Catch And Release Cell Sorting: Electrochemical Desorption Of T

Transcript

Available online at www.sciencedirect.com Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 Catch and release cell sorting: Electrochemical desorption of T-cells from antibody-modified microelectrodes He Zhu 1 , Jun Yan 1 , Alexander Revzin ∗ Department of Biomedical Engineering, University of California, Davis, 451 East Health Sciences Street #2619, Davis, CA 95616, United States Received 14 December 2007; received in revised form 3 February 2008; accepted 4 February 2008 Available online 10 March 2008 Abstract The development of integrated microsystems capable of interrogation, characterization and sorting of mammalian cells is highly significant for further advancement of point-of-care diagnostics and drug discovery fields. The present study sought to design a novel strategy for releasing antibody-bound cells through electrochemical disruption of the underlying antibody (Ab) layer. A microsystem for selective capture and release of cells consisted of an array of individually addressable gold microelectrodes fabricated on a glass substrate. Poly(ethylene glycol) (PEG) hydrogel photolithography was employed to make the glass regions non-fouling, thus, ensuring selective localization of proteins and cells on the microelectrodes. The gold surfaces were decorated with anti-CD4 Ab molecules using standard alkanethiol self-assembly and carbodiimide coupling approaches. The Ab-functionalized electrodes selectively captured model T-lymphocytes (Molt-3 cells) expressing CD4 antigen while minimal cell adhesion was observed on PEG hydrogel-modified glass substrates. Importantly, application of a reductive potential (−1.2 V vs. Ag/AgCl reference electrode) resulted in release of surface-bound T-cells from the electrode surface. Cyclic voltammetry and fluorescence microscopy were employed to verify that the detachment of captured T-cells was indeed due to the electrochemical disruption of the underlying alkanethiol-Ab layer. In the future, the cell sorting approach described here may be combined with microfluidic delivery to enable Ab-mediated capture of T-lymphocytes or other cell types followed by release of select cells for downstream gene expression studies or re-cultivation. © 2008 Elsevier B.V. All rights reserved. Keywords: Microfabrication; Cell micropatterning; Switchable biointerface; Microelectrodes; Cell sorting; Leukocyte immunophenotyping 1. Introduction Precise engineering of cell–substrate interactions is an important component in the development of microsystems for isolation and interrogation of living cells. While a number of surface modification approaches have been reported, the self-assembled monolayers (SAMs) of alkanethiols on gold arguably offer the most control and flexibility in designing the desired biointerfacial properties. Formation of SAMs with terminal ethylene glycol groups has been shown to render gold surfaces nonadhesive to cells or proteins, while assembly of alkanethiols with aliphatic or polar groups led to the formation of cell-adhesive biointerfaces [1–4]. In addition to designing biointerfacial properties through the choice of the end-functional groups, a significant empha- ∗ 1 Corresponding author. Tel.: +1 530 752 2383; fax: +1 530 754 5739. E-mail address: [email protected] (A. Revzin). Both authors contributed equally. 0927-7765/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2008.02.010 sis has been made on modulating spatial composition of SAMs by employing microfabrication approaches. Soft lithography [5] as well as traditional photolithography [6,7] have been employed to create substrates with micrometer-scale control over the spatial (in-plane) surface properties. While in the majority of the surface engineering approaches SAMs and proteins were micropatterned on substrates comprised of a single type of material, several reports suggested that the composition of the substrate as well as the choice of adsorbates may be used to define the in-plane micrometer scale properties [8–10]. For example, Laibinis et al. demonstrated selective formation of monolayers of alkanethiols and alkane carboxylic acids on regions of Au and alumina, respectively, that were microfabricated on the same substrate [8]. More recently, Zhang and co-workers fabricated silicon substrates containing gold micropatterns and modified these composite surfaces using alkanethiol assembly on gold and alkoxysilane assembly on silicon [10]. This surface engineering approach has been employed to create arrays of single cells for biosensor applications [9]. H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 In addition to being advantageous from the stand-point of biointerface design, gold is an electrode material of choice due to its excellent conductivity and chemical inertness. A number of approaches have utilized the electrical activity of gold in order to manipulate its interfacial properties in a spatiotemporal fashion [11–14]. Mrksich and co-workers synthesized alkanethiol molecules that became cell-adhesive upon applying an oxidative potential to the underlying gold substrate and employed this active biointerface design in assembling two different cell populations on the same substrate [13]. Recently, the same group proposed a more complex strategy whereby three types of alkanethiols, each decorated with cell-adhesive