Abstract
Background/Aim: Circulating tumor cells (CTCs) have been shown to have a correlation to metastasis and prognosis of patients with cancer. The enumeration and downstream gene analysis of CTCs have attracted efforts for personalized medicine. However, enumeration and phenotypic profiling in most capture devices are challenging due to the rarity and heterogeneity of CTCs. Here, we report an aptamer-cocktail strategy coupled in nano-microfluidic chip for enhancing isolation performance and characterizing the phenotypes of CTCs. Materials and Methods: Aptamer-cocktail recognizing EpCAM/Vimentin/EGFR/CD44 were bound to a nanopillar array on a nano-microfluidic chip. The recognition was validated with cancer cells by flow cytometry and the critical parameters were optimized with the nano-microfluidic chip. Finally, the system was applied to clinical samples. Results: The proposed aptamer-cocktail showed the predominant affinity with MDA-MB-231 and SK-BR-3. When utilized to capture artificial clinical samples, it showed 71% to 83% capture efficiency. CTC detection rate was 100% in five pre-treatment and five post-treatment breast cancer patients. The enumeration data ranged from 6-33 per 2 ml. The number of CTCs in breast cancer patients before therapy was 1.27-2 times higher than that after therapy. CTCs with epithelial and mesenchymal phenotype were both detected and identified; interestingly, the mean diameter of CTCpckpos acquired in these cases was much larger than that of CTCvimentinpos. Conclusion: The nano-microfluidic chip not only made it easier to phenotyping epithelial-like or mesenchymal-like CTCs, but can also be used to detect downstream genetic variation. The established platform can be applied in clinical research and facilitate auxiliary diagnosis with tumor recurrence and metastasis in advance.
Circulating tumor cells (CTCs) are considered as a potentially important carrier for the development of metastases; they are a direct and non-invasive circulating cancer marker for tracking cancer metastasis from peripheral blood (1, 2). In contrast to circulating tumor DNA (ctDNA), CTC carry the genome and protein expression of single-cells and provide a tool for complete living cell analysis (3). Thus, by capturing rare CTCs, downstream genetic heterogeneity and the protein expression on CTC can enable us to dynamically monitor the response to tissue resection and medication. However, CTCs with low abundance and heterogeneity have been technically challenging to capture from blood, and this has hampered clinical applications in precision oncology.
Microfluidics are a viable option to enhance CTC capture efficiency with a minimal consumption of blood sample. Various microfluidic devices such as the NanoVelcro (4), polymer fluidics (5), AP-Octopus-Chip (6) and CTC-iChip (7) have been developed for improving capture rates. Previous microfluidic devices suffer from the large vertical depth of their device feature [such as micro posts (8)], resulting in increased WBC retention and, more importantly, making removal of CTCs difficult. Although the removal of CTC pools trapped on micro posts could rely on a thiol exchange reaction (6), this would be enzyme-responsive (8, 9), or temperature responsive material (PIPAAm) (10, 11). However, these methods are not practical and suffer from poor purity. Additionally, the shedding of CTCs from solid tumors into the bloodstream is a highly discontinuous process, and they can undergo epithelial–mesenchymal transition (EMT) (12). During the migrating and disseminating in the blood, CTCs lose the original epithelial features and gain mesenchymal characteristics, resulting in diverse phenotypes. Accordingly, the alteration of cancer markers on cells affects the morphology (13), leading to the low sensitivity of CTC detection based on merely a single epithelial marker affinity or size exclusion strategy, which severely restrict the applicability to routine clinical practice and diagnostics.
