitative comparison (Fig. 3D) showed that the CMMS has 100[thin space (1/6-em)]000 fold resistance or repellence as that of the PS surface for blood cell adhesion. This means that the cell adhesion interaction on the CMMS is 100[thin space (1/6-em)]000 fold weaker than that on the PS surface. The dramatically decreased whole blood cell adhesion could be simply understood by the excellent anti-biofouling or anti-bioadhesion nature of the CMMS.

3.3. Enhancing tumor cell capture efficacy by the synergistic effect of ligands

FA ligand binding and optimization. The FA ligand is frequently linked at the end of PEG corona of PEGylated nanocarrier surfaces to enhance the tumor cell targeting efficacy.43 To improve the tumor cell–substrate interaction and enhance the tumor cell capture performance, it is highly desirable that the surfaces have cell specific adhesion. We tethered FA molecules on the CMMS by PEG linkers to impart the surface tumor cell capture ability. The surface density of FA ligands and the linker length of PEG chains on the CMMS were all optimized by SPR experiments to maximize the binding amount of folate receptor α (FRα) protein, which is overexpressed extracellularly on a variety of human cancers.40–43 As shown in Fig. 4, the adsorbed FRα amount increased remarkably as the PEG chain length increases. The increased length of the linker chain provides the ligand more degrees of freedom and longer distance to bind with FRα, thus enhancing the efficacy of FA-FRα interaction. On the other hand, a further increase of the PEG chain length results in few chain ends for FA coupling (with the same mass of PEG), or an increase of the PEG chain concentration results in more adsorption of non-specific proteins (Fig. 3A). Thus, we selected 5 kDa PEG as the FA linker for further optimizing specific interaction. Interestingly, the strongest interaction appeared at relatively low FA group density (around 15 ng cm−2 or 0.20 FA ligand per nm2), and then decreased significantly with the increase of FA density as shown in Fig. 4d. This result is much better than the case where more than 10% of FA ligand is required for the detectable tumor targeting effect.57,58 We have demonstrated that low concentration of FA ligands on micelles can give obvious tumor cell targeting performance, and the FA groups need to be released from burying under the hydrophobic core or shielding by the hydrophilic shell during nanocarrier formation.38,39 The CMMS can efficiently prevent the FA from being buried or shielded due to the densely packed cell outer membrane mimetic zwitterion structure, which repels the approach of the slightly hydrophobic FA group to the strongly hydrophilic surface. Meanwhile, the FA groups tethered at the PEG linker ends also have more opportunity to aggregate by the hydrophobic interaction with the increased surface densities. As the aggregated FA large structure does not fit into the ligand-binding pocket on FRα proteins,59 the increased FA density higher than 27 ng cm−2 results in a significantly reduced binding amount of FRα as shown in Fig. 4d. Therefore, it is important to optimize the surface composition for minimizing non-specific interactions and maximizing specific binding. By employing a sensitive and quantitative SPR technique, the step by step fabrication and performance of the CMMS were optimized. As listed in Table S1 (ESI), FA ligand (15.2 ng cm−2) decorated CMMSs showed much stronger binding/adsorption constant with FRα (KA = 3.45 × 109 M−1) than the non-specific BSA and Fg (KA: 1.25 × 104 and 2.11 × 105 M−1, respectively). The at least 4 orders of magnitude difference in the KA demonstrates highly specific binding with FRα on the optimized CMMS–FA. Furthermore, the FRα binding constant (3.54 × 109 M−1) is also 20-times higher than a recently reported KA value (1.7 × 108 M−1) for the binding of a well-designed multivalent FA-conjugated polyamidoamine dendrimer designed multivalent FA-conjugated polyamidoamine dendrimer G5(FA)6 on the FRα (0.26 FRα nm−2) immobilized SPR chip surface,60 suggesting a more promising surface for FRα receptor-targeted capturing and treatment of tumor cells.
image file: c9tb01649j-f4.tif
Fig. 4 FR α protein adsorption amounts on the surfaces decorated with different densities of FA groups tethered by different lengths of PEG linkers. The bars represent mean ± SD (n = 3).
Tumor cell capture efficiency and the synergistic effect of ligands. The tumor cell capture efficiency of FA decorated CMMSs under optimized conditions was firstly evaluated using human cervical carcinoma cell lines (HeLa). As shown in Fig. 5a, the capture efficiency reached a maximum value (39%) when the FA surface density was around 20 ng cm−2. This result is consistent with that measured from the FRα adsorption amount (Fig. 4d), supporting the effectiveness of the relationship between the FRα adsorption amount and the FA surface density determined by SPR measurements. On the other hand, although the FA ligand density effect was optimized on the CMMS–FA surfaces, the maximum HeLa cell capture efficiency (39%) was low. Thus, tumor cell integrin αvβ3 targeting cyclic RGD was decorated on the CMMS and CMMS–FA surfaces to prepare CMMS–RGD and CMMS–FA–RGD, respectively. The covalently bound RGD density was also measured by on-line SPR under the same reaction conditions as the off-line preparation. As shown in Fig. 5b, although the HeLa cell capture efficiency of CMMS–RGD is higher than that of CMMS–FA, the 78% efficiency is still not satisfactory. However, the FA and RGD dual ligand decorated surface CMMS–FA–RGD showed excellent capture efficiency (∼99%, Fig. 5c) for HeLa cells. Compared with the relatively low efficiency of CMMS–FA and CMMS–RGD, the extremely high efficiency of the CMMS–FA–RGD surface bearing an additional small amount of RGD (11 ng cm−2) is attributed to the synergistic action with the FA ligand for capturing the tumor cells. Since the contribution of the 11 ng cm−2 single RGD should be 47% calculated from the 13 ng cm−2 single RGD shown in Fig. 5b, the 17% extra efficiency of the CMMS–FA–RGD surface bearing 26 ng cm−2 FA and 11 ng cm−2 RGD can definitely be attributed to the synergistic effect of RGD and FA ligands. In order to confirm the synergistic effect of the dual ligands for capturing tumor cells, human breast cancer cell (MCF-7) and HeLa cell adhesion on differently decorated CMMSs was compared. Clearly, both HeLa and MCF-7 tumor cells were captured by the ligand effect (Fig. S5, ESI). They showed extra increased capture efficiencies by the additional small amount of RGD (11 ng cm−2) on the CMMS–FA–RGD surface cultured for 20 to 60 min (Fig. S6, ESI), suggesting that the synergistic effect of RGD and FA ligands exists in different tumor capture applications at different adhesion time points. The strong binding interaction with HeLa cells was further demonstrated in a microfluidic channel coated with CMMS–FA–RGD for the capture and enrichment of the fluorescent HeLa cells (Video S1, ESI).
image file: c9tb01649j-f5.tif
Fig. 5 HeLa cell capture efficiency on (a) CMMS–FA, (b) CMMS–RGD and (c) CMMS–FA–RGD surfaces decorated with different densities of ligands by changing the reaction time. The ligand densities at different reaction times were measured by SPS under the same conditions and were denoted on top of the capture efficiency bars. All the surfaces were immersed in the cell suspension for 60 min. The bars represent mean ± SD (n = 3).