ligands and possessing unique redox properties, were micropatterned on the same gold substrate [11]. Applying a specific potential allowed detaching the desired alkanethiol molecules along with bound cells. While electrochemical modulation of the chemical structure of alkanethiols is an elegant approach that allows sophisticated manipulation of biointerfacial properties [11,13,14], the synthesis of electroactive alkanethiols is complicated. Desorption of whole alkanethiol monolayers through electrochemical reduction of the thiolate (Au–S) bond offers a simpler avenue for controlling biointerfacial properties of gold electrodes. First reported by Widrig et al. [15], electrochemical desorption of alkanethiols has been employed to release cells from geometric confinement [16] and has recently been used to pattern three distinct cell types on the same surface [17]. The present paper describes capture and release of Tlymphocytes using an array of individually addressable Ab-modified gold microelectrodes fabricated on a glass substrate. The key design criteria established in this study were: development of a novel approach for integrating gold microelectrodes with a non-fouling biomaterial and demonstration of a strategy for releasing Ab-bound T-cells from electrodes. Biointegration or “packaging” of gold microelectrodes was achieved using poly(ethylene glycol) (PEG) photolithography—a method for micropatterning proteins and cells on glass substrates described by us previously [18–21]. Micropatterning gold electrodes on glass created a substrate with reflective metal patterns on transparent glass background. This substrate served as its own photomask, permitting “self-alignment” of gold and PEG gel microstructures through a simple backside flood exposure and, therefore, obviating the need for registration. During the subsequent anti-CD4 Ab-decoration and T-lymphocyte capture steps, the non-fouling PEG microstructures guided protein and cell adhesion exclusively to the gold electrodes. The T-cell release strategy employed here was based on electrochemical desorption of proteins from electrode surfaces reported by us previously [22]. Model T-lymphocytes expressing CD4 antigen were captured onto arrays of microelectrodes decorated with anti-CD4 Abs and packaged in PEG gel layer. Applying reductive potential (−1.2 V vs. Ag/AgCl reference) to the desired electrode resulted in disruption of the underlying Ab layer and detachment of Ab-bound T-cells only from actuated members of an electrode array. Overall, fabrication of individually addressable Ab-modified microelectrodes provided a simple and effective means of releasing surface-bound T-lymphocytes 261 with spatial and temporal control. In the future, this cell sorting strategy will be combined with microfluidic delivery to enable selective capture and subsequent release of cells for downstream analysis or re-cultivation. 2. Materials and methods 2.1. Materials 11-Mercaptoundecanoic acid (MUA), Nhydroxysuccinimide (NHS), 200-proof ethanol, Tris (hydroxymethyl) aminomethane buffer (Tris buffer), 99.9% toluene, poly(ethylene glycol) diacrylate (PEG-DA, MW 575), 2-hydroxy-2-methyl-propiophenone (photoinitiator), Tween 20, neutravidin conjugated with FITC and streptavidin conjugated with Alexa (for 2nd labeling) were purchased from Sigma–Aldrich (St. Louis, MO). 1-Ethyl3-(3-dimethylaminopropyl) carbodiimide (EDC) was from Pierce Biotechnology (Rockford, IL). Potassium ferricyanide, 10× phosphate buffered saline (PBS) without calcium and magnesium, trypan blue stain solution were from Fisher Scientific. Chromium etchant (CR-4S) and gold etchant (Au-5) were from Cyantek Corporation (Fremont, CA). Positive photoresist (AZ 5214-E IR) and its developer solution (AZ300 MIF) were brought from Mays Chemical (Indianapolis, IN). 3-Acryloxypropyl trichlorosilane, was purchased from Gelest, Inc. (Morrisville, PA). Biotinylated mouse anti-human CD4 Ab was from Beckman Coulter (Miami, FL). Molt-3 T-lymphocyte and 3T3 fibroblast cell lines were purchased from American Type Culture Collection (ATCC). RPMI-1640 and DMEM cell media were purchased from VWR. Fetal bovine serum (FBS) was from Invitrogen (Carlsbad, CA). 2.2. Fabrication of gold microelectrodes Standard (75 mm × 25 mm) glass slides coated with 15 nm Cr adhesion layer and 100 nm Au layer were purchased from Lance Goddard Associates (Foster City, CA). The substrates were O2 plasma treated for 5 min at 300 W and stored in an oven at 200 ◦ C prior to use. Au microelectrodes were fabricated using traditional photoresist lithography and wet etching processes. Briefly, positive resist (AZ 5214-E IR) was spin-coated at 800 rpm for 10 s followed by 4000 rpm for 30 s resulting in formation of a 4 ␮m thick layer of photoresist. The photoresist-coated glass slides were soft-baked on a hot-plate (Series 720, Barnstead International) at 100 ◦ C for 105 s, then placed in contact with a photomask and exposed to 365 nm, 10 mW/cm2 UV source for 55 s (Canon PLA-501F mask aligner). The substrates were then placed into a developer solution (AZ300 MIF) for 4 min. After development step, Au-coated glass slides were immersed in Au etching solution (1:3, v/v mixture of Au-5 etchant with H2 O) for ∼5 min, followed by 10 s immersion in Cr etching solution (1:1, v/v mixture of CR-4S etchant with H2 O). Metal was selectively removed from the regions not protected