In this study, we developed an aptamer-cocktail functionalized nanopillar array microfluidic chip to enhance the isolation of CTCs in breast cancer (BC). The surface of the microfluidic chip was constructed with a nanopillar array, which allows CTCs to be trapped in the microfluidic channel of PDMS and immobilized on the surface of the nanopillar array subtract. To address the heterogeneity of CTCs, a total of four universal cancer markers were selected, including EpCAM/CD44 and EGFR/Vimentin. They were involved to generate a special capture agent cocktail, representing epithelial (14,15) and mesenchymal markers (16, 17), respectively. In addition, proteins can be easily denaturated by pH, temperature, and other environmental factors. Therefore, we used aptamers (single-stranded DNA) instead of antibodies to capture CTCs, which could have sensitivity comparable to antigen-antibody reactions, while being easier to synthesize. The above strategy was applied to address the heterogeneity of CTCs, reflecting the different stages of cancer progression. The nanopillar array area was carefully restricted to the area covered by microchannels. On the surface of the nanopillars, the designed aptamer cocktail was introduced through Bioin/Streptavidin conjugation into the eight channels of the microfluidic chip arranged in parallel. Nanopillar-array increases substrate roughness and the contact area between cells and the substrate, while it is more conducive to the extension of cell pseudopodia. The selective enrichment of CTC was realized by gradient shear force and the biotin-streptavidin system. The aptamer assembled micro-nanofluidic chip has high selective recognition and affinity adsorption capture, which is expected to break through the technical bottleneck of traditional CTC detection requiring large amounts of blood (>8 ml) and high omission ratios. This CTC isolation system offers the opportunity of non-invasive and real-time monitoring during disease progression. Liquid biopsies can provide complementary roles for early detection of metastatic malignant tumors.
Materials and Methods
Materials. RPMI-1640, penicillin-streptomycin (P/S), and fetal bovine serum (FBS) were purchased from Gibco (Carlsbad, CA, USA). Hoechst 33342, DiD, DiO, biotin, streptavidin (SA) and glass antifade mountant were purchased from Invitrogen, Thermofisher (Waltham, MA, USA). Anti-CD45 antibody (#ab223850) and goat anti-chicken IgY H&L (Alex Fluor® 647, #ab150171) were purchased from Abcam (Cambridge, UK). Anti-Vimentin antibody (#PA110003), donkey anti-rabbit IgG H&L (Alex Fluor® 488, #A21206) and donkey anti-mouse IgG H&L (Alexa Fluor® 555, #A-31570) were purchased from Thermo (USA). Anti-pan Cytokeratin antibody (#Z0622) was purchased from Dako (Copenhagen, Denmark). Incubating buffer was Dulbecco’s phosphate-buffered saline [DPBS, G-CLONE (Beijing, PR China)] without calcium, magnesium and phenol red. Four aptamers (Biotin-EGFR: Biotin-C6spacer-TACCAGTGCGATGCTCAGTGCCGTTTCTTCTCTTTCGCTTTTT TTGCTTTTGAGCATGCTGACGCATTCGGTTGAC, Biotin-Vimentin: Biotin-C6 spacer-CACGCATAGCCTTTGCTCCTCGTCT GGAACGTCGCAGCTTTAGTTCTGGGCCTATGCGTG, Biotin-EpCAM: CACTACAGAGGTTGCGTCTGTCCCACGTTGTCATGG GGGGTTGGCCTG-Biotin-C6 spacer, Biotin-CD44: Biotin-C6 spacer -GGGATGGATCCAAGCTTACTGGCATCTGGATTTGCGC GTGCCAGAATAAAGAGTATAACGTGTGAATGGGAAGCTTCG ATAGGAATTCGG) were synthesized by Sangon Biotech (Shanghai, PR China). Paraformaldehyde (PFA) and donkey serum were purchased from Sigma-Aldrich (St. Louis, MO, USA). TBD D-Hank’s and TBD human peripheral blood lymphocyte separation solution (Ficoll) were purchased from Hao Yang Biological Manufacture Co., Ltd (Tianjin, PR China). Dimethyl sulfoxide (DMSO) was purchased from SINOPHARM (Beijing, PR China). Triton X-100 was purchased from Coolaber (Beijing, PR China).
Sample preparation. Three cell lines, i.e., SK-BR-3, MDA-MB-231 (breast cancer cell lines) and Jurkat T (immortalized human T lymphocyte cells) were employed. The above cell lines were purchased from the National Infrastructure of Cell Line Resource (Beijing, PR China). They were cultured in RPMI-1640 supplemented with 10% FBS and 1% P/S at 37°C and 5% CO2.
A total of 10 newly-diagnosed and post-treated patients with stage I-II breast cancer and 2 healthy donors were enrolled at the Medical Laboratory of Shenzhen Luohu People’s Hospital, the People’s Republic of China. After obtaining the consent of all patients and their families, and signing of informed consent, 2 ml of peripheral blood was collected using BD Vacutainer EDTA-2K. Then, the PBMCs of peripheral blood were obtained by Ficoll density gradient centrifugation. Samples were processed within 6 h. The blood sampling points of patients were all collected after diagnosis by aspiration biopsy. This work was approved by the Ethics Committee of Shenzhen Luohu People’s Hospital (2021-LHQRMYY-KYLL-025a) and performed according to the declaration of Helsinki.