The tumor cell capture efficiency and capture purity were then evaluated via mixed cell adhesion experiments by using mouse connective tissue fibroblast cell line (L929) and mouse embryonic fibroblast line (3T3) cells as the normal cell models. Representative microscopic images (Fig. S7, ESI) before and after rinsing with PBS solution demonstrated the selective capture of tumor cells on the CMMS–FA–RGD surface. As shown in Fig. 6, the capture efficiency for HeLa cells from a 200 μL L929[thin space (1/6-em)]:[thin space (1/6-em)]HeLa = 100[thin space (1/6-em)]:[thin space (1/6-em)]1 mixed cell suspension on the CMMS–FA–RGD surface increases rapidly from 35 ± 3% at 20 min incubation to 86 ± 2% at 60 min. Meanwhile, the capture purity decreases slowly from 100% at 20 min to 81 ± 3% at 60 min. Interestingly, the HeLa cell capture efficiency and capture purity at 60 min were all increased significantly to 91 ± 3% and 89 ± 3% respectively by extending the volume of the mixed cell suspension to 1000 μL (shown in Fig. 8b). These better results could be explained by more chance in larger volume to bind selectively with the tumor cells. These results are quite good compared with that obtained by using size-dictated immunocapture chips,61 or immunomagnetic nanospheres.62,63 Moreover, if we consider that the data were obtained simply on a flat surface without any assistance of complicated magnetic nanoparticles or micropost-patterned structure enhancement,5,31,63 this CMMS–FA–RGD should be one of the most efficient tumor cell capturing flat surfaces and the CTC capture efficiency and purity could be further improved on micro- or nano-structured substrates. Furthermore, this simply and efficiently captured CTC exhibits ∼90% viability (Fig. S8, ESI), which is important for molecular phenotyping, genotyping, and further biological characterization.