by photoresist, resulting in formation of Au micropatterns. Importantly, the photoresist layer was not removed immediately after etching 262 H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 Fig. 1. (A) PEG photolithography-based strategy for “self-alignment” of gold microelectrodes and hydrogel microstructures. Step 1: Gold microelectrodes are fabricated using standard photolithography and wet etching approaches. Step 2: PEG-DA prepolymer solution is spin-coated onto micropatterned substrates. Surfaces are silanized prior to spin-coating. Step 3: UV exposure from the backside results in formation of cross-linked PEG gel regions around the reflective microelectrodes. Step 4: Unpolymerized PEG residing on top of the electrodes is developed in water. (B) Schematic for T-cell capture and release. Step 1: Covalent binding of anti-CD4 Ab molecules onto gold microelectrodes via COOH-terminated thiols. Step 2: Capture of T-cells exclusively onto Ab-modified gold microelectrodes. Step 3: Applying reductive potential (1.2 V vs. Ag/AgCl reference) disrupts the underlying thiol layer and results in release of T-cells. but was employed to protect underlying Au regions during the silane modification protocol described below. 2.3. “Self-alignment” of PEG hydrogel microstructures with gold microelectrodes Prior to PEG gel immobilization, the glass substrates with photoresist-protected Au microelectrodes were modified using a silane coupling agent. The substrates were exposed to O2 plasma for 3 min at 300 W, placed into N2 filled glove bag and immersed in 2 mM solution of (3-acryloxypropyl) trichlorosilane in anhydrous toluene. The silane self-assembly was allowed to proceed for 1 h under N2 blanket, after which the substrates were removed, rinsed in toluene and dried using N2 gas. The substrates were then sonicated in acetone for 5 min to remove photoresist and placed in an oven for 3 h at 100 ◦ C to cure the silane layer. This silanization procedure has been used by us previously for anchoring PEG hydrogel microstructures to glass substrates [18,19]. The integration of Au microelectrodes with non-fouling PEG hydrogel micropatterns was performed as follows. Prepolymer solution, containing PEG-diacrylate (DA) (MW 575) and 2% (v/v) photoinitiator (2-hydroxy-2-methyl-propiophenone), was spin-coated at 800 rpm for 4 s onto glass slides containing Au electrode patterns. In contrast to traditional photolithographic registration, the uniform coat of prepolymer was flood-exposed from the back side (see Fig. 1A) with 365 nm, 5 mW/cm2 UV light source for 40 s. A substrate containing reflective metal regions and transparent glass background served as its own pho- tomask, causing photopolymerization and cross-linking of PEG prepolymer to occur around the electrodes (Fig. 1A). This “selfalignment” permitted integration Au electrodes and non-fouling PEG hydrogel regions without the need for fiduciary marks and registration. Regions of the prepolymer layer residing on metal patterns were not exposed to UV, did not become cross-linked and were removed from the substrate by developing in DI water for 5 min. 2.4. Decorating gold microelectrodes with antibodies To facilitate attachment of T-cells, gold electrodes were decorated with T-cell specific Abs. In order to avoid possibility of Ab denaturation due to heating, Cu wires were soldered to Au contact pads prior to thiol self-assembly and Ab immobilization steps. The electrode surfaces were then decorated with Ab molecules using self-assembly of mercaptoundecanoic acid (MUA) followed by activation of –COOH endgroups of this alkenthiol using EDC/NHS chemistry [22]. Briefly, Au electrodes with PEG hydrogel micropatterns were incubated with 1 mM MUA in 70% ethanol at room temperature for 18 h. After rinsing with copious amounts of 70% ethanol and DI water the self-assembled monolayer of MUA was activated with 0.4 M EDC/0.1M NHS in DI water for 15 min. Anti-human CD4 Ab solution (0.1 mg/ml in pH 4.5 sodium bicarbonate buffer) was reacted with activated MUA for 1 h. 50 nM Tris buffer was finally added to quench the remaining activated MUA. As an alternative to attaching Ab molecules directly to alkanethiols, neutravidin (0.1 mg/ml) was first immobilized on gold electrodes by a pro- H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 tocol identical to the one described above for Ab molecules. Biotinylated anti-CD4 Ab molecules were then incubated with neutravidin-treated Au microelectrodes for 1 h followed by rinsing with 1× PBS. 2.5. Patterning of cells on gold microdomains Molt-3 cells were cultured in suspension in RPMI 1640 media with 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin and 100 ␮g/ml streptomycin. The cells were spun down at 1200 rpm for 3 min and then re-suspended in 1× PBS to a final concentration of 5 × 106 cells/ml. Murine 3T3 fibroblasts were cultured in 175 cm2 tissue culture flasks at 37 ◦ C in humidified atmosphere with 10% CO2 /90% air. Fibroblasts were incubated in DMEM supplemented with 110 mg/l sodium pyruvate, 10% bovine calf serum, 200 U/ml penicillin and 200 ␮g/ml streptomycin. Cells were grown to pre-confluence and passaged by trypsinization in 0.05% trypsin/0.01% EDTA solution in PBS for 5 min at 37 ◦ C. The cell suspension was diluted 1:1 with fibroblast culture medium and centrifuged at 800 rpm for 5 min. After aspiration of the supernatant, cells were reconstituted in fresh fibroblast culture medium and counted using a hemocytometer. Fibroblast viability was typically better than 95% as determined by trypan blue exclusion. 