Microfluidic chip fabrication. The microchannel (20 mm×2 mm) pattern was prepared by photoetching with photoresist on the glass slide. The layout pattern was designed by 2D AutoCAD software (Autodesk, San Rafael, CA, USA) and printed on a 20,000 dpi photo mask, which was exposed, dissolved and corroded to remove the photoresist. Preparation of nanopillar was performed by RIE laser processing of glass wafers. After the nanopillar-array was prepared, scanning electron microscopy (SEM, HITACHI UHR FE-SEM SU8000 Series, SU8020, Tokyo, Japan) was used to confirm the morphological characterization of the nanoarray. As shown in Figure 1A-C, the nanoarray grew compact and uniform in thickness. The diameter of the nanopillar was about 1.5 μm, the height was about 615 nm, and the spacing was in the micron level (1.4 μm), meeting the requirements of cell capture. The microfluidic PDMS channel (20 mm*2 mm*100 μm), matching the above microchannel was designed and printed on the photomask, and the master mold was made with photoresist on the silicon wafer. Briefly, the cavity was made of photoresist on 4-inch diameter silicon wafers. The curing agent and alkali were mixed with PDMS in a ratio of 1:10 and degassed. After curing for 21 hours at 65°C, PDMS was cut and peeled off.
Design and characterization of the nano-microfluidic chip. (A) Picture of the microfluidic PDMS top and MN nanopillar chip. (B, C) Scanning electron microscopy (SEM) image of the nanopillar chip and the hierarchical structures on the bottom, side and top. (D, E) Schematic diagram of the streptavidin-modified nanopillar chip, and CD44, EGFR, Vimentin and EpCAM aptamers that specifically bind to the four proteins on tumor cells.
Activation of chip substrate and modification of cells. The cells were captured by biotin-labeled aptamers, pumped into the streptavidin modified chip system (see schematic diagram in Figure 1D, E). Briefly, SA was added with DPBS forming a final concentration of 10 μg/ml. Then, the SA (100 μl) was incubated with the chip for 60 min at 37°C. The SA-chips were triple washed with DPBS buffer and dried at 37°C. Biotin-labeled aptamers were added into the PBMC, with a final concentration of 4.20×10−4 M. After 40 min of incubation at 37°C, the aptamer labeled cells were triple washed with DPBS buffer and resuspended to 200 ul in RPMI-1640. The prepared cell suspension was injected into the single-cell isolation platform.
Flow cytometry analysis and confocal imaging. To verify the specific binding ability of aptamer and cancer cells, the breast cancer cell lines and control PBMC were split into equal aliquots, which contained about 2×106 cells. FAM (6-carboxy-fluorescein) labeled the four aptamers, which were added into separate aliquots with a final concentration of 4.20×10−4 M. After 40 min of incubation at 37°C, the aptamer labeled cells were centrifuged and washed with DPBS for three times. Fluorescent intensity of the FAM-aptamers on cells was characterized by flow cytometry (BD FACSAriaTM II, Franklin Lakes, NJ, USA).
To simulate clinical samples, aptamer-bound SK-BR-3 (1000 cells) and MDA-MB-231 (1,000 cells) were labeled with membrane dyes (DiD) and compared with control PBMC (2ml peripheral blood, DiO) suspensions, respectively. After enrichment on the CTC isolation platform, the capture efficiency of SK-BR-3 and MDA-MB-231 were calculated by confocal imaging (Zeiss LSM800, Oberkochen, Germany). The capture efficiency was calculated as the captured cells divided by the spiked cells. The clinical samples of BC patients at the end of the load test run were characterized by immunofluorescence and observed by LSM800. Briefly, anti-pan Cytokeratin, anti-CD45 and anti-Vimentin antibody mixture solutions were diluted 1:500 with DPBS buffer containing 2% donkey serum and 0.01% Triton X-100. Two hundred μl of primary antibody solution was added to the captured microarray, following 1 h of incubation at room temperature. They were then washed with DPBS for three times. The reaction system of the secondary antibody was the same with the primary antibody. The secondary antibody was dropped onto the captured microarray and shielded from light for 30 min at room temperature, and then washed with DPBS for three times. Finally, Hoechst 33342 was diluted with DPBS at a ratio of 1:5,000 and dropped onto the microarray at room temperature for 8 min. The microarray was then sealed with glass antifade mountant and observed by LSM800.
