TY - JOUR
AU - Powell, H., M.
AB - Abstract Personalized cancer therapies drive the need for devices that rapidly and accurately segregate cancer cells from solid tumors. One potential sorting strategy is to segregate populations of cells based on their relative strength of adhesion. To investigate the effect of surface hydrophilicity and cell phenotype on adhesion, primary human breast skin fibroblasts and keratinocytes and MCF-7 breast cancer cells were seeded onto air and CF4 plasma-treated nanofibers followed by exposure to three shear stresses (200, 275 and 350 dynes per cm2) 1 hour after inoculation. No difference in strength of adhesion was measured in either fibroblasts or keratinocytes on either plasma treated-surface: all exhibited >60% of the initial cell count after a 5 minute exposure to 350 dynes per cm2 of shear stress. In contrast, a significant difference between relative strength of adhesion on air versus CF4 plasma-treated surfaces was observed for MCF-7 cells: 26% and 6.6% of cells remained on the air and CF4 plasma-treated surfaces, respectively. The ability to sort this cancer cell line from two non-cancerous primary human cells was evaluated by inoculating a mixture of all three cell types simultaneously onto CF4 treated nanofibers followed by 1 hour of culture and exposure to 350 dynes per cm2 shear stress. The majority of MCF-7 cells were removed (0.7% remained) while a majority of fibroblasts and keratinocytes remained adhered (74 and 57%). Post-sorted MCF-7 viability and morphology remained unchanged, preserving the possibility of post-separation and analysis. These data suggest that the plasma treatment of electrospun scaffolds provides a tool useful in sorting cancer cells from a mixed cell population based on adhesion strength. Insight, innovation, integration Ongoing efforts developing personalized cancer therapies are hindered by a lack of platforms allowing rapid, accurate segregation of cancer cells from solid tumors. We demonstrate than electrospun fiber can – if properly surface-treated – distinguish between cancerous and non-cancerous cells based on their relative strength of adhesion. Post-sorted cell viability and morphology showed that these cells remained unchanged, allowing the possibility of post-separation proliferation and genetic analysis. Our data suggest that the plasma treatment of electrospun scaffolds provides a widely useful tool capable of sorting cancer cells from a mixed cell population based on adhesion strength. Introduction Properties common to metastatic cancers are local invasion and distant spread, though causative factors and molecular composition may differ greatly.1 Decades of clinical practice have indicated that standard treatment regimens are often not effective for specific components of the tumor cell population.2–4 This has shifted cancer therapy from a general, “one drug fits all” approach toward more personalized medicine.5–8 The ability to predict individual patient response and optimize chemotherapeutic drug choice, adjuvant therapies, and appropriate dosing could greatly increase effectiveness. However, personalized cancer treatments require swift, accurate and efficient diagnostics to assess disease states. Isolating patient cells of interest from biopsies is important to such diagnostics as the collected and analyzed cells will determine the course of treatment. Solid malignant tumors are heterogeneous masses whose composition is determined by the location of origin. Cells that comprise such tumors can include normal cell types, such as endothelial cells and fibroblasts, and cancer cells of multiple phenotypes,9 the latter of which must be effectively sorted from this initial mixed population for effective treatment development. Conventional cell sorting techniques – i.e., fluorescence- or magnetic-activated cell sorting (FACS or MACS) – rely upon labeling cells with either beads or dyes for separation.10–13 Recent improvements14 have not yet been widely adopted as the cost and size of the supporting equipment and the skill levels required limit widespread accessibility. Furthermore, a label specific to the cell of interest must be predetermined in order for it to be discerned from the population. The labeling process may alter the cell function but, more importantly, cancer cells of potential interest can be missed if they are not part of the specific cancer phenotype that was labeled. Such label-free cell sorting efforts seek to exploit differentials that exist between cells such as cell density,15 size,16 dielectric properties17,18 and refractive index.19 While label-free techniques can achieve efficiency or purity of 90% or greater,15,16,19 many are limited by small sample sizes and require either bulky or expensive supporting equipment or complex nanofabrication techniques. Cancer cell lines have also been sorted from normal cell lines based on their strength of adhesion, stemming from the characteristic of cancer cells to exhibit decreased adhesion from native cells.20 Kwon et al. showed that a breast cancer cell line, MCF7, could be segregated from a breast epithelial cell line, MCF10a, via adhesion onto polyurethane acrylate nanopatterned microfluidic channels.21 This technique represents a significant advance in cancer cell sorting but requires intricate