3T3 fibroblasts were seeded onto micropatterned substrates in order to ascertain our ability to guide cell attachment to the electrode regions and to eliminate non-specific cell adhesion from happening elsewhere on the surface. In these experiments, substrates containing hydrogel-enveloped Au pads were cut into 25 mm × 25 mm pieces and placed into 35 mm diameter Petri dishes. When seeding fibroblasts, 2 ml of cell suspension at 1 × 106 cells/ml in growth medium with 10% FBS were introduced into the Petri dishes containing micropatterned glass templates. The cells were incubated with patterned surfaces for 2 h at 37 ◦ C. After incubation, growth medium with unattached fibroblasts was removed and replaced with 2 ml of fresh media. Fibroblasts captured within PEG microwells were incubated at 37 ◦ C for 12 h and then fixed in 1% glutaraldehyde (v/v) in 1× PBS for 30 min. To capture model T-lymphocytes on Ab-decorated surfaces, a glass slide containing an array of Ab-decorated microelectrodes enveloped in PEG hydrogel layer was outfitted with a silicone gasket (13.5 mm × 9.5 mm × 1.0 mm, Grace Bio-labs, OR) so as to incubate cells with the region of the substrate containing microelectrodes. A 100 ␮l volume of Molt-3 cells suspended in 1× PBS at a density of 5 × 106 cells/ml was dispensed into a gasket. After incubating for 15 min at room temperature, a substrate with adherent T-cells was carefully transferred into a 100 mm diameter Petri dish containing 1× PBS and was washed by gentle agitation of the dish. Optical imaging of cells bound on the micropatterned surface was carried out using Nikon Eclipse LV 100. High-resolution images of the hydrogel microstructures were obtained using a JSM 5600LV scanning electron microscope (SEM) (JEOL Inc., Peabody, MA) operating at 10 mV accelerating voltage. In order to avoid charging effects, substrates were sputter-coated with gold-palladium to a thickness of 10 nm prior to SEM experiments. The same protocol was 263 followed for preparation and imaging of samples containing Au-PEG microstructures with or without the cells. 2.6. Desorbing Abs from microelectrodes Reductive desorption of immobilized Abs was carried out in a custom-made Plexiglass electrochemical cell using a three electrode system. Micropatterned Au regions serving as working electrodes were enclosed inside an electrochemical cell, creating a volume of ∼1 ml where 1× PBS served as an electrolyte solution. Ag/AgCl reference and Pt counter electrodes were positioned above the working electrode in the same volume. A potentiostat (CH Instruments, TX) was used to apply a reductive potential of (−1.2 V vs. Ag/AgCl) for 60 s to individually addressable Au regions. Cyclic voltammetry, employing ferricyanide as a redox reporter molecule, was used to characterize electrode surface properties before and after disruption of the alkanethiol-Ab layer. Cyclic voltammetry was performed from 0 to 500 mV (vs. Ag/AgCl) at a scan rate of 10 mV/s. Adsorption and desorption of immobilized Abs was also verified by fluorescence labeling. In these experiments, biotinylated anti-CD4 Ab molecules were either directly attached to MUA-modified electrodes using carbodiimide coupling or were bound to streptavidin-modified electrode surfaces. Visualization of biotin–Ab deposition was done by incubating micropatterned surfaces with 10 ␮g/ml streptavidin conjugated with FITC or Alexa 546 for 30 min. The micropatterned surfaces were washed with water to remove unbound streptavidin and imaged using a confocal microscope (Zeiss LSM 5 Pascal, Carl Zeiss, Inc.). Fluorescence imaging was performed before and after electrochemical desorption in order to confirm removal of the alkanethiol-Ab layer. 2.7. Releasing Ab-bound T-cells from microelectrodes When carrying out cell release experiments, a silicone gasket was placed around a region of microelectrode-bound T-cells, thus creating a working volume of 500 ␮l of 1× PBS. Ag/AgCl reference and Pt counter electrodes were immersed in the same volume, creating an electrochemical cell with Au micropattern serving as a working electrode. Cu wires soldered onto contact pads of microelectrodes prior to T-cell capture were used to create connection to a potentiostat. Reductive potential (−1.2 V) was applied for 60 s in order to disrupt the Ab layer and release T-cells. The process of electrochemical desorption of T-cells was monitored in situ using an upright microscope (MD 130 Electronic Eyepiece, Omano). Au microelectrodes that were not electrically activated served as a negative control in order to ensure that T-cell desorption was not a sample handling (shearing) artifact. After the desorption step, solution containing T-cells was carefully aspirated and a microfabricated surface was placed into and gently agitated in a 100 mm diameter Petri dish containing 1× PBS. The cells collected from the electrode surface were re-suspended in 100 ␮l 1× PBS and stained with Trypan Blue for testing cell viability. 