Single-cell collection and whole genomic amplification (WGA). CTC (CD45neg/PCKpos and/or Vimentinpos/Hoechest 33342pos) was selected as the selection object, allowing single CTCs to be identified and isolated by a fluorescent inverted microscope (Zeiss Observer3) and UnipicK Plus (NeuroInDx, Los Angeles, CA, USA). Once isolated at the single-cell level, CTCs underwent whole genome amplification. The DNA of isolated CTCs was amplified using the MALBAC@ Single Cell WGA Kit, according to the manufacturer’s instructions. The number of PCR cycles was increased to 21.
Droplet digital PCR. Primer sequences of HER2 and reference gene EFTUD2, as well as of droplet digital PCR (ddPCR) reaction system were referred to previous literature reports (18, 19). The CTC-WGA-DNA was prepared with 30 ng/μl in a total volume of 25 μl, and the total reaction was partitioned into ~20,000 droplets per sample in a JD Printer 100 droplet generator (CELLOMICS, ShenZhen, PR China). Emulsified PCR reactions were run on a thermal cycler (BIOER, HangZhou, PR China) incubating the plates at 95°C for 5 min followed by 33 cycles of 95°C for 15 s and 60°C for 60 s, followed by an infinite incubation at 35°C. The AnalyzeSystem software from CELLOMICS was used to assess the number of droplets positive for HER2 and EFTUD2.
Statistical analysis. One-way analysis of variance (ANOVA) was applied to test the significance of the same experimental object between different experimental groups. The Kruskal-Wallis test was used to compare multiple groups with uneven variances. Pearson’s correlation was used to compare correlations in the spiking experiment. All tests were repeated three times. Data are presented as means±standard deviation (SD). All statistical analyses were performed using IBM SPSS Statistics v24.0. All graphs were generated using GraphPad Prism 5. Statistical significance was defined as ns, not significant; *p<0.05; **p<0.01; ***p<0.001.
Results
Aptamer selection and recognition effect validation. Flow cytometry was applied to investigate the recognition ability of single aptamer (EpCAM, Vimentin, EGFR and CD44) and aptamer-cocktails against the cell lines. SK-BR-3 and MDA-MB-231 lines were incubated with FAM-apt (green fluorescence). The results of the flow cytometry (Figure 2A, B) showed that fluorescence intensity of single FAM-apt and FAM-apt-cocktail conjugated on the surface of SK-BR-3 and MDA-MB-231, respectively. Compared with single FAM-apt, the fluorescence intensity of FAM-apt-cocktail was notably higher, which indicated that there was increased FAM-apt binding to the cells. For MDA-MB-231, the binding ability of the aptamers was as following: EpCAM-apt > Vimentin-apt > EGFR-apt > CD44-apt. Although the aptamer had certain non-specific adsorption on control PBMCs, its fluorescence intensity was still lower than that of the blank group SK-BR-3 and MDA-MB-231.
The nano-microfluidic platform and performance verification. The fluorescence intensity of MDA-MB-231 (A) and SK-BR-3 (B) captured by four kinds of FAM-apt detected on flow cytometry. (C) The microfluidic holder consists of a PMMA window and MN chip slot, which is used to clamp microfluidic PDMS top and MN nanopillar chip to form a microflow channel through which the sample flows and is captured. (D) The optimal flow velocity of the nano-microfluidic platform was 0.5 ml/hr. (E) Using Jurkat cells as negative controls, the capture efficiency of cancer cells was verified by four chip designs. (F) Investigation of capture efficiency. The study was conducted by adding 100 to 2,000 cancer cells into PBMC of 2 ml of peripheral blood and to the Jurkat control group, respectively; the recovered cells were measured by counting. Data represent mean±SD of three or more independent replicates. Statistical significance was defined as ns, not significant; *p<0.05; **p<0.01; ***p<0.001.
