This protocol describes customizable surface functionalization of the desthiobiotin, streptavidin, and APTES system in order to isolate specific cell types of interest. In addition, this manuscript covers the applications, optimization, and verification of this process.
One of the limiting factors to the adoption and advancement of personalized medicine is the inability to develop diagnostic tools to probe individual nuances in expression from patient to patient. Current methodologies that try to separate cells to fill this niche result in disruption of physiological expression, making the separation technique useless as a diagnostic tool. In this protocol, we describe the functionalization and optimization of a surface for the cellular capture and release. This functionalized surface integrates biotinylated antibodies with a glass surface functionalized with an aminosilane (APTES), desthiobiotin and streptavidin. Cell release is facilitated through the introduction of biotin, allowing the recollection and purification of cells captured by the surface. This release is done through the targeting of the secondary moiety desthiobiotin, which results in a much more gentle release paradigm. This reduction in harsh reagents and shear forces reduces changes in cellular expression. The functionalized surface captures up to 80% of cells in a single cell mixture and has demonstrated 50% capture in a dual-cell mixture. Applications of this technology to xenografts and cancer separation studies are investigated. Quantification techniques for surface verification such as plate reader and ImageJ analyses are described as well.
Current bench-top cell separation approaches (e.g., fluorescence activated cell sorting1, laser capture micro-dissection2, immuno-magnetic bead separation1) can take several hours of preparation and sorting. These large time scales can affect physiological response and expression levels, resulting in analyses that are not representative of the physiological response3. Systems are needed that can rapidly and efficiently isolate specific cell types without disrupting cell-surface receptor-levels in order to improve cell isolation and enrichment for biomedical applications. Therefore, the rationale for our approach is to develop a gentle approach for cell isolation.
The "lab on a chip" concept offers the promise of orders of magnitude quicker (hours-to-minutes) cell isolation, and most frequently involves capturing cells onto a surface and releasing cells or intracellular contents through physical4,5 or chemical methods6. Although these approaches offer a few advantages such as identifying protein7,8 expression, identifying RNA expression9-11, or even providing cells for in vitro culture12,13, many of these techniques cannot be translated to diagnostics such as cell receptor profiling due to their non-physiological environments. Enzymatic lifting agents such as collagenases can also affect these receptor quantities14,15, meaning cell receptor quantification techniques that use these lifting agents will not generate accurate physiological data. Cellular lysis prevents differentiation between the native surface receptors, and those which were previously internalized16. This protocol describes a fast and gentle approach for cell isolation.
1. Cleaning the Glass Surface and Preparing Reagents
2. APTES and DSB Functionalization
3. Streptavidin Functionalization
4. Cell Capture and Release
5. Antibody Optimization: Antibody Titration
6. Cell Optimization: Cell Titration
7. Image Analysis
Note: The FIJI software package (http://fiji.sc/Fiji) is recommended for image analysis. Initially, the images were converted into grayscale image, and then the brightness/contrast was altered to bring out the cells.
Using this protocol we show cell capture (Figure 3A) and cell release (Figure 3C) of MCF7GFP cells as well as live cell controls (Figure 4). We quantified the cell capture as 60% and 80% were released (Figure 3C). When we extended this approach to a mixture of RAW 264.7 macrophages and MCF7GFP cells, 50% of RAW macrophages were captured (Fig. 3D) and 80% of RAW macrophages were release with 20 mM biotin (Figure 3B). Since excess antibody can decrease cell capture, we optimized via titrating 0-10,000 nM of HLA antibody, and observe that the ideal antibody concentration is between 100-1,000 mM antibody (Fig. 1). Similarly, we determine that the ideal concentration of cells that can be captured is 1 x 105 and 1 x 106, since below that number of cells, values are lower than the background17 (Figure 2). The fluorescence data from each well is processed as described below.