264 H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 Fig. 2. “Self-alignment” of hydrogel and gold microstructures using PEG photolithography. (A) and (B) SEM images demonstrating high-fidelity registration of 100 ␮m diameter gold pads with PEG gel structures. Patterning reflective gold patterns on transparent glass substrates followed by polymer coating and backside UV flood exposure obviates the need for fiduciary marks and alignment. (C) and (D) SEM images demonstrating selective attachment of 3T3 fibroblasts onto gold microelectrode surfaces. Fibroblasts were seeded everywhere on the surface containing gold and PEG gel micropatterns but were able to attach only to gold regions. Note that the white deposits on PEG gel regions are salt crystals remaining after the washing the surface with saline solution. 3. Results and discussion The present paper describes the development of a novel approach for capturing and releasing mammalian cells from microdevices. The cell release strategy was based on electrochemical disruption of an Ab layer bridging T-lymphocytes to the gold electrode surface. PEG hydrogel photolithography was employed in a novel “self-aligned” manner to effectively package Au microelectrodes within a non-fouling biomaterial (Fig. 1A). This design of the surface allowed to position Tlymphocytes exclusively on Ab-modified microelectrodes, with minimal cell adhesion occurring elsewhere. Importantly, electrical stimulation of select members of an electrode array led to the release of T-lymphocytes only from the actuated electrodes (Fig. 1B). In the future, the cell release mechanism described here may provide a convenient means of retrieving cells captured inside a microdevice for downstream gene expression studies or re-culturing. 3.1. “Self-alignment” of PEG hydrogel microstructures with gold microelectrodes Successful integration of biological systems with microfabricated devices requires methods for rendering surfaces of these devices resistant to non-specific cell and protein attachment. In this paper we present a surface modification procedure (Fig. 1A) whereby PEG-DA-containing prepolymer solution is photopatterned to fabricate hydrogel microstructures in registration with Au microelectrodes. A modification of the PEG photolithography method reported by us previously [18,19], the approach described here was ideally suited for integrating or packaging reflective metal patterns fabricated on a transparent glass substrate. Such a substrate effectively served as its own photomask, so that a flood exposure from the backside resulted in envelopment of Au microelectrodes in PEG hydrogel. As seen from Fig. 2A and B, this method led to high fidelity registration of biomaterial and metal microstructures without the need for fiduciary marks and alignment. The anchoring of the PEG gel to the glass substrate was ensured by modifying the surface with acrylated silane prior to the UV exposure step. Importantly, the silanization of exposed electrode surfaces was found to result in electrode passivation. Therefore, after fabricating Au microelectrodes, the photoresist layer was retained and used to mask the electrode interfaces during the silanization procedure. The microelectrodes packaged in PEG gel layer in this way were found to have comparable electrical properties to neat Au electrode surfaces. 3.2. Patterning of proteins and cells on gold microdomains The effectiveness of biological integration of Au microelectrodes was tested by incubating the micropatterned surfaces with H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 3T3 fibroblasts. These are anchorage dependent cells capable of adhering to various substrates by rapidly secreting their own adhesive ligands such as collagen. Therefore, interactions of fibroblasts with a substrate provide an excellent measure of the fouling properties of a material. In our experiments, microfabricated surfaces were incubated in a fibroblast suspension allowing these cells to sediment and form a contiguous monolayer spanning both Au and PEG gel regions. As seen from Fig. 2C and D, after 2 h incubation followed by washing in 1× PBS, fibroblasts were able to attach exclusively on Au electrodes with minimal cell adhesion occurring on the adjacent PEG hydrogel regions. It should be noted that white structures seen on PEG gel surfaces are salt crystals remaining from washing substrates with PBS, whereas dark objects residing on Au pads are fibroblasts. Given the “adhesiveness” of 3T3 fibroblasts, the ability to localize these cells exclusively to the electrode regions underscores the non-fouling properties of PEG hydrogel. While fibroblasts are anchorage dependent cells capable of non-specific attachment to a range of polymer and metal substrates, T-lymphocytes are circulating leukocytes that do not readily bind to surfaces. In order to promote T-cell attachment, Au electrode surfaces were decorated with Ab molecules specific to T-cell surface antigens. T-lymphocytes expressing CD4 antigen (CD4+ T-cells) are highly significant because of their importance as diagnostic markers of malignancies and infections, including HIV/AIDS [23–25]. Therefore, we wanted to demonstrate the ability to capture and release CD4+ T-cells. Self-assembly of MUA followed by activation of –COOH endgroups with EDC/NHS coupling agents were employed to covalently link neutravidin or anti-CD4 Ab molecules to the electrode surface. As seen from Fig. 3A, when microfabricated electrode surfaces with activated MUA layer were incubated with neutravidin–FITC, the fluorescence was localized only to the electrode regions. No protein deposition