Flow rate optimization in the single-cell isolation platform. The single-cell isolation platform is shown in Figure 2C. The platform consists of four parts: microfluidic holder, microfluidic PDMS top, micro/mano (MN) nanopillar chip and a liquid fluid system. The detection chip includes a MN nanopillar chip and microfluidic PDMS top, which were used to capture CTCs. We aimed to determine the optimal flow velocity, so as to improve the capture efficiency of rare cells. With the above four aptamers as capture markers, the flow rates varied from 0.2 to 2.0 ml/hr. As shown in Figure 2D, the capture efficiency of SK-BR-3 and MDA-MB-231 firstly increased and then decreased with the increase of flow rate, which was at 0.5 ml/h at maximum. For Jurkat cells, the capture efficiency did not change significantly with flow rate variations.
Capture efficiency and recovery of CTCs on the nano-microfluidic chip. In order to verify the capture efficiency by substrate roughness and aptamer modification in the microfluidic chip, the following 4 groups were designed: Flat w/o Biotin (A), Flat w/ Biotin (B), Structure w/o Biotin (C) and Structure w/Biotin (D) (as shown in Figure 2E). In test group A, cells not incubated with the biotin-aptamer were added to an ordinary slide, and the adsorption rate was 8.04% to 11.98%. In test group B, cells incubated with biotin-aptamer were added to an SA-ordinary slide, and the capture rate did not change significantly compared to group A. In test group C, cells not incubated with biotin-aptamer were added to the substrate of the nanopillar array chip; the capture efficiencies of SK-BR-3 and MDA-MB-231 were 5 times higher than that of the ordinary slides. The results showed that the nano-microfluidic chip substrate could intercept tumor cells using its special nanopillar structure. In test group D, the capture efficiency of SA-modified nanopillar array chip substrate was further increased by 1.5 times, compared to the unmodified structure. Moreover, in group C and D, the capture rates of SK-BR-3 and MDA-MB-231 were significantly different from that of the negative control Jurkat cells (ANOVN-test, F=76.84 and 68.86, respectively, p<0.001).
To evaluate the recovery of rare CTC with the nano-microfluidic platform, we added pre-stained DiD SK-BR-3 or MDA-MB-231 (ranging from 100 to 2000 cells) into pre-stained DiO PBMCs (separated from 2 ml of normal human donor blood samples). The mixed cells were isolated and counted. As shown in Figure 2F, the platform achieved an average capture efficiency at each dilution of cells equal to 75%, ranging from 71% to 83%. The results for both MDA-MB-231 and SK-BR-3 demonstrated that the enrichment process was linear, and the coefficient of correlation was 0.999.
Clinical application in breast cancer patients. To demonstrate the utility of the present platform regarding clinical efficiency, we investigated the ability of the micro-nanofluidic chip to isolate CTCs from breast cancer patients. Figure 3 is a representative image of CTC identification in a breast cancer patient using multiple antigenic markers vs. WBC in a healthy control. CTCs were identified as cells that were clearly visible under fluorescence microscopy. In this study, there are three types of CTCs: epithelial type (CD45neg, PCKpos, Hoechestpos), mesenchymal type (CD45neg, Vimentinpos, Hoechestpos) and epithelial-mesenchymal type (CD45neg, PCKpos, Vimentinpos, Hoechestpos). It can be seen from Figure 3 that the nuclei of CTCs were multi-nucleated and heterogenic, and the volume of epithelial CTC was more than 2 times larger than that of normal lymphocytes. But the other two types of CTCs are only slightly larger than lymphocytes. Therefore, the micro-nanofluidic chip can capture different types and sizes of CTCs without bias.
Immunofluorescence staining of the captured heterogeneous CTCs attached on the nano-membrane from the breast cancer patient samples. CTCs were isolated from breast cancer patients based on Pan-CK and Vimentin. No CTCs were found in healthy controls. Pan-CKpos and/or Vimentinpos, CD45neg, Hoechstpos were identified as CTC, while CD45pos and Hoechst33342pos were identified as WBC.
We captured CTCs in n=10/10 samples in patients, suggesting 100% sensitivity. Healthy volunteers (n=2/2) showed no presence of CTCs. All fluorescent images from patients are presented in Figure 4. In addition, we counted the white blood cells on the chip after each patient was captured. Figure 5 shows the images and pictures of the sample loading pipe, chamber chip and the waste liquid pipe generated by the above patients. Counting showed that the retention rate of WBC on the capture chip ranged from hundreds to thousands, which depends on the patient’s own factors. But the depletion rate of WBC could reach 99.5% in all clinical samples, indicating a depletion efficiency of 3.3 Logs.