image file: c9tb01649j-f6.tif
Fig. 6 (a) Fluorescence microscopic images of HeLa cell capture on the CMMS–FA–RGD surface from 200 μL mixed L929 (blue) and HeLa (green) cell suspensions (L929[thin space (1/6-em)]:[thin space (1/6-em)]HeLa = 100[thin space (1/6-em)]:[thin space (1/6-em)]1) incubated for 20, 40 and 60 min. (b) HeLa cell capture efficiency and purity change with incubation time. (c) HeLa and MCF-7 cell capture efficiency and capture purity on the CMMS–FA–RGD surface from mixed cell solutions (L929[thin space (1/6-em)]:[thin space (1/6-em)]tumor cell = 100[thin space (1/6-em)]:[thin space (1/6-em)]1) incubated for 60 min. The bars represent mean ± SD (n = 3).

In order to further understand the selective binding of the CMMS–FA–RGD surface with tumor cells, cell adhered morphology was recorded using both scanning electronic microscope (SEM) and optical microscope (Fig. S9, ESI). The morphologies of different cells on the differently modified surfaces were compared. As demonstrated in Fig. 7, the L929, 3T3 and tumor cells all showed obvious pseudopods on the bare glass, but all the cells including tumor cells did not show observable pseudopods on the CMMS, suggesting greatly reduced cell binding interaction of the CMMS. By introducing tumor cell targeting FA and RGD ligands on the CMMS, the decorated CMMS–FA, CMMS–RGD and CMMS–FA–RGD surfaces showed obviously enhanced adhesion force with HeLa and MCF-7 cells by their extended adhesion area and increased number of pseudopods. In contrast, all normal cells, including L929, 3T3 and white blood cells (WBC) in whole blood, did not demonstrate a significant morphology change by the ligands. Furthermore, the selective capture of tumor cells was also observed under slow shaking and microfluidic conditions. As all tumor cells have specific ligand receptor proteins on their membrane, ligand decorated antifouling surfaces could capture the tumor cells selectively under both flowing and static conditions.

image file: c9tb01649j-f7.tif
Fig. 7 SEM images of cells adhered from their suspensions for 40 min on different surfaces after drying.