Here, the average of 3 replicates is taken from a sample with no antibody (blank) and is subtracted from the fluorescence obtained from each sample. This value is then normalized to the average maximal fluorescence. This processing allows the researcher to clearly observe the low limit of cellular detection.
Figure 1. Normalized Blanked Antibody Titration. Human HLA-ABC is titrated across a constant quantity of MCF7GFP cells (125,000 cells per well) and the fluorescence of captured cells were quantified using a plate reader. Prior to exposure to the functionalized surface, the cells and antibodies were centrifuged to remove non-specific antibody attachment. Without this centrifugation, excess antibody will saturate surface and prevent cell pulldown, as shown with concentrations of 10,000 ng/ml where centrifugation was not sufficient to prevent antibody oversaturation of the surface, resulting in less cells binding to the surface. Gray line represents the control. Error bars represent standard error of the mean. Please click here to view a larger version of this figure.
Figure 2. Normalized Blanked Cell Titration. Using optimized antibody concentration from the antibody titration, MCF7GFP cells were titrated to find optimized capture concentration while keeping functionalized surface and antibody concentrations constant. This can be used to calibrate cell capture for different applications. Columns in this figure are replicates, while rows change the concentration of cells to find the idea range for cellular capture. Gray line represents control, and error bars represent standard error of the mean. Please click here to view a larger version of this figure.
Figure 3. Capture and Release Experiments. A single cell mixture was exposed to the functionalized surface (A), capturing MCF7GFP cells using HLA-ABC antibody. When washed with HBSS (B), cells remained captured, but when exposed to a solution of 20 mM biotin (C), cells were released. A cellular mixture of MCF7GFP cells and RAW 264.7 macrophages were exposed to a functionalized surface and were captured using mCD11b antibody (D). Non-specific MCF7GFP cells are labeled in green. Fluorescence intensity of the MCF7GFP cells decreased when exposed to a neutral wash (E), which implies that the surface did not target them, and that they attached non-specifically. When exposed to a biotin wash (F), the targeted RAW macrophages were released from the surface. Scale bars are 250 µm. Please click here to view a larger version of this figure.
Figure 4. Positive and Negative Cell Control for Dual Cell Isolation. The positive control RAW Macrophages are imaged on a fluorescence microscope, showing no fluorescent activity as well as the relative sizing of cells. Brightfield RAW macrophages show cell sizing (A), while under GFP excitation, there is no fluorescence (B), which is shown in the merged image (C). The negative control MCF7GFP cells are imaged on a fluorescence microscope showing large fluorescent activity, as well as the MCF7GFP cells' relative size. Cells are imaged on brightfield (D), showing sizing, and while under GFP excitation (E) show large fluorescence, which can be clearly shown in the merged image (F). Scale bars are 250 µm. Please click here to view a larger version of this figure.
Figure 5. Comparison of centrifuge usage in standard purification techniques. The pre-analysis centrifugation step is conserved in sample preparation. Magnetic bead isolation as well as FACS involves several additional centrifugation steps, which introduces stress onto cells and may alter expression levels. Please click here to view a larger version of this figure.
Supplemental Code File. Calculations. Please click here to download this file.