Fluorescence images of CTC isolation and detection in patient (P1~P10) with 2 ml of peripheral blood samples.
Microscopic images and pictures of the sample loading pipe, chamber chip and the waste liquid pipe. (A) Confocal imaging of the discarded cells in the waste pipe. The discarded cells were diluted with PBS by 1:1 ratio. (B) Immunofluorescence image of CTC captured on the nano-membrane after isolation by the micro-nanofluidic platform.
Table I shows the characteristics of 10 patients with breast cancer and the heterogeneous CTCs captured using the micro-nanofluidic chip. The results show that the number of CTCs captured in 2 ml of blood ranged from 6 to 33. The number of CTCs in breast cancer patients before therapy was 1.27-2 times higher than that after therapy. Before therapy, the number of CTCs in stage 2 patients (patients 7, 8, 9 and 10) was twice as high as that in stage 1 patients (patient 1). Only patient 4 had been undergoing postoperative treatment for more than two years but the CTC number was much higher than that of other patients undergoing the same treatment. An investigation of her case revealed that she had metastasis to the left supraclavicular lymph node and left axillary lymph node and would undergo salvage chemotherapy. In addition, the proportion of Vimentinpos CTCs in that patient was 58%, which further confirmed recurrence and metastasis. Except for patient 4, patients 5 and 7 had lymph node metastasis at initial diagnosis, and the proportion of Vimentinpos CTCs was above 65%. Surprisingly, patient 10 was newly diagnosed with breast cancer and Vimentinpos CTCs had reached 90%; it was significant that the patient is reviewed 3-6 months after surgery. Comparing both before and after therapy, we believe the micro-nanofluidic chip may predict the treatment response based on CTC enumeration. The proportion of Vimentinpos CTCs may indicate tumor recurrence and metastasis, as in patient 4. More patients need to be tested and followed up, as in patient 10.
Characteristics, tumor node metastasis (TNM) staging and circulating tumor cells (CTC) numbers with blood collection time points in breast cancer patients and healthy donors.
Determination of HER2 amplification status on CTC-WGA-DNA by droplet digital PCR. CTCs were randomly picked up from 10 patients by UnipicK Plus (Figure 6). In order to verify whether the CTCs sorted by our nano-microfluidic chip can be used for downstream genetic detection, we assessed the HER2 amplification status of CTC-WGA-DNA in the above 10 BC patients by droplet digital PCR (Table II). According to the cut off value set by Garcia-Murillas et al. (19), using a HER2:EFTUD2 copy number ratio of 2.0, it can be seen that patients 7 and 10 had HER2 amplification. The HER2:EFTUD2 copy number ratios of the above two samples were as high as 7.486 and 7.816, respectively. Figure 7 shows the output data from ddPCR of two patients with HER-2 amplification. The other patients did not have HER2 amplification, and the average HER2:EFTUD2 copy number ratio was 0.682.
Images of picking up identified CTC by the UnipicK Plus system. (A) Locate. Immunofluorescent stain was performed on the MN nanopillar chip for CTC identification (CD45neg, PCKpos or/and Vimentinpos, Hoechest 33342pos) under a 20× objective. (B) Collect. (C) Dispense. The empty area indicates that the target cell has been sucked out.
Results of HER-2 amplification status of CTC-WGA-DNA in breast cancer patients.
The output data from ddPCR of positive cases in the form of 1D-plots. (A) and (B) are droplet fluorescence distribution and droplet fluorescence histogram of HER2 amplification test results of patient 7, respectively. Similarly, (C) and (D) are the test results of patient 10, respectively.
Discussion
The gold standard for tumor diagnosis is tissue biopsy, which can reflect the molecular characterization of the tumor and help with clinical diagnosis and medication (20). However, intratumoral heterogeneity and fast evolving dynamics contribute to genetic alterations, sometimes leading to treatment failure, when using historical tissue biopsy data (21). These processes include microenvironmental factors, drug pressure, metabolic and homeostatic mechanisms, all of which influence the molecular characteristics of tumor cells (22, 23). Additionally, tissue biopsies are invasive and incapable of dynamic observation. Liquid biopsies can be performed at any time (such as CTC) and can dynamically reveal the genomic and proteomic information of CTCs; thereby, provide key information on tumor cell biology, so as to timely improve the treatment plan (24-26). However, the use of CTCs as clinical biomarkers is still on hold due to the technical difficulties in their isolation. In this work, we combine an aptamer and nanomaterial-based microfluidic chip to efficiently capture CTCs in whole blood from breast cancer patients. Our nano-microfluidic detection system shows advantages including (a) using peripheral blood volumes as small as 2 ml, (b) a high capture efficiency (71% to 83%) of rare tumor cells from PBMCs, (c) high sensitivity for CTC detection, (d) allowing the capture and characterization of heterogeneous cells, and (e) the availability for downstream inspection.