and potentially costly processing of the microfluidic device. To realize the greatest potential for clinical translation, inexpensive, high surface area, high throughput devices are needed. The goal of the present study is to utilize an inexpensive, high-throughput, electrospun fiber-based platform to sort cancer cells based on adhesion. First viability, spreading and strength of adhesion of primary human breast epithelial and fibroblast cells and MFC-7 cancer cells were quantified as a function of electrospun fiber hydrophilicity. Subsequently, the efficacy of the electrospun platform to sort the cancer cells from a mixed population of the primary human keratinocytes and fibroblasts via and applied shear stress was assessed. Additionally, the viability of the cells after exposure to shear stress and subsequent removal from the growth surface was quantified to ensure that the sorted cancer cells could still be used for subsequent downstream analyses. Materials and methods Polycaprolactone scaffolds Electrospun scaffolds were prepared using a solution of 10 wt% polycaprolactone (PCL; MW ∼65 000; Sigma-Aldrich, St. Louis, MO) in 1,1,1,3,3,3-hexafluoro-2-propanol (HFP; Oakwood Products, West Columbia, SC). HFP-PCL solutions were electrospun at a rate 10 ml h−1 (kd Scientific, Holliston, MA) and an electrical potential of 20 kV (Glassman High Voltage, High Bridge, NJ) onto glass slides or coverslips positioned on a grounding plate to a thickness of approximately 100 μm. Prior to cell inoculation, all electrospun scaffolds were subjected to a sterilizing pre-treatment of 70% ethanol for 20 min, followed by two 15 min medium rinses. Plasma surface modification As spun-PCL fibers were placed into a Harrick plasma cleaner (Harrick Plasma, Ithaca, NY, USA). Air plasma-treated samples were placed into the chamber under vacuum at 1000 mTorr, and a plasma radio frequency of 8–12 MHz for 2.5 minutes. Tetrafluoromethane (CF4) plasma-treated samples were placed into the chamber under vacuum at 400 mTorr, with the same radio frequency and time as the air samples. After 2.5 minutes, samples were removed from the chamber, and kept in a sealed container until use. Scaffold characterization The morphology of the PCL fibers, pre- and post-plasma treating, was qualitatively assessed using scanning electron microscopy (SEM; FEI Quanta, Hillsboro, OR). As-spun, air and plasma etched PCL scaffolds were affixed to aluminum SEM stubs using conductive carbon tape (Ted Pella, Reading, CA), and were subsequently sputter coated with gold to render the surface conductive. All samples were imaged in secondary electron mode at 5 kV. Surface hydrophobicity was quantified using goniometry. PCL fibrous scaffolds were cut into 5 × 1 cm segments, plasma treated with air or CF4 gas, as described previously, and water contact angle was immediately measured using a Krüss Easydrop DSA20 (Krüss, Hamburg, Germany) contact goniometer. A 300 μL drop of deionized water was placed on a dry area of the PCL fiber, and using the Easydrop software, water contact angle was measured using a sessile drop contact to surface measurement. Five measurements were made and the average ± standard deviation recorded. Cell culture Primary human breast fibroblasts and keratinocytes (passage 2) and MCF-7 breast cancer cells were maintained in a humidified incubator at 5% CO2–95% air and 37 °C. Fibroblasts and keratinocytes were maintained in Dulbecco's Modified Eagles Medium (DMEM; Sigma) supplemented with 4% fetal bovine serum (FBS; Invitrogen, Portland, OR), 10 ng mL−1 epidermal growth factor (EGF; Peprotech, Rocky Hill, NJ), 5 μg mL−1 insulin (Sigma), 0.5 μg mL−1 hydrocortisone (HC; Sigma), 100 μM ascorbic-acid-2-phosphate (Sigma) and 1% penicillin-streptomycin (PSF; Invitrogen), and Medium 153 (Sigma) supplemented with 0.2 vol% bovine pituitary extract (Gemini Bioproducts, West Sacramento, CA), 1 ng mL−1 EGF, 5 μg mL−1 insulin, 0.5 μg mL−1 HC and 1% PSF, respectively. MCF-7 cells were cultured in DMEM supplemented with 10% FBS, 5 μg mL−1 insulin, 3.51 mg mL−1d-glucose (Sigma) and 1% PSF. Medium for all cells was changed every other day. Cell proliferation Once each cell type reached approximately 70% confluence, cells were harvested from culture flasks using trypsin–ethylenediaminetetraacetic acid (EDTA) at a concentration of 2.65 × 103 Units per mL trypsin + 0.01% EDTA (Sigma) and inoculated onto 12 mm diameter, electrospun scaffolds at a density of 20 000 cells per cm2 (6, 12 mm disks per group). At days 1, 3, 5, and 7, a 4 mm biopsy was removed from each sample (n = 6 per group per time point) and cell proliferation was assessed using a CellTiter 96 AQueous Non-reactive Cell Proliferation Assay (MTS) (Promega Corp.; Madison, WI). Briefly, each punch biopsy was incubated with medium containing 10% MTS-PMS solution for 3 hours at 37 °C and 5% CO2 according to manufacturer's protocol. Following this incubation, the medium was removed and its absorbance was read at 490 nm using a plate reader (Gemini Spectramax). Average absorbance ± standard deviation was reported. Confocal microscopy To quantify cell spreading as a function of surface modification and cell type, cells were harvested and inoculated onto electrospun scaffolds at a density of 50 000 cells per cm2. PCL-cell constructs were removed