Improvements in cell isolation techniques furthers scientific studies in structure-function relationships in neuroscience18, stem cell programming in regenerative biology, and angiogenic signaling in vascular biology19. Indeed, primary cell culture20 (e.g., HUVECs) in vascular biology is primarily done through the use of cell isolation techniques. Cell isolation was also recently used for quantitative flow (qFlow) cytometry analysis of plasma membrane receptors3,14,15,19,21. However, existing cell isolation methodologies affect cell-surface receptor levels and are costly in both personnel and reagents. We have advanced a new method in surface functionalization17, which allows for the creation of a gentle system of reproducible capture of a single cell type from a mixture of cell types to meet these shortcomings. This technique can be integrated into an iterative system, allowing the possibility of capturing different cell types in stages using specific antibodies. To further clarify the procedure, a number of troubleshooting questions and answers are offered below:
Functionalization Optimization: The functionalization processes has been optimized for improved uniformity and pull down of MCF7gfp cells. This protocol can alternatively be optimized and customized for any cell type and antibody configuration. This can be achieved through altering the concentrations of the base components in the capture surface, or by changing the type or concentrations of the antibodies. Most antibody types can be biotinylated using the above protocol. Once biotinylated, they can then be titrated (Figure 1) to find ideal concentration for pull down. The cell concentration can be titrated (Figure 2) to find the ideal concentration of cells to capture. Using this, the feasibility of capturing a certain cell type with the capture surface can be ascertained. For example, if a cell type of interest is on the scale of 100 cells per million cells, capturing concentrations of cells in that range using the functionalized surface would be ideal. If during the titration, the surface cannot capture cells of that small a concentration, then optimization of the surface or antibody is necessary in order to be able to filter out cells within that concentration range.
Which cell types are used in this protocol? How could someone modify this protocol to use a different cell type?
Currently, this protocol uses MCF7-GFP cells, and RAW 264.7 macrophages are often included to show an ability to pull out one cell type in a dual-cell mixture. These cells were selected as they were two of the most relevant cell types to mouse xenograft model (human tumor within murine environment). In order to calibrate the process for a variety of other cells, antibody specificity is paramount. There are several online resources available for selecting antibodies22.
What are some advantages of this method?
Gentle approach: The lack of force makes this approach gentle. Indeed, our calculated shear force is maximally, 24 x 10-6 pN17, which should not cause significant disruption to biomarkers, since hydrodynamic stresses of 2.09 Pa (656 pN assuming 314 pM2 cell surface area) induce necrosis while values of stress below 0.59 Pa (185 pN assuming 314 pM2 cell surface area) do not23. This gentle approach is thus advantageous over some commercially available options such as centrifugation-based approaches24, which exert up to 0.78 Pa25 of peak shear stress due to the sudden acceleration at values around 600 G, which may change protein expression patterns and morphologies25,26 and even induce cell necrosis23. Thus a process that could reduce the amount of centrifuge usages would reduce the amount of stresses that the cells would experience through purification and ultimately ensure more physiological data. Our current protocol uses centrifuges to purify and reconstitute the cell concentration prior to capture, however, this preparation process is standard in many purification techniques such as magnetic bead isolation27,28, flow cytometry29, and FACS30 (Figure 5). Our approach reduces all downstream centrifugation processes that the other techniques use in addition to the preparation step.
Biotin-avidin approach: The application of commonly used biomaterials (DSB-SAv and biotin-SAv) also renders this approach advantageous31,32. Avidin family proteins bind extremely selectively to the biotin family proteins and are used for a range of scientific and medical applications including: antibody-fluorophore attachment33, quantitative Qdot-polystyrene bead attachment34, and the creation of hydrogels that respond to stimuli in the environment to release encapsulated drugs32. Additionally, the relative ease of acquiring biotinylated antibodies makes this approach widely accessible and customizable.
Desthiobiotin approach without beads: The capture surface is able to implement the use of desthiobiotin without the use of harsh reagents and forces to release the cells. DSB has been used for reversible cell attachment and release by other systems in conjunction with DSB-antibodies and magnetic beads35. However, the harsh effects of the separation technique as well as the large cellular loss associated with preparing the samples27,28 result in differential receptor and chemokine expression28,36. This approach aims to mitigate and overcome these limitations by eliminating the use of both magnet and beads to create a much more gentle wash and release methodology.
What are the critical steps in this protocol? What can cause variability? How can that variability be controlled?