The CellSearch system of immunomagnetic bead positive enrichment method based on EpCAM antibody detects CTCs in up to 60% of metastatic cancer patients (27). Vimentin is overexpressed in multiple epithelial cancers, including breast cancer, and its overexpression correlates well with biological behavior. Fasanya et al. (28) found that Vimentin had high presence on MG63 (77%) by FCM. A recent study by Costello et al. (29) reported seven vimentin thioated aptamers with high specificity for IGROV cells. The fluorescence intensity of FITC-vimentin binding cells was up to 5×102. Results of Scharpenseel et al. (30) showed that the detection rate of CTCs in non-small cell lung cancer patients with immunomagnetic bead technology based on the combination of EGFR and EpCAM was 1.76 times higher than that of a single marker. CD44 is highly expressed in many cancer types, which is a proven surface biomarker of cancer-initiating cells in breast cancer and other tumors (31). The combination of the above four markers was selected as the capture tag in our study. As shown in Figure 2A and B, its binding ability with cancer cells was significantly enhanced with fluorescence intensity up to 5×103-104. Previously, Zhao et al. (32) connected the aptamer to the microfluidic chip embedded with silicon nanowire substrate (SiNS); the results showed that the binding efficiency of the synergistic aptamer-cocktail with cancer cells was five times higher than that of the single aptamer. Similarly, Liu et al. (33) modified multimarker aptamer cocktail DNA nanostructures (TP-multimarker) on the deterministic lateral displacement (DLD) microfluidic chip, and the average capture rate of CTCs reached over 80%, which was similar to our research results. In addition, compared with a single aptamer, TP-multimarker can improve the capture efficiency by 50%. Therefore, it is feasible and effective to improve the capture efficiency of CTCs by using the synergistic effect of an aptamer-cocktail.
To assess the performance of the multiaptamer assembled nano-microfluidic chip, spiking experiments were carried out. The result demonstrated a high capture efficiency (75%), in contrast to several immunomagnetic bead technologies that display efficiencies of 40-60% (34, 35). Of course, there are many size-based microfluidic methods, such as DLD, shear-induced diffusion (SID) and other techniques, whose capture efficiency can reach more than 90% or even 99% (36, 37). However, our study results (see Figure 3) showed that CTCs were not always larger than WBCs, such as Vimentin expressed-CTCs were not much different from lymphocytes. Previous studies have used large cell lines to verify their experimental scheme, which could lead to certain misses in detection rates in the CTC capture of clinical samples.
Peripheral blood CTCs of clinical patients are heterogeneous at multiple levels, including size, morphology, and molecular markers (38). In order to verify the application of the nano-microfluidic detection system in real samples of cancer patients, 10 cases of breast cancer patients were selected for CTC sorting and identification. After observation of 197 CTCs from 10 patients, typical epithelial CTCs (large), mesenchymal CTCs (small) and epithelial and mesenchymal CTCs (small) were selected, as shown in Figure 3. Our experimental results are consistent with the research report of Khoo et al. (13), where the majority of “small” CTCs were positive for Vimentin, further confirming their mesenchymal phenotype. In addition, “small” CTCs are more likely to form CTC clusters, which usually display mixed epithelial and mesenchymal characteristics and are surrounded by more platelets, which can change the tumor microenvironment for distant metastasis (39). When determining CTCs with different solid tumors in patients, the heterogeneity of CTC morphology can still cause some difficulties in interpretation. However, up until now, testing techniques and interpretation standards are not standardized, due to the lack of a guide specifying the detailed definition and interpretation criteria of CTCs.
