Ulex Europaeus Agglutinin-I-Based Magnetic Isolation for the Efficient and Specific Capture of SW480 Circulating Colorectal Tumor Cells.

The efficient and specific capture of circulating tumor cells (CTCs) from patients' peripheral blood is of significant value in precise cancer diagnosis and cancer therapy. As fine targeting molecules, lectins can recognize cancer cells specifically due to the abnormal glycosylation of molecules on the cancer cell membrane and the specific binding of lectin with glycoconjugates. Herein, a Ulex europaeus agglutinin-I (UEA-I)-based magnetic isolation strategy was developed to efficiently and specifically capture α-1,2-fucose overexpression CTCs from colorectal cancer (CRC) patients' peripheral blood. Using UEA-I-modified Fe3O4 magnetic beads (termed MB-UEA-I), up to 94 and 89% of target cells (i.e., SW480 CRC cells) were captured from the cell spiking complete cell culture medium and whole blood, respectively. More than 90% of captured cells show good viability and proliferation ability without detaching from MB-UEA-I. In combination with three-color immunocytochemistry (ICC) identification, MB-UEA-I has been successfully used to capture CTCs from CRC patients' peripheral blood. The experimental results indicate a correlation between CTC characterization and tumor metastasis. Specifically, MB-UEA-I can be applied to screen early CRC by capturing CTCs when served as a liquid biopsy. The presented work offers a new insight into developing cost-effective lectin-functionalized methods for biomedical applications.

Introduction

Circulating tumor cells (CTCs) refer to a general term for various types of tumor cells shed from the primary site into peripheral blood with the ability to metastasize to distant organs and form metastatic tumors.[1−3] Thus, capturing and detecting CTCs may hold great clinical significance for noninvasive tumor profiling, efficacy monitoring, prognosis prediction, and early diagnosis of metastasis like a liquid biopsy.[4−6] Nevertheless, efficient enrichment of high-purity CTCs is a tremendous challenge because of high heterogeneity and intrinsic rarity. Generally, there are only a few-to-several hundred CTCs in 106–109 hematologic cells per milliliter of peripheral blood.[7,8]

To date, researchers have developed numerous methods to highly efficiently capture and detect CTCs, which are mostly in view of their physical properties (density, size, deformability, and adhesion preference)[9−13] and/or biological characteristics (antibody–antigen, aptamer, peptides, and E-selectin).[14−34] For instance, a microfluidic immunoaffinity assay for CTC isolation from breast cancer patients’ blood was proposed by Park et al.[18] The related isolation strategy obviously improves both capture efficiency and recovery of the isolated CTCs, which is very important for performing sensitive downstream assays. Wu et al. designed an efficient and reliable multifunctional nanosphere system with chip assistance for biomarker phenotyping of single heterogeneous CTCs based on the size and specific targeting of human epidermal growth factor receptor 2 (HER2) and the epithelial cell adhesion molecule (EpCAM).[19] A signal amplification detection method based on Ag2S fluorescent nanodots combined with immunomagnetic spheres (IMNs) was proposed by Ding et al., which has been used for the magnetic capture of CTCs with high efficiency (six CTCs per milliliter of the mimic whole blood) as well as ultrasensitive fluorescence labeling.[20] Especially, CTC enrichment based on antibody (e.g., epidermal growth factor receptor (EGFR), EpCAM, HER2, etc.)-modified immunomagnetic beads are the most commonly applied for CTC isolation from the peripheral blood of patients.[25−30] This magnetic isolation method shows several advantages including simple magnetic operation, specific and effective coupling with target cells, and easy integration with consecutive identification and analysis (such as immunocytochemistry (ICC) and Raman imaging). Importantly, the CellSearch system based on magnetic isolation has been approved by the U.S. Food and Drug Administration (FDA), which is used to assess the diagnostic and prognostic efficacy in patients with metastatic colorectal cancer, breast cancer, and prostate cancer.[31,32]