This process has four critical steps that can result in the decreased efficiency of the capture surface. The first critical step is the prevention of water to the APTES surface during the APTES functionalization steps, which would result in the destruction of the self-assembled surface. This is remedied by using ethanol as a solvent and baking the APTES in the oven to reduce hydrolysis from later aqueous solutions. The second critical step is the EDC reaction steps in which EDC catalyzes the reaction of the DSB with the APTES: allowing cross-linking of the two layers. If the EDC is excluded or not sufficiently added, the DSB will not be able to attach the floor, which will compromise the release mechanism. The third critical step is during the SAv functionalization, as maintaining the SAv layer is critical. Non-uniformity of the streptavidin layer results in the reduction in capture efficiency. The fourth critical step is in the biotinylation of the antibody, as the binding of the antibody to the capture surface is facilitated via the biotin-streptavidin interaction of the capture surface and antibody. If the antibody is non-biotinylated, then the streptavidin layer will be unable to capture it and pull it down, making the capture surface useless. Variability and inconsistency in the surface capture can be attributed to concentration non-regularity as well as stochasticity of layer attachment. This variation can be controlled by using precise concentrations of layer components calibrated to the surface and the intended targets of capture. In this design, concentrations are calibrated specifically to the capture of MCF7-GFP cells.
What are some drawbacks of this method?
Currently, the APTES functionalization is done by using a liquid phase immersion of the surface for 55 minutes followed by a baking step for up to 2 hours. Although this allows for a complete layer of APTES silane to bind to the surface, a more uniform process would be the vapor phase silanization37. This decreases APTES contact time, thus saving the researcher time, as well as increasing uniformity. Additionally, improved pull-down has been observed17 when overnight incubation of SAv is used, and currently, the surface functionalization requires two, overnight incubations, making the overall process take three days. So, optimizing the chemistry would offer significant time-savings for the researcher.
What are some warnings or precautions that may be helpful?
When functionalizing surfaces, it is imperative that proper care is taken to the risk of respiratory damage. Both APTES and mercaptoethanol can be extremely dangerous to the lungs and associated organs, thus it is necessary to do the functionalization process in a chemical hood to reduce interaction with the chemical fumes. Mercaptoethanol is a thiol which is especially pungent in odor. When using mercaptoethanol, allow for contaminated waste to sit in hood for a day or two before disposal.
What are the future aims and applications of this surface?
Our future aims involve customizing this surface to work with a variety of relevant cell types for angiogenic related diseases. Particularly, we intend to focus on Human Umbilical Vein Endothelial Cells in order to segue to using blood samples and separating out cells of interest from there. Additionally, we plan to integrate separation modalities such as aptamers into the design of the functionalized surface to further increase our capture and release percentages.
The authors have nothing to disclose.
We would like to thank the American Cancer Society, Illinois Division (282802) and the National Science Foundation CBET (1512598) for funding support. We also would like to thank Dr. Dianwen Zhang from the University of Illinois Beckman Institute for microscopy training. Finally, we would like to thank Jared Weddell, Stacie Chen, and Spencer Mamer for insightful discussions.