As shown in Table I, patients with low CTC numbers after surgery and therapy might indicate an effective treatment effect, while patients with a large number of CTCs might mean that they are at risk of recurrence, or that extra combined treatments are required (such as in patient 2). The results of more than 6000 CTC-based prostate cancer diagnoses suggest that CTCs can be used as an indicator of early efficacy evaluation in metastatic castration-resistant prostate cancer (mCRPC) patients (40). The dynamic advantage of quantitative detection has been verified in the clinical application of efficacy evaluation. This conclusion was also confirmed in patient 4 in this study. The latter developed disease progression (recurrence and metastasis) after eight courses of AC-T chemotherapy within two years after surgery. Synchronously, the CTC number reached 26/2ml, 2 to 4 times higher than other patients after the treatment. In addition, the proportion of Vimentinpos CTCs in the patient was 58%. Vimentin is mainly associated with poor prognosis and a metastatic phenotype (41). A study reported by Yu et al. (42) showed that dynamic changes of EMT types on CTCs are involved in tumor metastasis. Continuous CTC monitoring in 11 breast cancer patients suggests that interstitial CTC is associated with disease progression. However, the correlation between CTC phenotype and disease outcome is still under investigation and requires a large sample size of tracking and recording. In addition, different technologies may favor a capture tendency for a certain CTC subtype and these biases should be considered when CTCs are used for clinical application.
The reference gene EFTUD2 could be detected by ddPCR for CTC-WGA-DNA after sorting, picking-up and WGA, indicating that the CTCs sorted by our nano-microfluidic chip could be used for downstream gene detection. The results of HER2 amplification in 2 cases were also consistent with the immunohistochemical results, suggesting that CTCs could replace tissue to detect gene variation in cancer patients. Interestingly, patient 3 was identified as HER2 expression positive by immunohistochemistry at initial diagnosis. But after 10 months of surgery, there was no corresponding HER2 amplification on the CTC, and the number of CTC was lower than that of the patient before surgery. The number of CTCs and HER amplification status may indicate that the patient is in remission and treatment is effective. However, large, multicenter studies are needed to provide sufficient evidence of consistency between CTCs and tumor tissue, and fluid biopsies can be used instead of tissue biopsies to dynamically monitor disease progression.
With these potential clinical applications of our nano-microfluidic detection system, a few limitations remain. Isolated single cells restricted studies at the transcriptome level due to fixation with PFA and membrane lysis with Triton X-100. Recently Shi et al. (43) developed a cell fixation agent [the effective constituent is 3,3′-Dithiobispropanoic acid bis and N-succinimidyl-3-(2-pyridyldithiol) propionate] that can maintain the complete structure and conformation of the protein antigens, while increasing the cell permeability; this is compatible to a variety of single-cell nucleic acid sequencing methods, so as to complete single-cell RNA sequencing analysis. Secondly, the problem of non-specific binding of leukocytes has not been fully worked out, which has increased our workload (e.g., selecting CTC by microscopy for downstream experiments). Likewise, the PEG-PLGA-nanofiber reported by Zhu et al. (44) was modified with specific biomarkers for identifying LUAD and SCC lung cancers, respectively. Compared to these results, the markers identified in this procedure are expected to help in the optimization for monitoring of disease evolution of breast cancer during treatment. In addition, increasing the enrolled number of patients for verification the clinical application of the CTC test is promising. Further studies on better cell fixation, purification methods with micro-nanochip with CTC isolation, identification and downstream sequencing tests in the integrated system are expected to be developed.
In conclusion, new technologies need to be systematically validated and tested in multi-center clinical trials to compare their performance before they can be used in clinical practice. The aptamer-cocktail assembled nano-microfluidic chip achieved a higher performance in capturing CTCs in breast cancer, and the system is expected to be applied in other cancers. Future research may reveal the clinical significance of dynamic CTC detection by nano-microfluidic chip and facilitate the discovery of new drug therapy targets combined with sequencing technology.
Acknowledgements
We would like to thank all the participants and donors of this study for their valuable contribution.
Footnotes
↵* These Authors contributed equally to this work.
Authors’ Contributions
Min Li and Dan Liu: designed and performed the experiments, analyzed the data, and wrote the manuscript. Jian Zhang: collected clinical samples and analyzed the flow cytometric data. Xiuming Zhang: formulated the idea and provided financial support of the study. Xiaowen Dou: designed the experiments, contributed to experimental execution and revision of manuscript.
Conflicts of Interest
The Authors declare that there are no conflicts of interest in relation to this study.
- Received June 27, 2022.
- Revision received July 14, 2022.
- Accepted July 18, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.