3.4. Tumor cell enrichment and detection from spiked whole blood

After certification of the normal cell repellence and tumor cell capture ability of CMMS–FA–RGD surfaces from mixed cell samples, we further investigate the surface performance of capturing spiked tumor cells in whole blood samples. For demonstrating the strong binding and high capture capacity, a 1 cm2 CMMS–FA–RGD surface was immersed in 1 mL whole blood spiked with a large number (9400) of prestained HeLa cells. As shown in Fig. 8, the spiked HeLa cells were captured on the CMMS–FA–RGD surface with high efficiency (91.5%) due to the strongly enhanced synergistic binding interactions of FA and RGD ligands. More excitingly, the captured tumor cells were bonded so strong that they were motionless on the CMMS–FA–RGD surface during washing (Video S2, ESI), which removed almost all the blood cells and reserved all the adhered HeLa cells as shown in Fig. 8a by comparing the microscopic images before and after washing. Thus, the capture purity was improved up to ∼100% by more careful rinsing in some cases (Fig. S10 and S11, ESI). The high capture purity could be attributed to the more than 99.999% prevention of the whole blood cell adhesion compared with that of the bare PS cell culture plate surfaces (Fig. 3). Thus, the enrichment factor, derived by the division of the cell ratios on the surface (HeLa[thin space (1/6-em)]:[thin space (1/6-em)]WBC = 18[thin space (1/6-em)]:[thin space (1/6-em)]1 in Fig. 8) and in the spiked whole blood (9400[thin space (1/6-em)]:[thin space (1/6-em)]1 × 107, by considering 1.0 mL whole blood contains 1 × 107 WBCs), is increased to 19[thin space (1/6-em)]000-fold from the HeLa cell spiked whole blood samples, suggesting the excellent performance of our method for highly efficient enrichment of tumor cells in whole blood samples. To evaluate the applicability of the surface for measuring unlabeled tumors, a 100 HeLa cell spiked whole blood sample was tested. The immunocytochemical identification results (Fig. 8c) indicate clearly the capture of the tumor cells and no observable WBCs. In the case of blood samples containing 20 tumor cells, the fluorescence microscopic observation of the captured tumor cells need to be performed with a 4× objective lens.
image file: c9tb01649j-f8.tif
Fig. 8 (a) Microscopic images of HeLa cell capture on the CMMS–FA–RGD surface from 9400 HeLa cells (stained green) spiked in 1.0 mL whole blood (WBC[thin space (1/6-em)]:[thin space (1/6-em)]HeLa = 1060[thin space (1/6-em)]:[thin space (1/6-em)]1) incubated for 60 min. (b) HeLa cell capture efficiency and capture purity on the CMMS–FA–RGD surface from 1.0 mL of the spiked whole blood and from 1.0 mL mixed cell suspensions (L929[thin space (1/6-em)]:[thin space (1/6-em)]HeLa = 100[thin space (1/6-em)]:[thin space (1/6-em)]1) incubated for 60 min. (c) Captured cells were stained with Hoechst 33342 (blue), PE-labeled anti-CD45 mAb (red) and FITC-labeled anti-CK19 mAb (green). The CK19 positive and CD45 negative cells are confirmed to be tumor cells. The bars represent mean ± SD (n = 3).

It is worthy of note that attractive results were obtained simply by direct adhesion from the spiked whole blood samples. The selective capture strategy of the FA and RGD ligand decorated CMMS flat surface showed high efficiency, simple device and operation over current immunomagnetic nanosphere assisted methods, which improved the enrichment of MCF-7 cells to 4368-fold via complicated multiple processes to remove a huge number of erythrocytes and white blood cells (WBCs) in the blood sample before the tumor cell capture by anti-EpCAM antibody and anti-EGFR antibody cocktail modified immunomagnetic nanospheres.62 In addition to the demonstrated high efficacy, simple process and device, our ligand decorated CMMS does not involve any antibody and thus can significantly reduce the cost.8,28 Furthermore, the fabrication of this ligand decorated CMMS is versatile on material independent substrates with a variety of ligands, scalable and applicable to large area devices through the mild reactions in aqueous solutions. Obviously, such a CMMS–FA–RGD surface can also be constructed at the interfaces of nano-drugs and nano-diagnostic reagents to enhance their selective targeting performances. All these advantages pave a promising way for developing low cost, easy operation, but high efficiency CTC detection and tumor cell targeted therapy techniques, including integration into and improving the current microfluidic platform,64 filtration-enriched capture,65 or tumor cell targeted nanodrug systems.66

