A duplexed microsphere-based cellular adhesion assay

ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 337 (2005) 246–255 www.elsevier.com/locate/yabio

A duplexed microsphere-based cellular adhesion assayq Wendy Lee Connorsa, Jyrki Heinob,* a

Department of Medical Biochemistry and MediCity Research Laboratory, University of Turku, FI-20014 Turku, Finland b Department of Biochemistry and Food Science, University of Turku, FI-20014 Turku, Finland Received 14 August 2004 Available online 7 December 2004

Abstract Interactions of integrin cellular adhesion molecules with matrix proteins play important roles in complex bidirectional signaling pathways. To investigate these interactions, a novel flow-cytometry-based cellular adhesion assay has been developed. Based on the concept of microcarrier cell culture, derivatized polystyrene microspheres (9.6 lm) are used as a substrate for the immobilization of type I collagen to which cells then adhere. Using cytometric detection, the extent of cellular adhesion can be precisely determined by comparison of adhered and nonadhered populations based on the side scatter properties of the microspheres. In combination with immunostaining, the novel format of this assay enables the correlation of adhesive function to other cellular characteristics such as surface expression. In this work, the protein kinase C activator 12-O-tetradecanoylphorbol-13-acetate (TPA) was used to stimulate increased adhesion in Chinese hamster ovary cells stably transfected with the collagen receptor integrin a2b1. Multiple clones of varying expression distributions were analyzed, and correlations of adherent populations versus receptor distributions show a threefold increase in functional cellular adhesion to collagen upon treatment with TPA. Probability binning analysis of duplexed data revealed subtle changes in adhesion versus receptor distribution mediated by TPA which otherwise would not have been detectable.
2004 Elsevier Inc. All rights reserved. Keywords: Adhesion assay; Integrin; Collagen; Cytometry

Integrins are a family of heterodimeric, membranespanning glycoproteins which function as both signaling and adhesion molecules, participating in cell–matrix and cell–cell interactions [1,2]. They consist of noncovalently linked a and b subunits, which each comprise extracellular, transmembrane, and cytoplasmic domains. Integrinmediated cellular adhesion is subject to multiple modes of regulation including outside-in activation via contact with matrix proteins and inside-out stimulation through complex cellular signaling pathways [3]. For this reason, the simultaneous determination of functional markers such as cell surface expression or activation of signaling q This work was financially supported by the Academy of Finland, the Finnish Cancer Association and the Sigrid Juse´lius Foundation. * Corresponding author. Fax: +358 0 2 333 6860. E-mail address: jyrki.heino@utu.fi (J. Heino).

0003-2697/$ - see front matter
2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2004.10.044

pathways with cellular adhesion is an important capability. Conventional cellular adhesion studies use well plate methods in which a layer of matrix molecules is applied to the bottom of a well, cells are pretreated in suspension and added to the wells, and nonadherent cells are removed by washing. Adherent cells are then quantified, at times by manual counting but more commonly by radiolabeling [4,5], staining with a colorimetric [6] or fluorescent [7] reagent, or by enzymatic techniques in which a substrate is cleaved, resulting in a signal proportional to the number of cells present [8,9]. The signal is then detected by an automated plate reader. These methods can work quite well; however, they lack the essential capability for simultaneous determination of multiple variables. Thus, they yield very limited information.

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

The use of derivatized microcarrier supports as the basis for automated biological assays has been rapidly expanding for several years [10–12]. However, cellular adhesion assays have not, until now, made use of their potential. To investigate integrin-mediated cellular adhesion, a flow-cytometry-based adhesion assay in which type I collagen is adsorbed to carboxylated polystyrene microspheres (9.6 lm) has been devised. Plated cells interact with the collagen-coated beads,

247

and the extent of cellular adhesion can be precisely determined by comparison of adhered and nonadhered populations based on the side scatter properties of the microspheres (Fig. 1). Since no fluorescent labeling is required for the fundamental adhesion assay, subsequent immunostaining enables assessment of adhesion as a function of surface antigen expression or assessment of biochemical properties as a function of adhesion.

Fig. 1. Microsphere-based cellular adhesion assay. (A) Flow cytometry analysis of CHO a2 cells capable of forming functional adhesions to collagen-coated beads. Cellular adhesion is determined by comparison of adhered (upper right quadrant) and nonadhered (lower right quadrant) populations based on the forward and side scatter properties of cells and beads. Nonadhered beads appear in the upper left quadrant. (B) Nonspecific interaction is assessed using BSA-coated beads as a negative control. (C) A duplexed adhesion/expression assay was devised to determine the relative distributions of a2 integrins on adherent cells within a heterogeneous population. The bead-based adhesion assay was performed as described in the text, followed by fixation with formalin and staining with mAb 12F1 and a FITC-labeled secondary antibody. Using the gating functions of cytometry software, the fluorescence signal representing receptor distributions for adherent and total cell populations can be compared. (D) A CHO cell transfected with the a2 integrin subunit adheres to a type I collagen-coated bead.