from culture at 1, 3, 6 and 24 h (n = 4 per time point). Constructs were rinsed with phosphate buffered saline (PBS) three times for five minutes each, fixed in 4% paraformaldehyde in PBS for 1 hour and again rinsed twice with PBS. Fibroblast-PCL samples were stained with phalloidin (AlexaFluor phalloidin 488; Invitrogen) and DAPI (Invitrogen), while MCF-7 and keratinocyte-PCL samples were immunostained with basic cytokeratin (Invitrogen) and DAPI. All samples were imaged with an Olympus FV1000-Spectral Confocal microscope (Olympus, Center Valley, PA) at 20–63× optical magnification. Cell area (n ≥ 100 per groups per time point) was measured using ImageJ and the average cell area (μm2) ± standard deviation was reported. Computational modeling of fluid flow and surface shear in nanofiber parallel plate device The fluid velocity profile and surface shear profile within the nanofiber parallel plate device was modeled to ensure uniformity in the image capture locations selected for the adhesion strength assay. First, surface roughness of the nanofiber platform was quantified using a Wyko NT9000 optical profilometer operating in vertical scanning interferometry (VSI) mode, with 20× objective and 1.0× field-of-view lens. The surface roughness value used to inform the model was averaged from three surface scans. Finite element fluid flow models were then constructed using Comsol 4.2a (Comsol, Inc., Burlington, MA) using a roughened lower surface to simulate the influence on the flow due to the presence of nanofibers on the base surface. To generate the surface roughness, a MATLAB script was developed to generate a series of thousands of circles, each of radius 2.789 μm having a variable separation between centers. The model chosen used 4 times the radius as the center-to-center separation. The circles were imported into COMSOL, and a rectangular block was generated for the flow channel. The Boolean Difference operator was used to subtract the circles from the area of the rectangular flow channel, giving a textured base surface. Several levels of triangular mesh refinement were used; ultimately the model with 493 236 degrees-of-freedom was chosen for the analysis presented here. As a result, the hemispheres were meshed coarsely, giving a saw-tooth texture for the base surface of the textured model used for Case 2 (Fig. 1). A surface velocity profile and a surface shear contour were calculated for an inlet velocity of 8.536 × 10−4 m s−1, with no slip at the top or base surface and an outlet pressure of 0. Locations for cell imaging before and after shear were selected in areas having equivalent surface shear. Fig. 1 Open in new tabDownload slide Finite element mesh of the rough nanofiber surface for fluid flow modeling. Fig. 1 Open in new tabDownload slide Finite element mesh of the rough nanofiber surface for fluid flow modeling. Strength of adhesion Prior to harvesting, cells for adhesion testing were stained with CellTracker™ Red CMTPX (Invitrogen), according to manufacturer's instructions, in order to quantify the relative percentage of cells remaining after testing. Each harvested cell type was inoculated onto individual electrospun slides at 20 000 cells per cm2, incubated for one hour and then tested for adhesion strength. Electrospun fiber-coated slides were imaged at equal intervals along the center line of the device, 12 mm from the inlet to 10 mm from the outlet port, using fluorescent microscopy (Nikon Eclipse LV150) before and after testing (seven images were taken at 10× magnification for each sample, 6 samples per group). Slides were loaded into a parallel-plate device following Kawamoto et al.22 and Powell et al.,23 and were exposed to a shear stress of 200, 275, or 350 dynes per cm2 for 5 min each (n = 6 per shear stress). Image analysis was performed to quantify cell number on each scaffold before and after shear exposure, average percent remaining ± standard error of the mean was reported. Cell sorting Adhesion testing was then performed with the three cell types mixed and seeded onto slides to determine whether cancer cells could preferentially be removed from the mixed population as a function of adhesion strength. To identify each cell type, fibroblasts were stained with CellTracker™ Green CMFDA (Invitrogen) and keratinocytes with CellTracker™ Red CMTPX (Invitrogen) prior to inoculation, according to manufacturer's protocols. Unstained MCF-7s and live stained fibroblasts and keratinocytes were inoculated onto CF4-treated scaffolds, incubated in blended culture medium for one hour and exposed to a shear stress of 200 dynes per cm2 for 5 min (non-shear exposed samples served as a control). All samples were fixed in 4% PFA, stained with DAPI and imaged using confocal microscopy (Olympus FV1000-Spectral Confocal). MCF-7 cells were identified by presence of DAPI nuclear staining and absence of any additional staining. The number of fibroblasts, keratinocytes and MCF-7s per field of view was quantified in non-exposed and sheared samples. Post-shear viability To determine if shear stress exposure reduced cell viability, cells were inoculated onto CF4 treated scaffolds (n = 6) and tested as described above with a shear stress of 350 dynes per cm2. The cells removed from the scaffolds by shear flow were collected and inoculated