4. Conclusions

A blood cell repelling and tumor cell capturing flat surface was fabricated by a facile step by step coating strategy, which could be applied on a variety of materials due to the universal adhesion of a mussel inspired polydopamine mediated adhesion layer. The FA and RGD ligand decorated cell membrane mimetic surface CMMS–FA–RGD could capture HeLa and MCF-7 tumor cells directly from the cell containing samples. The tumor cell capture efficiency could be enhanced by optimizing the surface ligand density, linker length, and the synergistic effect of different kinds of ligands. As a result, this tumor cell capture surface offers distinct advantages, such as (a) versatile fabrication on material independent substrates; (b) scalable and applicable to large area devices due to the mild reactions in aqueous solutions; (c) simple tumor cell capture assay, the direct adhesion in whole blood samples and then a rinsing procedure makes it possible to deploy in point-of-care testing; (d) high capture efficiency and purity, due to the excellent anti-adhesion for normal cells and targeted binding for tumor cells; and (e) inexpensive, due to the use of chemically synthesized small ligands, without any antibody or immunomagnetic nanospheres. In summary, the combination of the blood cell repellent surface based on cell membrane mimetic modification and tumor cell capture from ligand decoration provides an attractive solution for the design and versatile construction of cost-effective materials and devices for cancer diagnosis and targeted therapy.

Conflicts of interest

The authors declare no conflict of interest.


The authors are thankful for the funds of the National Natural Science Foundation of China (NSFC, grant no. 21374087 and 21774100).