248

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

Materials and methods Generation of cell lines Preparation of the construct for expression of the a2 integrin subunit and generation of Saos-2 a2b1 cells has been previously described [13]. Chinese hamster ovary (CHO)1 cells were obtained from the American Type Culture Collection, and transfection was performed as described [14]. The cells were maintained in DulbeccoÕs modified EagleÕs medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum (FCS; GibcoBRL), 2 mM glutamine, 100 IU mL
1 penicillin G, and 100 lg mL
1 streptomycin. Briefly, cells were transfected by electroporation (0.3 kV, 960 microfarad, 0.4cm cuvette in RPMI plus 1 mM sodium pyruvate, 2 mM L -glutamine, without serum) using 20 lg of expression plasmid consisting of cDNA in pAW [15]. Transfected cells were plated and allowed to recover for 1 day in culture medium. Neomycin analog G418 (Life Technologies) was added to the medium at a concentration of 1 mg mL
1. Cell surface expression levels of the integrin a2 subunit were verified using anti-integrin mAb 12F1 (BD Biosciences) and flow cytometry. Receptor staining Cells were grown to early confluence and released by trypsinization. Trypsin was inhibited by addition of DMEM containing 10% FCS, cells were centrifuged, washed once in PBS, pH 7.4, containing 1% FCS, and resuspended in the same buffer to 2 · 106 cells mL
1. Anti-a2 integrin mAb 12F1 was added to a final concentration of 2.5 lg mL
1, and cells were incubated for 30 min at 4 C with rotation. Cells were stained with a FITC-conjugated rabbit anti-mouse IgG secondary antibody (Dako; 1:20) for 30 min at 4 C, then washed twice in PBS, pH 7.4, and assayed by flow cytometry (FACScan; BD, Franklin Lakes, NJ, USA) Generation of derivatized bead supports Carboxylated polystyrene microspheres 9.6 lm in diameter (Polymer Laboratories, Church Stretton, UK) were washed 2· in 0.01 M HCl, resuspended in Cellon bovine dermal collagen (Cellon S.A., Luxembourg) at 3 mg mL
1, pH 2, and incubated with rotation at 4 C for 48 h. Beads were washed 2· in serum-free DulbeccoÕs modified EagleÕs medium (SF-DMEM), 1 Abbreviations used: CHO, Chinese hamster ovary; DMEM, DulbeccoÕs modified EagelÕs medium; FCS, fetal calf serum; PBS, phosphate-buffered saline; BSA, bovine serum albumin; RT, room temperature; TPA, 12-O-tetradecanoylphorbol-13-acetate; FITC, fluorescein isothiocyanate; PKC, protein kinase C; PBA, probability binning analysis.

resuspended in PBS + 1% bovine serum albumin (BSA; Sigma), and incubated for 24 h with rotation at 4 C. Finally, beads were washed 2· and resuspended in SF-DMEM. Bead concentration was adjusted to approximately 1.6 · 108 beads mL
1. BSA-coated negative control beads were produced similarly, except that beads were incubated in PBS + 1% BSA, pH 7.4, with rotation at RT for 2 h. For antibody-coated beads (16B4, Serotec; rat IgG, Sigma), bead suspension was washed 2· in PBS, and supernatant was removed. Antibody was added to a final concentration of 200 lg Ab/ mL 10· bead suspensions, and final volume was adjusted to 2 mL with PBS. Beads were incubated with rotation at 4 C for 48 h, washed 2· in PBS, resuspended in PBS + 1% BSA, and incubated for 1 h with rotation at RT. Finally, beads were washed 2· and resuspended to 7.5% v/v in PBS + 0.05% BSA. All beads were stored at 4 C and used within 30 days of coating. Microsphere-based cellular adhesion assay Stably transfected CHO cells were grown to early confluence and released by trypsinization. Trypsin was inhibited by addition of DMEM + 10% FCS, cells were centrifuged, washed once in DMEM + 10% FCS, plated into 96-well plates at 2 · 105 cells/well, and allowed to attach for 1 h at 37 C. Cells were washed once with SF-DMEM and medium was replaced with 45 lL SF-DMEM per well. Cells were then treated according to a general protocol; primary treatment followed by addition of collagen-coated beads (5 lL of 1.6 · 108 beads mL
1) with or without 12-O-tetradecanoylphorbol-13-acetate (TPA; 100 nM except as where noted; Calbiochem), incubation, trypsinization, fixation, and analysis. The following primary treatment protocols were used. (1) For LY294002 experiments (Calbiochem), inhibitor was added to each well in 5-lL aliquots of 10· concentration to a final volume of 50 lL, final concentration as listed, and cells were incubated at 37 C for 30 min before addition of beads. (2) For inhibitory antibody experiments using P1H5 (Santa Cruz) or A2-IIE10 (a kind gift from Dr. F. Berditchevski) [16], primary treatment was the same except that incubation was 15 min at 37 C. (3) For TPA concentration series experiments, no primary treatment was used. After incubation with beads, overlying medium was removed, cells were released by trypsinization (25 lL of 0.5% v/v trypsin well
1), trypsin was inhibited by addition of 50 lL DMEM + 10% FCS, cells were fixed (20 lL of 10% v/ v formalin well
1), and cytometry was performed. Neither trypsinization nor fixation affected the cellular adhesions to collagen-coated beads consistent with the insensitivity of integrins to trypsin as previously shown [17]. The extent of cellular adhesion was determined by comparison of adhered and nonadhered populations based on the side and forward scatter properties of the