Despite the fact that antibody functionalization has highly improved CTC capture efficiency, the antibody-dependent CTC isolation strategy remains limited in widespread clinical applications due to its bench-to-bench stability, reproducibility, mass production, and high cost.[5,35,36] Besides, antibodies (such as EpCAM) expressed on the surface of CTCs are probably more heterogeneous than initially expected (that is, the epithelial–mesenchymal transition (EMT) leads to not only the decreased expression of epithelial markers (EpCAM) but also increased expression of mesenchymal markers (EGFR))[5,37] and are even missing in certain tumor classes (e.g., sarcoma and melanoma).[38] Thus, other targeting molecules, like DNA aptamers,[39,40] recognition peptides,[41] or specific binding lectins,[42,43] have also been tested in CTC isolation. Compared to antibodies, lectins are ubiquitous in a variety of plants including beans and grains, which are amenable to mass production and play important roles in identification of glycoprotein and the phase of the disease. They have been widely used in cytochemistry, histochemistry, and immunochemistry for the detection and characterization of glycosylated residues and different glycoconjugates in human or animal cells and tissue surfaces.[44,45] For instance, Xie et al. put forward trifunctional nanospheres modified with wheat germ agglutinin to identify DU-145 cells with sialic acid and N-acetylglucosamine expressed on the surface.[42] Therefore, the lectin-modified magnetic isolation strategy could be a promising technique for CTC isolation. Our previous work demonstrates that Ulex europaeus agglutinin-I (UEA-I) can conjugate with α-1,2-fucose on the surface of SW480 cells effectively and selectively[46,47] and thus shows potential for the early diagnosis of colorectal cancer (CRC) as well as cancer phenotype identification as a biomarker.

Herein, a cell-sorting strategy is proposed to efficiently capture and isolate CTCs, which is based on UEA-I-modified Fe3O4 magnetic beads (termed MB-UEA-I). The resulting MB-UEA-I can capture SW480 cells spiked in peripheral blood highly efficiently. In addition, the captured cells show good viability, which satisfies requirements for downstream mutation detection and cell proliferation. Furthermore, the practicability of MB-UEA-I is demonstrated by CTC capture and isolation from unprocessed CRC patients’ blood samples. This work also identifies that primary CRC and metastatic CRC have different CTC characterization.

Experimental Section

Preparation of MB-UEA-I

The carboxyl group of MBs was activated for conjugation with UEA-I through covalent attachment to obtain MB-UEA-I. The experimental details are illustrated in the Supporting Information (SI).

Capturing SW480 Cells in Cell Spiking Samples

One hundred SW480 cells were spiked into 1 mL of complete culture medium (L-15 + 10 v%/v% FBS) and incubated with 0.05 mg of MB-UEA-I at 37 °C with gentle shaking for 45 min. Subsequently, an external magnet was used to separate the MB-UEA-I-captured cells, while the uncaptured cells in supernatants were transferred into a 48-well plate and counted by an inverted microscope. As control experiments, 0.05 mg of MB-UEA-I or blank MBs were incubated with 100 HCT116 cells, 100 NCM460 cells, or 100 SW480 cells spiked in 1 mL of complete culture medium, separately, and the captured cells were separated and counted as previously described. Besides, the Hoechst 33342-prestained SW480 cells, Rhodamine B-prestained HCT116 cells, and Rhodamine B-prestained NCM460 cells were also incubated with MB-UEA-I or blank MBs. After separation, the captured and uncaptured cells were observed via a fluorescence microscope. In addition, MB-UEA-I or MBs after interactions with cells were also observed through scanning electron microscopy (SEM) characterization.

For investigating the capture sensitivity of MB-UEA-I toward SW480 cells, various amounts of SW480 cells were spiked into 1 mL of complete culture medium (L-15 + 10 v%/v% FBS) and incubated with 0.05 mg of MB-UEA-I at 37 °C with gentle shaking for 45 min. After separation, the uncaptured cells in supernatants were transferred into a 48-well plate and counted by an inverted microscope. Furthermore, the capture sensitivity of MB-UEA-I toward SW480 cells spiking in the whole blood sample was also investigated. Various amounts of Hoechst 33342-prestained SW480 cells were spiked into 1 mL of the fresh healthy blood sample. The capture process was carried out as mentioned above but the MB-UEA-I-captured cells were characterized via a fluorescence microscope.

Cell Viability and Proliferation Study

Calcein AM and propidium iodide (PI) were chosen to study the cell viability, while the proliferation study was conducted by reculturing the MB-UEA-I-captured SW480 cells (see the SI for experimental details).

Capture and ICC Identification of SW480 and CTC Cells

The mimic clinical sample was obtained by spiking 10 SW480 cells into 1 mL of the healthy whole blood. Then, 0.05 mg of MB-UEA-I were added to the sample, incubated under gentle shaking at 37 °C for 45 min, and subsequently separated by an external magnet. The three-color ICC was used to recognize MB-UEA-I-captured cells. For the clinical practicability test, blood samples collected from 14 CRC patients before surgery and 5 healthy individuals were separately incubated with 0.05 mg of MB-UEA-I to capture CTCs, which were further identified by ICC.