  1. J. Massagué and A. C. Obenauf, Nature, 2016, 529, 298–306 CrossRef PubMed.
  2. C. L. Chaffer and R. A. Weinberg, Science, 2011, 331, 1559–1564 CrossRef CAS PubMed.
  3. B. Weigelt, J. L. Peterse and L. J. van't Veer, Nat. Rev. Cancer, 2005, 5, 591–602 CrossRef CAS PubMed.
  4. E. I. Galanzha, E. V. Shashkov, T. Kelly, J. W. Kim, L. Yang and V. P. Zharov, Nat. Nanotechnol., 2009, 4, 855–860 CrossRef CAS PubMed.
  5. S. Nagrath, L. V. Sequist, S. Maheswaran, D. W. Bell, D. Irimia, L. Ulkus, M. R. Smith, E. L. Kwak, S. Digumarthy, A. Muzikansky, P. Ryan, U. J. Balis, R. G. Tompkins, D. A. Haber and M. Toner, Nature, 2007, 450, 1235–1239 CrossRef CAS PubMed.
  6. L. De Mattos-Arruda, J. Cortes, L. Santarpia, A. Vivancos, J. Tabernero, J. S. Reis-Filho and J. Seoane, Nat. Rev. Clin. Oncol., 2013, 10, 377–389 CrossRef CAS PubMed.
  7. C. Alix-Panabières and K. Pantel, Nat. Rev. Cancer, 2014, 14, 623–631 CrossRef PubMed.
  8. Z. Shen, A. Wu and X. Chen, Chem. Soc. Rev., 2017, 46, 2038–2056 RSC.
  9. S. Sharma, R. Zhuang, M. Long, M. Pavlovic, Y. Q. Kang, A. Ilyas and W. Asghar, Biotechnol. Adv., 2018, 36, 1063–1078 CrossRef CAS PubMed.
  10. J. M. Jackson, M. A. Witek, J. W. Kamande and S. A. Soper, Chem. Soc. Rev., 2017, 46, 4245–4280 RSC.
  11. B. J. Green, T. Saberi Safaei, A. Mepham, M. Labib, R. M. Mohamadi and S. O. Kelley, Angew. Chem., Int. Ed., 2016, 55, 1252–1265 CrossRef CAS PubMed.
  12. T. M. Becker, N. J. Caixeiro, S. H. Lim, A. Tognela, N. Kienzle, K. F. Scott, K. J. Spring and P. De Souza, Int. J. Cancer, 2014, 134, 2523–2533 CrossRef CAS PubMed.
  13. C. Alix-Panabieres and K. Pantel, Clin. Chem., 2013, 59, 110–118 CrossRef CAS PubMed.
  14. L.-L. Wu, M. Tang, Z.-L. Zhang, C.-B. Qi, J. Hu, X.-Y. Ma and D.-W. Pang, Anal. Chem., 2018, 90, 10518–10526 CrossRef CAS PubMed.
  15. E. S. Johnson, R. K. Anand and D. T. Chiu, Anal. Chem., 2015, 87, 9389–9395 CrossRef CAS PubMed.
  16. P. G. Schiro, M. Zhao, J. S. Kuo, K. M. Koehler, D. E. Sabath and D. T. Chiu, Angew. Chem., Int. Ed., 2012, 51, 4618–4622 CrossRef CAS PubMed.
  17. K. Xiong, W. Wei, Y. Jin, S. Wang, D. Zhao, S. Wang, X. Gao, C. Qiao, H. Yue, G. Ma and H. Y. Xie, Adv. Mater., 2016, 28, 7929–7935 CrossRef CAS PubMed.
  18. E. A. Punnoose, S. Atwal, W. Liu, R. Raja, B. M. Fine, B. G. M. Hughes, R. J. Hicks, G. M. Hampton, L. C. Amler, A. Pirzkall and M. R. Lackner, Clin. Cancer Res., 2012, 18, 2391–2401 CrossRef CAS PubMed.
  19. G.-S. Park, H. Kwon, D. W. Kwak, S. Y. Park, M. Kim, J.-H. Lee, H. Han, S. Heo, X. S. Li, J. H. Lee, Y. H. Kim, J.-G. Lee, W. Yang, H. Y. Cho, S. K. Kim and K. Kim, Nano Lett., 2012, 12, 1638–1642 CrossRef CAS PubMed.
  20. S.-K. Lee, D.-J. Kim, G. Lee, G.-S. Kim, M. Kwak and R. Fan, Biosens. Bioelectron., 2014, 54, 181–188 CrossRef CAS PubMed.
  21. H. J. Yoon, T. H. Kim, Z. Zhang, E. Azizi, T. M. Pham, C. Paoletti, J. Lin, N. Ramnath, M. S. Wicha, D. F. Hayes, D. M. Simeone and S. Nagrath, Nat. Nanotechnol., 2013, 8, 735–741 CrossRef CAS PubMed.