onto a polystyrene 96-well plate at a density of 1000 cells per well. Viability was quantified 1 and 24 h after inoculation using an MTS assay, as previously described, and compared to cells which had not been exposed to shear stress. Brightfield images of the plated cells were taken at 1 hour post inoculation. Results PCL scaffolds Utilizing scanning electron microscopy, no morphological differences between the as-spun and plasma-treated (air or CF4) electrospun fibers were observed (Fig. 2). Fiber shape and topography suggest that plasma treatment had no effect on fiber morphology at either the micron or submicron level. In contrast, plasma treatment did have a significant effect on the hydrophilicity of the PCL fibers. Air and CF4 plasma treatment resulted in contact angles of 0 ± 0° and 157.6 ± 6.9°, respectively (Fig. 3). Fig. 2 Open in new tabDownload slide Scanning electron micrographs of as-spun electrospun PCL (A) and PCL after plasma-treatment with (B) air and (C) CF4. Inset: higher magnification images of individual nanofibers showing surface roughness. Scale bar = 500 nm. Fig. 2 Open in new tabDownload slide Scanning electron micrographs of as-spun electrospun PCL (A) and PCL after plasma-treatment with (B) air and (C) CF4. Inset: higher magnification images of individual nanofibers showing surface roughness. Scale bar = 500 nm. Fig. 3 Open in new tabDownload slide Contact angle of water droplets on as-spun electrospun PCL and PCL plasma-treated with air and CF4 plasma. Fig. 3 Open in new tabDownload slide Contact angle of water droplets on as-spun electrospun PCL and PCL plasma-treated with air and CF4 plasma. Cell viability and spreading No statistically significant difference in proliferation was observed between cells cultured on the CF4 plasma-treated and air plasma-treated PCL scaffolds at days 1–5 (Fig. 4). Keratinocyte and MCF-7 proliferation was significantly lower on CF4 plasma-treated scaffolds at day 7 (Fig. 4B and C). Proliferation of fibroblasts on air treated scaffolds increased rapidly from days 1 to 5, after which the rate of proliferation decreased. Fibroblasts on CF4 plasma-treated scaffolds proliferated slightly slower, with lower average absorbance values from days 1–5, but matched that of fibroblasts on air treated scaffolds by day 7. MCF-7 proliferation on both substrates closely followed the same trend until day 7 where cells on air displayed a more rapid increase in proliferation. Keratinocyte proliferation, in contrast, was slow over the course of the 7 day culture period on both scaffold types (Fig. 4). Fig. 4 Open in new tabDownload slide MTS proliferation assay of (A) primary human fibroblasts, (B) primary human epidermal keratinocytes and (C) human breast cancer carcinomas cultured on air or CF4 plasma-treated electrospun PCL scaffolds. Fig. 4 Open in new tabDownload slide MTS proliferation assay of (A) primary human fibroblasts, (B) primary human epidermal keratinocytes and (C) human breast cancer carcinomas cultured on air or CF4 plasma-treated electrospun PCL scaffolds. While scaffold hydrophobicity had minimal effects on cell proliferation, the wettability of the scaffolds significantly altered the speed and extent to which cells spread. At 1 h post inoculation, fibroblasts on air-plasma-treated scaffolds showed increased spreading compared to the CF4 plasma-treated scaffolds. Fibroblasts on air treated scaffolds exhibited spread, extended filopodia, and had visible stress fibers 1 hour post inoculation. In contrast, fibroblasts on the CF4 plasma-treated scaffolds at that time point were in the initial stages of attachment and most cells appeared rounded (Fig. 5). After 6 hours, fibroblasts on CF4 plasma-treated scaffolds began to spread more rapidly such that only a slight decrease in fibroblast size was observed at 24 hours (p > 0.05 at 24 hours). Statistically significant difference in fibroblast spreading was seen until 24 hours in culture. Prominent stress fibers were observed in the fibroblasts on air-plasma treated fibers at 24 hours whereas F-actin staining in fibroblasts on CF4-plasma treated nanofibers remains punctate with only thin stress fibers observed (Fig. 5). Keratinocyte cultures did not generally exhibit any differences in cell size (Fig. 6B) but a small increase in colony size was observed (Fig. 5). MCF-7 cells were approximately 30% larger on air treated scaffolds after 24 hours in culture and were in more tightly packed colonies than the CF4-treated scaffold group. Fig. 5 Open in new tabDownload slide Confocal image of human keratinocytes, human fibroblasts and MCF-7 cells cultured on air and CF4 plasma-treated PCL scaffolds. Cell morphology is shown after 1 and 24 hours of culture (blue = nuclei, green = actin). Scale bar = 50 μm. Fig. 5 Open in new tabDownload slide Confocal image of human keratinocytes, human fibroblasts and MCF-7 cells cultured on air and CF4 plasma-treated PCL scaffolds. Cell morphology is shown after 1 and 24 hours of culture (blue = nuclei, green = actin). Scale bar = 50 μm. Fig. 6 Open in new tabDownload slide Cell area as a function of culture time and PCL fiber surface treatment. (A) Fibroblasts, (B) keratinocytes and (C) MCF-7 breast cancer cells. Fig. 6 Open in new tabDownload slide Cell area as a function of culture time and PCL fiber surface