Results and Discussion

Preparation and Characterization of MBs and MB-UEA-I

As shown in Scheme , UEA-I was modified on the surface of MBs by carbodiimide chemistry through covalent attachment. The as-obtained MB-UEA-I can specifically recognize α-1,2-fucose positive cancer cells, such as SW480 cells and CTCs of primary CRC. The average size of MBs is 1.02 ± 0.06 μm in diameter. After being modified with UEA-I, the MBs exhibit a negligible change in morphology and dispersibility (as shown in Figure a,b). To confirm the combination of MBs with UEA-I, FL-UEA-I was employed to react with MBs. Under 488 nm excitation, MB-UEA-I-FL emits an obvious fluorescence at 520 nm (as shown in Figure c,d). Besides, we figured out the calibration curve of FL-UEA-I and measured the fluorescence spectra of 25 μg of FL-UEA-I before and after reaction with 0.5 mg of MBs. According to the change in fluorescence intensity, we found that the coupling quantity of FL-UEA-I on MBs was about 16 μg mg–1 (as shown in Figure S1).

Scheme 1

Schematic Illustration of CTC Capture and Identification Based on MB-UEA-I

Figure 1

SEM micrographs of (a) MBs and (b) MB-UEA-I, and the fluorescence images (c) and fluorescence spectra (d) of MBs and MB-UEA-I-FL.

Specific Capture of SW480 Cells with MB-UEA-I

The concentration of MB-UEA-I and incubation time were optimized to obtain the maximum capture efficiency. Figure S2a shows that the capture efficiency increases as the concentration of MB-UEA-I increases. More than 90% of spiked SW480 cells are captured when the concentration of MB-UEA-I is higher than 0.05 mg mL–1. In addition, as the incubation time of MB-UEA-I with cells increases, the capture efficiency also increases and more than 90% of SW480 cells can be captured when the incubation time is longer than 45 min (as shown in Figure S2b). Considering that a higher concentration of MB-UEA-I and longer incubation time may result in more nonspecific capture of untargeted cells, 0.05 mg mL–1 MB-UEA-I and an incubation time of 45 min are chosen to capture targeted cells in the following experiments.

For testing their specificity, MB-UEA-I were first incubated with two CRC cell lines (SW480 cells and HCT116 cells) and one normal colorectal cell line (NCM460 cells) under the optimized conditions. In comparison with HCT116 and NCM460 cells, SW480 cells express α-1,2-fucose residues on the cell membrane surface with an aberrantly high level, which have a high binding affinity with UEA-I through the combination with α-1,2-fucosylated lactodifucotetraose-derived neoglycolipid.[46−48] As expected, MB-UEA-I can capture more than 94% of spiked SW480 cells in complete culture medium, which is much higher than those of spiked HCT116 cells (12%) and NCM460 cells (11%) (as shown in Figure ). Moreover, few SW480 cells were captured by the unmodified MBs. Besides, the fluorescence imaging studies of MB-UEA-I- or MB-treated cells also verify the specific binding between MB-UEA-I and SW480 cells (Figure S3). The SEM micrographs in Figure S4 also show the successful attachment of MB-UEA-I on the SW480 cell surface. The results suggest that MB-UEA-I can be used to efficiently and specifically capture SW480 cells.

Figure 2

Capture efficiencies of MB-UEA-I to SW480 cells, HCT116 cells, and NCM460 cells, and the capture efficiency of MBs to SW480 cells. Error bars mean standard deviations (SD, n = 3).

Capturing SW480 Cells Spiked in Mimic CTC Samples

Different amounts (ranging from 3 to 1000) of SW480 cells were added to 1 mL of complete culture medium or healthy whole blood to investigate the capture sensitivity of MB-UEA-I. Under the optimal experimental conditions, the calibration curves were obtained by plotting the linearity of the captured cell numbers against the spiked cell numbers. As shown in Figure a, the linear equations are y1 = 0.94x – 0.92 (R2 = 0.9998) for SW480 cells in complete culture medium and y2 = 0.89x + 0.08 (R2 = 0.9987) for SW480 cells in the whole blood, indicating that capture sensitivity are 94 and 89% in complete culture medium and whole blood, respectively. In particular, four SW480 cells were captured from six SW480 cells spiking 5 mL of whole blood (as shown in Figure S5), demonstrating the high capture sensitivity of MB-UEA-I. These results reveal that the as-constructed MB-UEA-I can be directly applied to sort CTCs of primary CRC in clinical samples. Furthermore, the captured cells from SW480 cells spiking the whole blood were identified with ICC by the use of Hoechst 33342 (a marker for nucleus), Alexa Fluor 488-labeled anti-CK19 (a marker for epithelial cells) monoclonal antibody, and Alexa Fluor 594-labeled anti-CD45 (a marker for white blood cells (WBCs)) monoclonal antibody.[49] Hoechst 33342+/CK19+/CD45– cells are recognized as SW480 cells, while Hoechst 33342+/CK19–/CD45+ are sorted as WBCs. As shown in Figure b, SW480 cells and WBCs can be effectively separated by MB-UEA-I. The result confirms that the cell-sorting method based on MB-UEA-I enables efficient and reliable detection of CTCs.