  22. J. Sekine, S.-C. Luo, S. Wang, B. Zhu, H.-R. Tseng and H.-H. Yu, Adv. Mater., 2011, 23, 4788–4792 CrossRef CAS PubMed.
  23. N. Zhang, Y. Deng, Q. Tai, B. Cheng, L. Zhao, Q. Shen, R. He, L. Hong, W. Liu, S. Guo, K. Liu, H.-R. Tseng, B. Xiong and X.-Z. Zhao, Adv. Mater., 2012, 24, 2756–2760 CrossRef CAS PubMed.
  24. W. Han, B. A. Allio, D. G. Foster and M. R. King, ACS Nano, 2010, 4, 174–180 CrossRef CAS PubMed.
  25. W. Chen, S. Weng, F. Zhang, S. Allen, X. Li, L. Bao, H. W. R. Lam, J. A. Macoska, S. D. Merajver and J. Fu, ACS Nano, 2013, 7, 566–575 CrossRef CAS PubMed.
  26. W. Chen, S. G. Allen, A. K. Reka, W. Qian, S. Han, J. Zhao, L. Bao, V. G. Keshamouni, S. D. Merajver and J. Fu, BMC Cancer, 2016, 16, 614 CrossRef PubMed.
  27. L. S. Wong, F. Khan and J. Micklefield, Chem. Rev., 2009, 109, 4025–4053 CrossRef CAS PubMed.
  28. G. K. Olivier, A. Cho, B. Sanii, M. D. Connolly, H. Tran and R. N. Zuckermann, ACS Nano, 2013, 7, 9276–9286 CrossRef CAS PubMed.
  29. K.-C. Chang, D. F. J. Tees and D. A. Hammer, Proc. Natl. Acad. Sci. U. S. A., 2010, 97, 11262–11267 CrossRef PubMed.
  30. L. Rao, Q.-F. Meng, Q. Huang, Z. Wang, G.-T. Yu, A. Li, W. Ma, N. Zhang, S.-S. Guo, X.-Z. Zhao, K. Liu, Y. Yuan and W. Liu, Adv. Funct. Mater., 2018, 28, 1803531 CrossRef.
  31. H.-Y. Lou, W. Zhao, L. Hanson, C. Zeng, Y. Cui and B. Cui, Langmuir, 2017, 33, 1097–1104 CrossRef CAS PubMed.
  32. D. M. Zhu, L. Wu, M. Suo, S. Gao, W. Xie, M. H. Zan, A. Liu, B. Chen, W. T. Wu, L. W. Ji, L. B. Chen, H. M. Huang, S. S. Guo, W. F. Zhang, X. Z. Zhao, Z. J. Sun and W. Liu, Nanoscale, 2018, 10, 6014–6023 RSC.
  33. Y. Dang, M. Quan, C. M. Xing, Y. B. Wang and Y. K. Gong, J. Mater. Chem. B, 2015, 3, 2350–2361 RSC.
  34. Q. Xue, H. Cao, F. Meng, M. Quan and Y. K. Gong, J. Membr. Sci., 2017, 528, 1–11 CrossRef CAS.
  35. Y. Iwasaki and K. Ishihara, Sci. Technol. Adv. Mater., 2012, 13, 064101 CrossRef PubMed.
  36. S. Jiang and Z. Cao, Adv. Mater., 2010, 22, 920–932 CrossRef CAS PubMed.
  37. J. Zhao, Y. D. Chai, J. Zhang, P. F. Huang, K. Nakashima and Y. K. Gong, Acta Biomater., 2015, 16, 94–102 CrossRef CAS PubMed.
  38. H. T. Jiang, K. Ding, F. N. Meng, L. L. Bao, Y. D. Chai and Y. K. Gong, J. Mater. Chem. B, 2016, 4, 5464–5474 RSC.
  39. K. Ding, R. Li, Y. Ma, N. Li, T. Zhang, X. C. M. Xing, H. T. Jiang and Y. K. Gong, Langmuir, 2019, 35, 1257–1265 CrossRef CAS PubMed.
  40. L. E. Kelemen, Int. J. Cancer, 2006, 119, 243–250 CrossRef CAS PubMed.
  41. C. Luan, H. Wang, Q. Han, X. Ma, D. Zhang, Y. Xu, B. Chen, M. Li and Y. Zhao, ACS Appl. Mater. Interfaces, 2018, 10, 21206–21212 CrossRef CAS PubMed.
  42. C. S. Kue, A. Kamkaew, K. Burgess, L. V. Kiew, L. Y. Chung and H. B. Lee, Med. Res. Rev., 2016, 36, 494–575 CrossRef CAS PubMed.
  43. L. Nie, F. Li, X. Huang, Z. P. Aguilar, Y. A. Wang, Y. Xiong, F. Fu and H. Xu, ACS Appl. Mater. Interfaces, 2018, 10, 14055–14062 CrossRef CAS PubMed.
  44. K. Chen and X. Chen, Theranostics, 2011, 1, 189–200 CrossRef CAS PubMed.
  45. T. Konno, J. Watanabe and K. Ishihara, Biomacromolecules, 2004, 5, 342–347 CrossRef CAS PubMed.