treatment. (A) Fibroblasts, (B) keratinocytes and (C) MCF-7 breast cancer cells. Strength of adhesion The finite element analysis results, shown as color contours representing velocity and shear rate (Fig. 7), indicate that the device generated a maximum level of surface fluid velocity and surface shear starting 9 mm from the inlet and traveling down the center line of the device to the outlet. Within this region, a uniform area of shear rate and fluid velocity can be seen (Fig. 7, white dashed line). All images for the strength of adhesion were collected from this region. Fig. 7 Open in new tabDownload slide Velocity fringe and shear rate contours for the nanofiber-based rough mesh. Dashed box indicates imaging area for the strength of adhesion testing. Fig. 7 Open in new tabDownload slide Velocity fringe and shear rate contours for the nanofiber-based rough mesh. Dashed box indicates imaging area for the strength of adhesion testing. For fibroblasts and keratinocytes, no significant difference in strength of adhesion was observed between cells cultured on the air or CF4 plasma-treated PCL scaffolds and the percent of cells remaining was relatively constant across the three values of shear stress (Fig. 8A and B). MCF-7 cells, however, showed a significant decrease in cell adhesion on CF4 plasma-treated scaffolds as compared with air plasma-treated scaffolds. The CF4 group exhibited an average of 19.7% fewer cells remaining than the air group after shear stress exposure (Fig. 8C). Additionally, with increasing shear stress from 200 to 350 dynes per cm2, cell adhesion decreased from 55.8% to 26.1% cells remaining for air treated samples and from 32.2% to 6.6% cells remaining for CF4 plasma-treated samples (Fig. 8C). Comparing the retention of the three cell types on CF4 plasma-treated substrates, significant differences were found between MCF-7 cancer cells and the two primary normal tissue cells (Fig. 8D). The largest difference was found after exposure to 350 dynes per cm2, where the retention rates for fibroblasts, keratinocytes and MCF-7 cells were 61.3%, 51.9% and 6.6%, respectively. Fig. 8 Open in new tabDownload slide Strength of cell adhesion to air and CF4 plasma-treated electrospun PCL fibers. Cells were exposed to 200, 275 and 350 dynes per cm2 of shear stress for 5 min. Percent of cells remaining after the exposure to shear stress shown for (A) fibroblasts, (B) keratinocytes and (C) MCF-7 breast cancer cells. (D) Direct comparison of adhesion strength between cell types on CF4 plasma-treated electrospun fibers. Note that the MCF-7 cancer cells are very sensitive to shear stress while the epithelial and mesenchymal cell types adhere much more strongly with no change in percent of cells remaining between any of the shear stress levels. Fig. 8 Open in new tabDownload slide Strength of cell adhesion to air and CF4 plasma-treated electrospun PCL fibers. Cells were exposed to 200, 275 and 350 dynes per cm2 of shear stress for 5 min. Percent of cells remaining after the exposure to shear stress shown for (A) fibroblasts, (B) keratinocytes and (C) MCF-7 breast cancer cells. (D) Direct comparison of adhesion strength between cell types on CF4 plasma-treated electrospun fibers. Note that the MCF-7 cancer cells are very sensitive to shear stress while the epithelial and mesenchymal cell types adhere much more strongly with no change in percent of cells remaining between any of the shear stress levels. Cell sorting Confocal images of the mixed cell population on CF4 plasma-treated substrates after exposure to 350 dynes per cm2 revealed that a large percentage of fibroblasts and keratinocytes remained (Fig. 9B) compared to those before shear (Fig. 9A), approximately 74% and 57%, respectively. However, the shear-exposed substrate was almost completely devoid of MCF-7 cells, with an average of only 0.7% of MCF-7 cells remaining. Fig. 9 Open in new tabDownload slide Representative confocal images of mixed population of cells, fibroblasts (green), keratinocytes (red) and MCF-7 breast cancer cells (blue), seeded onto CF4 plasma-treated PCL fibers (A). Preferential removal of MCF-7 cells from the mixed population after exposure to 350 dynes of shear stress for 5 minutes (B). Fig. 9 Open in new tabDownload slide Representative confocal images of mixed population of cells, fibroblasts (green), keratinocytes (red) and MCF-7 breast cancer cells (blue), seeded onto CF4 plasma-treated PCL fibers (A). Preferential removal of MCF-7 cells from the mixed population after exposure to 350 dynes of shear stress for 5 minutes (B). Post-shear viability MCF-7 cells removed from CF4 plasma-treated electrospun substrates by shear stresses of 200 and 350 dynes per cm2 and MCF-7 cells plated directly without shear exposure showed no significant differences in metabolic activity either one or 24 h post inoculation (Fig. 10A). Additionally, no significant difference in cell shape was observed between the control and shear conditions (Fig. 10B–D). Fig. 10 Open in new tabDownload slide (A) MTS assay of MCF-7 cell metabolisms one hour and 24 hours after removal from CF4 plasma-treated PCL fibers using 200 and 350 dynes per cm2 shear stress versus MCF-7 cells that were not subjected to shear stress (0 dynes per cm2). Brightfield images of the sorted population after 