was observed on non-adhesive PEG hydrogel regions, once again confirming the effectiveness of our bio-packaging strategy in defining interactions between biological and microfabricated components of the microsystem. Molt-3 cells (model T-lymphocytes expressing CD-4 antigen) were employed to demonstrate cell capture on the electrodes containing anti-CD4 Abs. As described in the preceding section, microelectrodes were enveloped in a non-fouling PEG hydrogel layer to enable Ab assembly and cell binding to occur exclusively at the electrode–solution interface. In order to demonstrate our ability to modulate a number of captured T-cells, Au microdomains of dimensions varying from 20 to 40 ␮m in diameter were fabricated, packaged in PEG hydrogel and decorated with anti-CD4 Abs. As seen from Fig. 3B, single Molt-3 cells became bound on 20 ␮m diameter Au pads. When cells were incubated with substrates presenting a range of Au pad sizes, the number of captured T-lymphocytes varied from 5–6 on 40 ␮m diameter attachment pads down to single cells residing on 20 ␮m diameter attachment pads (Fig. 3C). Because this study was focused on the development and validation of a novel cell release strategy, the electrode patterns of greater area capable of capturing larger groups of T-cells were implemented for cell desorption experiments. However, as evidenced 265 Fig. 3. Designing microelectrodes for T-cell attachment. (A) Upon incubation with the micropatterned electrode substrate FITC-labeled neutravidin (50 ␮g/ml) became localized on 40 ␮m diameter gold pads. No protein deposition on PEG gel domains was observed. (B) An array of 20 ␮m diameter gold pads was modified with anti-CD4 Abs and incubated with Molt-3 cells—model T-lymphocytes expressing CD4 antigen. Individual T-cells became bound onto Ab-modified gold pads. (C) A substrate containing Ab-modified gold pads ranging in dimensions from 40 to 20 ␮m diameter (in 5 ␮m increments) was incubated with T-cells. SEM micrograph demonstrates attachment varying the number of T-cells on attachment pads of different dimensions. 20 ␮m diameter gold pads supported attachment of single cells. 266 H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 Fig. 4. (A) Cyclic voltammetry of a single member of a microelectrode array (300 ␮m diameter) performed at 10 mV/s in 5 mM ferricyanide solution. A microelectrode served as a working electrode with Ag/AgCl reference and Pt counter electrodes constituting a 3-electrode cell. Application of reductive potential (−1.2 V vs. Ag/AgCl) resulted in desorption of the alkanethiols-Ab layer and regeneration of cathodic and anodic peak typical of ferricyanide. (B) Anodic peak currents from ferricyanide cyclic voltammograms of four separate desorption experiments (n = 4). These data show reproducible regeneration of electron exchange across the electrolyte–electrode interface after reductive desorption of the alkanethiol-Ab layer. by Fig. 3B and C, it is feasible to develop microelectrode arrays for “panning” and sorting of individual T-cells. 3.3. Desorbing anti-CD4 Abs from microelectrodes Adsorption and desorption of Ab layer was characterized by cyclic voltammetry with ferricyanide serving as a redox reporter molecule. Cyclic voltammetry allowed to monitor electron exchange across the electrode–ferricyanide solution interface as a function of biomolecular assembly occurring on the electrode surface. Fig. 4A shows typical ferricyanide cyclic voltammograms collected from a single 300 ␮m diameter Au electrode. As seen from these data, bare Au electrode exhibited typical anodic and cathodic peaks associated with redox activity of ferricyanide. MUA self-assembly followed by Ab immobilization resulted in disappearance of the redox peaks, pointing to limited electron transfer across the electrode–solution interface (see Fig. 4A). A number of studies reported reductive desorption of alkanethiols to occur in the potential range of −1 to −1.5 V [16,26,27]. In the present study a potential of −1.2 V (vs. Fig. 5. Using fluorescence labeling to demonstrate detachment of Ab molecules from microelectrodes. (A) Gold microelectrodes fabricated on glass were modified with COOH-terminated thiols and biotinylated anti-CD4 Ab molecules. The surface was then incubated with neutravidin–Alexa 546 resulting in uniform fluorescence signal. Note that the PEG hydrogel packaging strategy was not employed hence appearance of fluorescence on glass regions. (B) Applying reductive potential to electrode 1 regions and washing the surface resulted in loss of the fluorescence signal from the activated-electrode regions only. Electrode 2 and glass substrates were still fluorescent pointing to the presence of Alexa–neutravidin–biotin–anti-CD4 complex. Ag/AgCl reference) was applied for 30 s to effectively remove MUA and protein layers. Applying reductive potential to Abmodified microelectrodes resulted in regeneration of anodic and cathodic redox peaks that were comparable in shape and magnitude to bare microelectrodes (Fig. 4A). Fig. 4B presents anodic peaks from ferricyanide cyclic voltammetry performed before and after reductive desorption experiments. As seen from these data, regeneration of the electron exchange across the electrolyte–electrode interface upon applying negative potential was highly repeatable. The electrochemistry experiments provided direct evidence of disruption