  46. Y. K. Gong, L. P. Liu and P. B. Messersmith, Macromol. Biosci., 2012, 12, 979–985 CrossRef CAS PubMed.
  47. M. Tanaka, T. Sawaguchi, Y. Sato, K. Yoshioka and O. Niwa, Tetrahedron Lett., 2009, 50, 4092–4095 CrossRef CAS.
  48. N. J. de Mol and M. J. E. Fischer, Handbook of Surface Plasmon Resonance, The Royal Society of Chemistry, Cambridge, UK, 2008, vol. 5, p. 123 Search PubMed.
  49. C. M. Xing, F. N. Meng, M. Quan, K. Ding, Y. Dang and Y. K. Gong, Acta Biomater., 2017, 59, 129–138 CrossRef CAS PubMed.
  50. H. Lee, S. M. Dellatore, W. M. Miller and P. B. Messersmith, Science, 2007, 318, 426–430 CrossRef CAS PubMed.
  51. Z. Cao and S. Jiang, Nano Today, 2012, 7, 404–413 CrossRef CAS.
  52. L. Vroman, J. S. Mattson and C. A. Smith, Science, 1974, 184, 585–586 CrossRef CAS PubMed.
  53. H. S. Lee, S. J. Stachelek, N. Tomczyk, M. J. Finley, R. J. Composto and D. M. Eckmann, J. Mater. Chem. B, 2013, 1, 6382–6391 RSC.
  54. Y. Chang, Y. J. Shih, C.-J. Lai, H. H. Kung and S. Jiang, Adv. Funct. Mater., 2013, 23, 1100–1110 CrossRef CAS.
  55. H. Chen, J. Yang, S. Xiao, R. Hu, S. M. Bhaway, B. D. Vogt, M. Zhang, Q. Chen, J. Ma, Y. Chang, L. Li and J. Zheng, Acta Biomater., 2016, 40, 62–69 CrossRef CAS PubMed.
  56. W. Feng, J. L. Brash and S. Zhu, Biomaterials, 2006, 27, 847–855 CrossRef CAS PubMed.
  57. J. H. Lee, A. Sahu, C. Jang and G. Tae, J. Controlled Release, 2015, 209, 219–228 CrossRef CAS PubMed.
  58. Z. Tang, D. Li, H. Sun, X. Guo, Y. Chen and S. Zhou, Biomaterials, 2014, 35, 8015–8027 CrossRef CAS PubMed.
  59. C. Chen, J. Ke, X. E. Zhou, W. Yi, J. S. Brunzelle, J. Li, E. L. Yong, H. E. Xu and K. Melcher, Nature, 2013, 500, 486–489 CrossRef CAS PubMed.
  60. P. T. Wong, D. Chen, S. Tang, S. Yanik, M. Payne, J. Mukherjee, A. Coulter, K. Tang, K. Tao, K. Sun, J. R. Baker Jr. and S. K. Choi, Small, 2015, 11, 6078–6090 CrossRef CAS PubMed.
  61. M. G. Ahmed, M. F. Abate, Y. Song, Z. Zhu, F. Yan, Y. Xu, X. Wang, Q. Li and C. Yang, Angew. Chem., Int. Ed., 2017, 56, 10681–10685 CrossRef CAS PubMed.
  62. X. Y. Ma, L. L. Wu, L. Chen, Y. H. Qin, J. Hu, M. Tang, C. M. Xu, C. B. Qi, Z. L. Zhang and D. W. Pang, ACS Appl. Nano Mater., 2018, 1, 4019–4027 CrossRef CAS.
  63. Z. Li, G. Wang, Y. Shen, N. Guo and N. Ma, Adv. Funct. Mater., 2018, 28, 1707152 CrossRef.
  64. Y. Xiao, M. Wang, L. Lin, L. Du, M. Shen and X. Shi, Mater. Chem. Front., 2018, 2, 891–900 RSC.
  65. E. Pailler, M. Oulhen, F. Billiot, A. Galland, N. Auger, V. Faugeroux, C. Laplace-Builhé, B. Besse, Y. Loriot, M. Ngo-Camus, M. Hemanda, C. R. Lindsay, J.-C. Soria, P. Vielh and F. Farace, BMC Cancer, 2016, 16, 477 CrossRef PubMed.
  66. L. Qiu, T. Chen, I. Öçsoy, E. Yasun, C. Wu, G. Zhu, M. You, D. Han, J. Jiang, R. Yu and W. Tan, Nano Lett., 2015, 15, 457–463 CrossRef CAS PubMed.


Electronic supplementary information (ESI) available: Fig. S1–S10, Table S1, Videos S1 and S2 for tumor cell capture and blood cell removal on the CMMS–FA–RGD surface. See DOI: 10.1039/c9tb01649j
These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2019