24 hours in culture. (B) Control pure MCF-7 population, (C) 200 dynes per cm2, and (D) 350 dynes per cm2. All cells observed exhibited MCF-7 morphology. Cell fragments were sparse in the 200 dynes per cm2 group but more prevalent in the 350 dynes per cm2. Fig. 10 Open in new tabDownload slide (A) MTS assay of MCF-7 cell metabolisms one hour and 24 hours after removal from CF4 plasma-treated PCL fibers using 200 and 350 dynes per cm2 shear stress versus MCF-7 cells that were not subjected to shear stress (0 dynes per cm2). Brightfield images of the sorted population after 24 hours in culture. (B) Control pure MCF-7 population, (C) 200 dynes per cm2, and (D) 350 dynes per cm2. All cells observed exhibited MCF-7 morphology. Cell fragments were sparse in the 200 dynes per cm2 group but more prevalent in the 350 dynes per cm2. Discussion Plasma surface treatments are commonly utilized to modify the surface of polymer materials to change their wettability and to alter cell-material interactions.24–27 As anticipated, air plasma treatment of PCL electrospun substrates resulted in a significant decrease in wetting angle (increased hydrophilicity) compared to untreated PCL while CF4 plasma treatment increased the wetting angle (increased hydrophobicity) (Fig. 2). Increased hydrophilicity following air plasma treatment has been seen previously in polyethylene terephthalate (PET) films23 and PCL nanofibrous scaffolds28 and is attributed to the incorporation of polar hydroxyl, carbonyl, and carboxyl functional groups on the surface.29,30 CF4 plasma treatment has also been shown to decrease the wettability (i.e., increased contact angle) of PET films,24 polyvinylidene fluoride, polyimide and polyamidoimide,31 cotton fabrics,32 and polyurethane substrates33 as a result of the incorporation of fluorine-containing functional groups into the polymer surface.24 In contrast to prior reports showing changes in substrate architecture in response to plasma treatment,25,34,35 no difference in nanofiber morphology was observed between as-spun, air and CF4 plasma-treated PCL scaffolds (Fig. 1). Previous reports showed an increase in fiber diameter and slight fiber melting of PCL nanofibers after air plasma treatment at 30 W for 5 minutes.28 PCL films plasma treated with a mixture of oxygen and air were also four times as rough as non-treated PCL films.36 The maintenance of PCL fiber morphology following plasma treatment in the current study was attributed to use of low power during the treatment process that reduces overall sample temperature and prevents sample melting and deformation. Regarding the behavior of seeded cells on our plasma-treated substrates, imaging of the primary human fibroblasts, primary human keratinocytes and MCF-7 cells during the initial adhesion and spreading phase revealed a significant increase in fibroblast spreading on the air-treated PCL fibrous substrates at all time points up to 6 hours (Fig. 5). In contrast, differences in keratinocyte and MCF-7 cell spreading were more evident at later time points. Surface wettability has previously been shown to alter cell spreading and morphology by controlling protein adsorption.37,38 In general, hydrophobic surfaces increase overall protein adsorption. However, strong adhesion of proteins to hydrophobic surfaces often leads to a significant change in protein conformation reducing the availability of adhesion binding sites.39–41 As a result, hydrophobic surfaces are often associated with slow cell adhesion and spreading.42,43 For example, hexamethyldisiloxane plasma treated with oxygen showed a reduction in contact angle to zero along with improved fibroblast spreading.44 Decreased fibroblast spreading on CF4 plasma-treated PCL fibers at early time points is attributed to changes in protein conformation following adsorption to the hydrophobic fibers. With time, additional proteins from the medium adsorb to the PCL surface and likely render the surface more hydrophilic as has been seen previously with poly(d,l lactidic acid) films soaked on serum containing medium for 30 minutes.45 Thus the effects of surface wettability are less significant at later time points. MCF-7 cell area on air plasma-treated PCL fibers was significantly larger than on CF4 plasma-treated fibers only after 24 hours. As these cells are slow to adhere and spread, it is likely that changes in adhesion and spreading rates at the early time points were too small to be detected as the cells remained very rounded in each condition. However, with increased culture time the cells were able to spread on the culture surface and differences in morphology could be more easily observed. In the early 1940's cancer cells were reported to display weaker intercellular adhesions than normal native cells.20 More recently, metastatic cancer cells have been shown to exhibit fewer E-cadherin adhesion molecules when detaching from the epithelium and primary tumor leading to a shift into a motile phenotype46 that can correspond to decreased adhesion strength. For example, the onset of breast cancer cell motility in response to insulin-like growth factor I (IGF-I) correlates with lowering of adhesion strength from 2.52 ± 0.20 to 1.52 ± 0.13 dynes per μm2 in cells attached to fibronectin.47 In the current study, non-cancerous primary human cells