and desorption of the passivating alkanethiol-Ab layer from the microelectrode surface. Stripping of the Ab molecules was also visualized by fluorescence microscopy. Fabrication of multiple individually addressable electrodes on the same substrate allowed to activate specific electrodes while maintaining integrity of the Ab layer H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 267 Fig. 6. Release of T-cells from Ab-modified microelectrodes. (A) Model T-cells seeded onto substrates consisting of 300 ␮m diameter electrodes enveloped in PEG hydrogel. Cells formed a monolayer and interacted with both Ab-modified gold regions as well as hydrogel regions. (B) After washing, only T-cell exposed to Ab-gold regions were retained. No cells remain on PEG hydrogel domains. (C) Application of negative potential (−1.2 V vs. Ag/AgCl reference) for 60 s followed gentle agitation of the surface resulted in release of T-cells. (D) An array of three individually addressable (300 ␮m diameter) electrodes presented in the same field of view (×40 magnification). Initially, T-lymphocytes were captured on all three electrodes, however, upper and lower electrodes were activated by applying reductive potential. The electrode in the center was not electrically activated. As seen from this image, the central electrode contains surface-bound T-cells while the activated electrodes are void of cells. Trypan blue staining of T-lymphocytes released from the electrode surfaces revealed that ∼90% of cells were viable. on other conductive regions. Fig. 5A demonstrates immobilization of biotinylated anti-CD4 Abs on the substrate containing two individually addressable electrode regions. The presence of Ab–biotin complex was verified by incubating the surface with neutravidin–Alexa 546 complex. It should be noted that PEG packaging was not utilized in this case, hence the uniform distribution of fluorescence on both Au and glass regions. As seen from Fig. 5B, actuating of electrode 1 resulted in disappearance of fluorescence signal only from the actuated electrode region (dark region) but not from the other electrode or the glass substrate. Overall, electrochemistry and fluorescence microscopy results provide conclusive evidence pointing to desorption of the Ab layer from the gold electrode surface. 3.4. Releasing Ab-bound T-lymphocytes from microelectrodes Leukocytes are white blood cells that provide important diagnostic information about malignancies, infections and autoimmune disorders [28]. T-lymphocytes are responsible for orchestrating adaptive immune response and, therefore, represent a particularly important subset of leukocytes. For example, CD4+ T-cells are preferentially destroyed by HIV/AIDS, resulting in compromised immunity and greatly increasing patient’s susceptibility to opportunistic infections [23,24]. The diagnos- tic importance of leukocytes in general and CD4+ T-cells in particular provides the impetus for developing new miniature devices capable of rapid capture and characterization of these cells. These cytometry microdevices most commonly employ a “panning” approach whereby the cells are captured onto Abdecorated regions within the device, and then immunolabeled and/or enumerated [21,29–32]. Therefore, a biological goal of the present study was to develop an approach for releasing T-cells captured inside a microdevice. In the cell release experiments described schematically in Fig. 1B, model T-lymphocytes were seeded onto a five member array of 300 ␮m diameter Au electrodes connected to contact pads by 10 ␮m leads. This electrode array was imbedded into a non-fouling PEG gel layer using “self-alignment” process described above. Fig. 6A demonstrates that introduction of cells onto a substrate resulted in formation of a dense, contiguous monolayer of T-cells interacting with both Au and PEG hydrogel microstructures. However, as demonstrated in Fig. 6B, gentle aspiration of the cell seeding medium and agitation of the substrate in 1× PBS revealed that T-cells attached exclusively on Ab-modified electrode regions with minimal cell attachment on PEG hydrogel. Importantly, T-cell attachment to Au leads connecting electrodes to contact pad was eliminated by adjusting UV exposure so as to over-polymerize PEG gel around 10 ␮m wide Au line. 268 H. Zhu et al. / Colloids and Surfaces B: Biointerfaces 64 (2008) 260–268 After cell capture, a microfabricated substrate was transferred into an electrochemical cell and immersed in ∼1 ml of 1× PBS serving as electrolyte solution. Pt counter and Ag/AgCl reference electrodes were introduced into the electrochemical cell with T-lymphocyte capturing Au surface serving as a working electrode. Applying reductive potential (−1.2 V vs. Ag/AgCl for 60 s), followed by gentle agitation of the electrochemical cell resulted in desorption of T-lymphocytes from the electrode (Fig. 6C). Fig. 6D shows an array of three gold microelectrodes employed to capture model T-cells. Upper and lower electrodes were electrically activated while a central electrode was not. As seen from this image, the electrode that was not activated contains captured T-cells, whereas activated electrodes are void of cells. These data provide additional evidence that T-cell release was due to electrode activation and was not a cell handling (e.g. shearing off) artifact. In the future, the cell sorting mechanism demonstrated here will be incorporated into a microfabricated cytometry platform for capture and interrogation of cells to enable release of specific cells based on phenotype and function determination made in the microdevice. 