Figure 3

(a) Capture sensitivities of MB-UEA-I to SW480 cells. Error bars mean SD (n = 3). (b) Typical fluorescence images of the cells identified by ICC ((i) SW480 cells spiking the whole blood sample, (ii) WBC in the supernatant, and (iii) MB-UEA-I-captured SW480 cells). Scale bars are 10 μm.

Cell Viability Analysis and Proliferation Study

The viability of MB-UEA-I-captured SW480 cells was analyzed through Calcein AM and PI costaining. Most captured cells exhibit green fluorescence emission, and the viability rate is calculated to be 93%, which is similar to the viability rate (95%) of the control cells (Figure S6). Besides, the proliferation ability of MB-UEA-I-captured cells was further explored. The captured cells with MB-UEA-I on the surface can well adhere to the culture plates and exhibit strong proliferation ability (as shown in Figure ). In particular, the proliferation rate of MB-UEA-I-captured SW480 cells is as rapid as that of normal cultured SW480 cells (Figure S7). The results distinctly illustrated that the captured cells retained both the cell viability and proliferative ability, enabling trusty downstream studies, for example, gene expression, mutation, sequencing, etc.

Figure 4

Fluorescence images of MB-UEA-I-captured SW480 cells after re-cultured. From left to right: adhered to the plate, proliferated, just reached confluence, reached confluence after one passage and five passages, respectively. The scale bars are 50 μm.

CTC Capture and Identification from Clinical CRC Patients’ Peripheral Blood

The clinical practicability of MB-UEA-I was demonstrated by directly profiling CTCs in peripheral blood samples, which were collected from 14 CRC patients before surgery and 5 healthy individuals. The Hoechst 33342+/CK19+/CD45– cells characterized by ICC were classified as CTCs. As illustrated in Figure a and Table S1, the CTC numbers ranged from 3 to 20 cells mL–1 blood from CRC patients, while there are no CTCs in the blood samples of healthy donors. As a typical sample, the fluorescence images of 16 CTCs captured from the blood sample of the No. 5 patient are shown in Figure b. In addition, the numbers of CTCs (mean values of 13 ± 4 CTCs mL–1) in blood samples from primary CRC patients (nos. 1–7) are higher than that in blood samples (mean values of 5 ± 2 CTCs mL–1) from metastatic CRC patients (nos. 8–14). The phenomenon is caused by decreasing the fucosylation level of the CRC cell membrane with tumor progression, i.e., the cell membrane is highly fucosylated at an early stage of CRC and gradually defucosylated through genetic mutation in the metastasis stage of CRC.[50] The result indicates the characterization of CTCs associated with disease progression and demonstrates that the MB-UEA-I-based magnetic isolation strategy is a promising tool for the early diagnosis of CRC. But it still needs to be combined with other diagnostic techniques to determine the CRC stage accurately.

Figure 5

(a) Quantitative detection of CTCs from 14 CRC patients (nos. 1–14) and 5 healthy individuals (nos. 15–19) blood samples. (b) Fluorescence images of the MB-UEA-I-captured CTCs from patient no. 5 by three-color ICC identification. Scale bars are 5 μm.

Conclusions

In summary, a convenient and rapid lectin-based magnetic isolation strategy is developed to highly efficiently and specifically capture CTCs of CRC through strong binding affinity between UEA-I and α-1,2-fucose. The captured cells are completely suitable for downstream studies such as CTC culture, gene expression, and mutation detection because the as-proposed MB-UEA-I magnetic isolation platform exhibits a negligible effect on cell viability and proliferation ability. Heterogeneous populations/features of CTCs at different disease progression stages have been demonstrated by profiling CTCs in whole blood samples from 14 CRC patients, which may offer helpful information for clinical decisions. Therefore, the lectin-based magnetic isolation strategy demonstrates the practical use of lectin-based isolation of rare cells and has a great promise to be translated to the clinic as a cost-effective liquid biopsy method.