were shown to be adhered more strongly to the PCL nanofibers versus the MCF-7 cancer cell line. Additionally, the MCF-7 cancer cell strength of adhesion was significantly affected by fiber hydrophilicity to which primary cells were not highly sensitive. Confocal imaging and subsequent cell area measurements indicate that the MCF-7 cells were slower to attach and spread upon the nanofibers than the non-cancerous cells (Fig. 4 and 5). It is likely that both the fibroblasts and keratinocytes form focal adhesions to the surface more rapidly and that these focal adhesions are more stable than those of the MCF-7 cells. This would result in a larger number of cell-matrix adhesion sites at a given time and provide a stronger bond between the non-cancerous cells and the provisional matrix (i.e., PCL nanofibers) than the MCF-7 cancer cells. The use of plasma treated PCL nanofibers has been shown to sort cancer cells from a mixed population of cells based on adhesion. Strength of adhesion has been utilized in one prior study to sort the MCF-7 breast cancer cell line from a breast epithelial cell line.21 This study showed that adhesion can be utilized to sort these two cell types; however the microfluidic chamber utilized required a significant amount of labor for fabrication, analyzes a relatively small amount of cells (∼1 200 cells per 4 channel device), and has been validated using only cell lines. Our current platform is rapid, requiring only one hour of adhesion, inexpensive, and high throughput; more than 150 000 cells per run were segregated in this relatively small platform. The PCL-nanofiber platform sorts cancer cells from the population without any significant change to cell viability (Fig. 8), preserving the possibility of post-separation expansion and analysis. Another potential advantage of this nanofiber-based platform is that its porous nature provides an open exchange for soluble factors. This enables studies of the effect of chemical communication between cells on adhesion strength via membrane-separated co-culture.48,49 Additionally, in the current context MCF-7 cells were effectively sorted from clinically relevant primary human cells. These data suggest that creation and incorporation of nanofiber-based substrates into larger device platforms could allow for important advances in cancer cell isolation from mixed populations. Conclusion The current study presents an electrospun polycaprolactone fiber-based device for cancer cell sorting from a mixed population of primary human cells. CF4 plasma-treatment of the PCL nanofibers decreased fiber wettability and subsequently reduced strength of cancer cell adhesion, allowing them to be more easily segregated from the general cell population following shear stress. This device is inexpensive, easy to use and efficiently sorts these breast cancer cells without any apparent damage. Though future studies are needed to validate this device using primary tumor resections and biopsies, the current data suggests significant potential for contributing to the development of personalized cancer treatments. Acknowledgements This material is based upon work supported by the National Science Foundation under Grant Nos. CMMI-0928315 and EEC-0914790. We would also like to acknowledge Professor Patricia Morris (OSU) for the use of her contact angle goniometer. References 1 N. B. La Thangue and D. J. Kerr, Nat. Rev. Clin. Oncol. , 2011 , 8 , 587 – 596 . Crossref Search ADS PubMed 2 D. J. Slamon , B. Leyland-Jones, S. Shak, H. Fuchs, V. Paton, A. Bajamonde, T. Fleming, W. Eiermann, J. Wolter, M. Pegram, J. Baselga and L. Norton, New Engl. J. Med. , 2001 , 344 , 783 – 792 . Crossref Search ADS 3 Early Breast Cancer Trialists' Collaborative Group , Lancet , 2005 , 365 , 1687 – 1717 . Crossref Search ADS PubMed 4 R. Stupp , W. P. Mason, M. J. van der Bent, M. Weller, B. Fisher, M. J. B. Taphoorn, K. Belanger, A. A. Brandes, C. Marosi, U. Bogdahn, J. Curschmann, R. C. Janzer, S. K. Ludwin, T. Gorlia, A. Allgeier, D. Lacombe, J. G. Cairncross, E. Eisenhauer and R. O. Mirimanoff, New Engl. J. Med. , 2005 , 352 , 987 – 996 . Crossref Search ADS 5 R. L. Schilsky , Nat. Rev. Drug Discovery , 2010 , 9 , 363 – 366 . Crossref Search ADS 6 B. M. Fine and L. Amler, Clin. Pharmacol. Ther. (N. Y., NY, U. S.) , 2009 , 85 , 535 – 538 . Crossref Search ADS 7 M. J. Duffy and J. Crown, Clin. Chem. (Washington, DC, U. S.) , 2008 , 54 , 1770 – 1779 . Crossref Search ADS 8 T. S. Mok , Nat. Rev. Clin. Oncol. , 2011 , 8 , 661 – 668 . Crossref Search ADS PubMed 9 D. Bonnet and J. E. Dick, Nat. Med. (N. Y.) , 1997 , 3 , 730 – 737 . Crossref Search ADS 10 A. Molino , M. Colombatti, F. Bonetti, M. Zardini, F. Pasini, A. Perini, G. Pelosi, G. Tridente, D. Veneri and G. L. Cettto, Cancer , 1991 , 67 , 1033 – 1036 . Crossref Search ADS PubMed 11 L. Patrawala , T. Calhoun, R. Schneider-Broussard, H. Li, B. Bhatia, S. Tang, J. G. Reilly, D. Chandra, J. Zhou, K. Claypool, L. Coghlan and D. G. Tang, Oncogene , 2006 , 25 , 1696 – 1708 . Crossref Search ADS PubMed 12 R. Handgretinger , P. Lang, M. Schumm, G. Taylor, S. Neu, E. Koscielnak, D. Niethammer and T. Klingebiel, Bone Marrow Transplant. , 1998 , 21 , 987 – 993 . Crossref Search ADS PubMed 13 M. Geens , H. Van de Velde, G. De Block, E. Goossens, A. Van Steirteghem and