4. Conclusions In this paper, we describe a novel approach of capturing Tcells onto Ab-modified electrode arrays and then releasing cells by electrochemical desorption of the underlying Ab layer. Selective placement of T-cells on Ab-modified electrodes was ensured by “self-alignment” of non-fouling PEG hydrogel microstructures with the electrode regions. The detachment of anti-CD4 Ab molecules linked to gold electrodes via alkanethiols was demonstrated by electrochemistry as well as fluorescence microscopy. Importantly, model T-lymphocytes attaching to an array of Ab-decorated electrodes were released using the same strategy of disrupting the underlying cell-adhesion layer. In contrast to the cell release strategy reported by Mrksich and co-workers [11,12], the method described here is considerably simpler as it employs standard alkanethiol/protein conjugation procedures and obviates the need for synthesis of specialized thiols. The use of microelectrodes fabricated on an insulating glass substrate ensures that individual regions of the substrate may be activated and may release cells independently of each other. The microfabricated substrate capable of capturing antigen-specific cells and releasing them upon interrogation may offer a valuable tool for immunology, cancer research and other application where diagnosis is frequently based on the emergence of malignant cells or changes in the numbers/function of normal cells. When integrated with a microfluidic delivery system, this electroactive substrate will enable selective capture of T-lymphocytes or other cell types and subsequent release of the cells of interest for downstream analysis and re-culturing. Acknowledgements We thank Prof. Louie’s lab in the Department of Biomedical Engineering at UC Davis for providing assistance with confocal microscopy. This publication was supported in part by the California Research Center for the Biology of HIV in Minorities, California HIV/AIDS Research Program #CH05D-606. HZ was supported through an NIH Training Grant (EB003827). References [1] C. Pale-Grosdenmange, E.S. Simon, K.L. Prime, G.M. Whitesides, J. Am. Chem. Soc. 113 (1991) 12–20. [2] K.L. Prime, G.M. Whitesides, J. Am. Chem. Soc. 115 (1993) 10714–10721. [3] K. Prime, G. Whitesides, Science 252 (1991) 1164–1167. [4] E. Ostuni, R.G. Chapman, M.N. Liang, G. Meluleni, G. Pier, D.E. Ingber, G.M. Whitesides, Langmuir 17 (2001) 6336–6343. [5] Y. Xia, G.M. Whitesides, Angew. Chem. Int. Ed. 37 (1998) 550–575. [6] S.K. Bhatia, J.L. Teixeira, M. Anderson, L.C. Shriverlake, J.M. Calvert, J.H. Georger, J.J. Hickman, C.S. Dulcey, P.E. Schoen, F.S. Ligler, Anal. Biochem. 208 (1) (1993) 197–205. [7] C.S. Dulcey, J.H. Georger, V. Krauthamer, D.A. Stenger, T.L. Fare, J.M. Calvert, Science 252 (1991) 551–554. [8] P.E. Laibinis, J.J. Hickman, M.S. Wrighton, G.M. Whitesides, Science 245 (1989) 845–847. [9] M. Veiseh, M.Q. Zhang, J. Am. Chem. Soc. 128 (4) (2006) 1197– 1203. [10] M. Veiseh, M.H. Zareie, M.Q. Zhang, Langmuir 18 (17) (2002) 6671–6678. [11] W.S. Yeo, M. Mrksich, Langmuir 22 (25) (2006) 10816–10820. [12] W.S. Yeo, C.D. Hodneland, M. Mrksich, ChemBioChem 2 (2001) 590. [13] M.N. Yousaf, B.T. Houseman, M. Mrksich, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 5992–5996. [14] J. Lahann, S. Mitragotri, T.N. Tran, H. Kaido, J. Sundaram, I.S. Choi, S. Hoffer, G.A. Somorjai, R. Langer, Science 299 (2003) 371–374. [15] C.A. Widrig, C. Chung, M.D. Porter, J. Electroanal. Chem. 310 (1–2) (1991) 335. [16] X. Jiang, R. Ferrigno, M. Mrksich, G.M. Whitesides, J. Am. Chem. Soc. 125 (2003) 2366–2367. [17] Y. Li, B. Yuan, H. Ji, D. Han, S.Q. Chen, F. Tian, X.Y. Jiang, Angew. Chem. 46 (7) (2007) 1094–1096. [18] A. Revzin, R.G. Tompkins, M. Toner, Langmuir 19 (2003) 9855–9862. [19] A. Revzin, R.J. Russell, V.K. Yadavalli, W.-G. Koh, C. Deister, D.D. Hile, M.B. Mellott, M.V. Pishko, Langmuir 17 (2001) 5440–5447. [20] A. Revzin, P. Rajagopalan, A.W. Tilles, F. Berthiaume, M.L. Yarmush, M. Toner, Langmuir 20 (2004) 2999–3005. [21] A. Revzin, K. Sekine, A. Sin, R.G. Tompkins, M. Toner, Lab Chip 5 (2005) 30–37. [22] S.G. Balasubramanian, A. Revzin, A.L. Simonian, Electroanalysis 18 (2006) 1885. [23] R.F. Siliciano, T. Lawton, C. Knall, R.W. Karr, P. Berman, T. Gergory, E.L. Reinherz, Cell 54 (1988) 561–575. [24] R.S. Veazey, M. DeMaria, L.V. Chalifoux, D.E. Shvetz, D.R. Pauley, H.L. Knight, M. Rosenzweig, R.P. Johnson, R.C. Desrosiers, A.A. Lackner, Science 280 (1998) 427–431. [25] J. Guarner, P. Montoya, C. del Rio, G. Hernandez-Tepichin, Cytometry 30 (1997) 178–180. [26] T. Sumi, H. Wano, K. Uosaki, J. Electroanal. Chem. 550 (2003) 321–325. [27] V.M. Mirsky, Trends Anal. Chem. 21 (2002) 439–450. [28] M.L. Turgeon, Clinical Hematology, 3rd ed., Lippincott Williams & Wilkins, 1999. [29] X.H. Cheng, D. Irimia, M. Dixon, K. Sekine, U. Demirci, L. Zamir, R.G. Tompkins, W. Rodriguez, M. Toner, Lab Chip 7 (2) (2007) 170– 178. [30] S.K. Murthy, A. Sin, R.G. Tompkins, M. Toner, Langmuir 20 (2004) 11649–11655. [31] K. Sekine, A. Revzin, R.G. Tompkins, M. Toner, J. Immunol. Methods 313 (2006) 96–109. [32] A. Sin, S.K. Murthy, A. Revzin, R.G. Tompkins, M. Toner, Biotechnol. Bioeng. 91 (2005) 816–826.