(3-Aminopropyl) triethoxysilane (APTES) | Acros Organics | 919-30-2 | Used to make 2% APTES solution |
Plasma Cleaner Pico | Diener | Model 1 | Cleans surfaces and allows for bonding of PDMS to glass |
d-Desthiobiotin (DSB) | Sigma | D20655 | Used as the releasing mechanism in the cellular capture surface. |
dimethyl sulfoxide (DMSO) | British Drug Houses (BDH) | BDH1115-1LP | Dissolves the DSB into solution |
1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) | Thermo-Scientific | 5g: 22980 25g: 22981 |
Activates Carboxylic Acids and allows binding of proteins to glass surface. |
uncoated 8-well culture slide | BD Falcon | Case of 24: 354118 Case of 96: 354108 |
Used in cellular experiments involving Zeiss fluorescence microscope such as initial capture and release quantification experiments |
Glass bottom 24-well plates | MatTek | P24G-0-13-F | Used in cellular experiments involving the plate reader such as antibody and cellular titration experiments |
Mercaptoethanol | Science Lab | 60-24-2 | Used to quench reaction between EDC and DSB |
4-Morpholinoethanesulfonic acid hydrate (MES Hydrate 99%) |
Fisher Scientific | AC172590250 | Used to make 0.1 M MES Buffer for use in EDC reaction |
Precision Oven | Thermo Scientific | 11-475-153 | Used in curing of PDMS and APTES layer. |
Titramax 1000 Shaker | Heidolph | 13-889-420 | Used to ensure even distribution of APTES on surfaces. |
1X Streptavidin 5mg [e7105-5mg] |
Proteo Chem | 9013-20-1 | Biotin-binding protein May cause irritation |
5 cm Glass Dish | Fisher Scientific | 08748A | Used in HUVEC studies as well as future profiling studies. |
14 cm Petri Dish with Cover | Sigma-Aldrich | Z717231 | Used to hold samples being functionalized and transport them. |
MCF7-GFP cells | Cell Biolabs | AKR211 | Stored in liquid nitrogen |
RAW264.7 mouse macrophages |
ATCC | TIB-71 | Gifted to us from Smith lab at the University of Illinois. Stored in liquid nitrogen |
TrypLE | Life Technologies | 12605036 | Stored in 100mL at room temperature |
Dulbecco’s modified Eagle medium | Cell Media Facility at School of Chemical Sciences at UIUC | 50003PC | Supplier: Corning |
Nonessential amino acids | Cell Media Facility at School of Chemical Sciences at UIUC | 25-025-CI | Already added into DMEM by facility. Supplier: Corning |
10% fetal bovine serum | Fisher Scientific | 03-600-511 | Stored in 500mL at < -10⁰C |
1% Penicillin–Streptomycin | Life Sciences Storeroom at UIUC | 17602e | Supplier: VWR Stored in 100 ml at 4⁰C |
Cell scraper | Fisher Scientific | 12-565-58 | Small 23cm 50 pack |
Cell Dissociation Solution | Corning | MT-25-056CI | Used to lift cells non-enzymatically for the use in cell experiments |
Hemacytometer | Hausser | 02-671-54 | Used to count cells for quantification of cell solutions and capture and release effectivity. |
Biotin | Amresco | 58-85-5 | Used to release cells from surface. |
HBSS | Created from Recipe | N/A | Used to keep cells alive in suspension as well as wash surfaces of non-specific binding. (Adapted from Cold Spring Harbor Protocols): In 500 mL, use 4 g NaCl, .2 g KCl, .0402 g Na2PO4*7H2O, .03 g KH2PO4 and .5 g Glucose. Add DI water to get to 500 mL, filter, and then refrigerate. |
HLA-ABC Antibody | BioLegend | 311402 | Antibody used to capture MCF7gfp cells |
hIgG Antibody | BioLegend | HP6017 | Antibody used to capture MCF7gfp cells |
MCF7 GFP cells | Cell Biolabs | AKR-211 | Luminal Breast Cancer line that has been transfected with green fluorescent protein. |
Assorted Conicals | Thermo-Scientific | 15mL: 12-565-268 | 50/15 mL plastic conicals for storing solutions and aliquots. |
Mini-Tube Rotators (End over End Mixer) | Fisher Scientific | 05-450-127 | Used to incubate antibody and mix other cellular solutions in order to mix |
Axiovert 200M (Fluorescence Microscope) | Zeiss | N/A | Zeiss Axiovert 200 M inverted florescence microscope. |
Zeba Desalting columns | Thermo-Scientific | PI-87770 | Used to purify newly biotinylated antibodies after the use of the Biotinylation Kit. Instructions provided at: http://www.funakoshi.co.jp/data/datasheet/PCC/89894.pdf |
EZ Link Sulfo NHS Low Weight Biotinylation Kit | Thermo- Scientific | Used to biotinylate antibodies to allow them to integrate with the capture surface | |
Plate Reader | BioTek | Synergy HTX Multimode Reader | Used to quantitatively measure fluorescent intensity in the titration experiments. |