H. Tournaye, Hum. Reprod. , 2007 , 22 , 733 – 742 . Crossref Search ADS PubMed 14 S. Y. Yang , S. K. Hsiung, Y. C. Hung, C. M. Chang, T. L. Liao and G. B. Lee, Meas. Sci. Technol. , 2006 , 17 , 2001 – 2009 . Crossref Search ADS 15 C. Hsu , D. Di Carlo, C. Chen, D. Irimia and M. Toner, Lab Chip , 2008 , 8 , 2128 – 2134 . Crossref Search ADS PubMed 16 S. Zheng , H. Lin, J. Liu, M. Balic, R. Datar, R. J. Cote and Y. Tai, J. Chromatogr., A , 2007 , 1162 , 154 – 161 . Crossref Search ADS 17 F. F. Becker , X. B. Wang, Y. Huang, R. Pethig, J. Vykoukal and P. R. Gascoyne, Proc. Natl. Acad. Sci. U. S. A. , 1995 , 3 , 860 – 864 . Crossref Search ADS 18 P. R. C. Gascoyne , J. Noshari, T. J. Anderson and F. F. Becker, Electrophoresis , 2009 , 30 , 1388 – 1398 . Crossref Search ADS PubMed 19 M. P. MacDonald , G. C. Spalding and K. Dholakia, Nature , 2003 , 426 , 421 – 424 . Crossref Search ADS PubMed 20 D. R. Coman , Cancer Res. , 1944 , 4 , 625 – 629 . 21 K. W. Kwon , S. S. Choi, S. H. Lee, B. Kim, S. N. Lee, M. C. Park, P. Kim, S. Y. Hwang and K. Y. Suh, Lab Chip , 2007 , 7 , 1461 – 1468 . Crossref Search ADS PubMed 22 Y. Kawamoto , A. Naoko, Y. Ito, N. Wada and M. Kaibara, J. Mater. Sci.: Mater. Med. , 1997 , 8 , 551 – 557 . Crossref Search ADS PubMed 23 H. M. Powell , D. A. Kniss and J. J. Lannutti, Langmuir , 2006 , 22 , 5087 – 5094 . Crossref Search ADS PubMed 24 C. Wen , M. Chuang and G. Hsiue, Thin Solid Films , 2006 , 503 , 103 – 109 . Crossref Search ADS 25 A. Martins , E. D. Pinho, S. Faria, I. Pashkuleva, A. P. Marques, R. L. Reis and N. M. Neves, Small , 2009 , 5 , 1195 – 1206 . PubMed 26 J. Yang , G. Shi, J. Bei, S. Wang, Y. Cao, Q. Shang, G. Yang and W. Wang, J. Biomed. Mater. Res. , 2002 , 62 , 438 – 446 . Crossref Search ADS PubMed 27 A. Dekker , K. Reitsma, T. Beugeling, A. Bantjes, J. Feijen and W. G. Vanaken, Biomaterials , 1991 , 12 , 130 – 138 . Crossref Search ADS PubMed 28 S. Siri , P. Wadbual, V. Amornkitbamrung, N. Kampa and S. Maensiri, Mater. Sci. Technol. , 2010 , 26 , 1292 – 1297 . Crossref Search ADS 29 S. W. Myung , C. S. Lee and H. S. Choi, J. Colloid. Interface Sci. , 2006 , 295 , 409 – 416 . Crossref Search ADS PubMed 30 E. D. Yilidirim , M. Gandhi and A. Fridman, NATO Science for Peace and Security Security Series A: Chemistry and Biology , 2008 , pp. 191 – 201 . 31 Y. Momose , M. Noguchi and S. Okazaki, Nucl. Instrum. Methods Phys. Res., Sect. B , 1989 , 39 , 805 – 808 . Crossref Search ADS 32 M. G. McCord , Y. J. Hwang, Y. Qui, L. K. Hughes and M. A. Bourham, J. Appl. Polym. Sci. , 2003 , 88 , 2038 – 2047 . Crossref Search ADS 33 T. L. Sterrett , R. Sachdeva and P. Jerabek, J. Mater. Sci.: Mater. Med. , 1992 , 3 , 402 – 407 . Crossref Search ADS 34 Q. F. Wei , W. D. Gao, D. Y. Hou and X. Q. Wang, Appl. Surf. Sci. , 2005 , 245 , 16 – 20 . Crossref Search ADS 35 D. Sun and G. K. Stylios, J. Mater. Process. Technol. , 2006 , 173 , 172 – 177 . Crossref Search ADS 36 H. Lee , Y. Jeong, S. Jeong, S. Park, S. Bae, H. Kim and C. Cho, Appl. Surf. Sci. , 2008 , 254 , 5700 – 5705 . Crossref Search ADS 37 P. Roach , D. Eglin, K. Rohde and C. Perry, J. Mater. Sci.: Mater. Med. , 2007 , 18 , 1263 – 1277 . Crossref Search ADS PubMed 38 C. J. Wilson , R. E. Clegg, D. I. Leavesley and M. J. Pearcy, Tissue Eng. , 2005 , 11 , 1 – 18 . Crossref Search ADS PubMed 39 B. J. Papenburg , E. D. Rodrigues, M. Wessling and D. Stamatialis, Soft Matter , 2010 , 6 , 4377 – 4388 . Crossref Search ADS 40 K. E. Michael , V. N. Vernekar, B. G. Keselowsky, J. C. Meredith, R. A. Latour and A. J. Garcia, Langmuir , 2003 , 19 , 8033 – 8040 . Crossref Search ADS 41 P. Roach , D. Farrar and C. C. Perry, J. Am. Chem. Soc. , 2005 , 127 , 8168 – 8173 . Crossref Search ADS PubMed 42 Z. Gugala and S. Gogolewski, J. Biomed. Mater. Res., Part A , 2006 , 76 , 288 – 299 . Crossref Search ADS 43 C. M. Alves , Y. Yang, D. L. Carnes, J. L. Ong, V. L. Sylvia, D. D. Dean, C. M. Agarwal and R. L. Reis, Biomaterials , 2007 , 28 , 307 – 315 . Crossref Search ADS PubMed 44 J. Wei , M. Yoshinari, S. Takemoto, M. Hattori, E. Kawada, B. Liu and Y. Oda, J. Biomed. Mater. Res., Part B , 2007 , 81 , 66 – 75 . Crossref Search ADS 45 R. I. Chen , N. D. Gallant, J. R. Smith, M. J. Kipper and C. G. Simon, J. Mater. Sci.: Mater. Med. , 2008 , 19 , 1759 – 1766 . Crossref Search ADS PubMed 46 J. M. Lee , S. Dedhar, R. Kalluri and E. W. Thompsom, J. Cell Biol. , 2006 , 172 , 973 – 981 . Crossref Search ADS PubMed 47 L. Lynch , P. I. Vodyanik, G. Boettiger and M. A. Guvakova, Mol. Biol. Cell , 2005 , 16 , 51 – 63 . Crossref Search ADS PubMed 48 D. Gallego-Perez , N. Higuita-Castro, S. Sharma, R. K. Reen, A. F. Palmer, K. J. Gooch, L. J. Lee, J. J. Lannutti and D. J. Hansford, Lab Chip , 2010 , 10 , 775 – 782 . Crossref Search ADS PubMed 49 N. Ferrell , D. Gallego-Perez, N. Higuita-Castro, R. T. Butler, R. K. Reen, K. J. Gooch and D. J. Hansford, Anal. Chem. , 2010 , 82 , 2380 – 2386 . Crossref Search ADS PubMed This journal is © The Royal Society of Chemistry 2012
TI - Plasma surface modification of electrospun fibers for adhesion-based cancer cell sorting
JF - Integrative Biology
DO - 10.1039/c2ib20025b
DA - 2012-08-21
UR - https://www.deepdyve.com/lp/oxford-university-press/plasma-surface-modification-of-electrospun-fibers-for-adhesion-based-nKMLRCaSFc
SP - 1112
VL - 4
IS - 9
DP - DeepDyve
ER -