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

microspheres and cells, respectively. Each data point represents 3–10 replicates and is reported as means ± SD. CellQuest, WinMDI, Microsoft Excel, and SPSS software were used for analysis of cytometry data. Plate reader cellular adhesion assay The 96-well plates were coated with Cellon (50% v/v in PBS; 25 lL per well) or BSA (1% in PBS; 25 lL per well) by overnight incubation at 37 C followed by 2· wash with PBS, pH 7.4. Stably transfected CHO cells were grown to early confluence and released by trypsinization. Trypsin was inhibited by addition of DMEM + 10% FCS, cells were centrifuged, washed once and resuspended to 4 · 105 cells mL
1 in SFDMEM. Aliquots of cell suspension were added to Eppendorf tubes containing inhibitory antibody, and cells were incubated at 4 C with rotation for 30 min. Cells were then plated into precoated 96-well plates at 2 · 104 cells well
1 and allowed to attach for 30 min at 37 C. Unbound cells were removed by washing 2· with SF-DMEM and medium was replaced with 100 lL SFDMEM. BSA-coated wells were used as a negative control. The number of attached cells was determined using the CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (Promega) according to manufacturerÕs instructions. The absorbance at 492 nm was recorded using an automated plate reader. Reported values are the average of four replicates ±SD. Duplexed cellular adhesion/expression assay The microsphere-based adhesion assay was performed as above. After incubation with microspheres, detached fixed cells were washed in 0.5 mL PBS + 1% FCS and resuspended in 100 lL PBS + 1% FCS with or without anti-a2 integrin mAb (12F1) to a final concentration of 2.5 lg mL
1. Samples were incubated for 30 min at 4 C with rotation. FITC-conjugated rabbit anti-mouse IgG secondary antibody (5 lL) was added to each sample, followed by incubation for a further 30 min. Samples were washed twice, resuspended in 200 lL PBS, and analyzed by flow cytometry.

Results and discussion A novel cytometry-based cellular adhesion assay gives rapid and reliable functional analysis of collagen receptors To investigate integrin-mediated cell–matrix interactions, a flow-cytometry-based assay specific for adhesion has been developed. Carboxylated polystyrene microspheres are used as a substrate for the adsorption of type I collagen, to which cells then adhere; the microspheres

249

are incubated with collagen for P48 h. Cantarero et al. [18] have shown that there is only a small advantage to incubating proteins with polystyrene beyond 16 h and, once adsorbed, very little protein is lost throughout multiple enzyme-linked immunosorbent assay steps, suggesting that collagen-coated polystyrene beads will be highly stable. Also, Lee et al. [11] have shown that the amount of collagen immobilized this way onto polystyrene beads is somewhat dependent on the initial concentration of collagen. However, by using a commercially available preparation, the concentration of collagen in our immobilization procedure is constant. Microspheres are coated in large batches to further minimize interexperimental variation. The detection of forward and side scatter characteristics of particles is a fundamental function of flow cytometry. Forward scatter measures diffracted light and can be used to estimate the size of a particle, while side (or orthogonal) scatter measures refracted and reflected light resulting from differences in refractive index (e.g., granularity of cells or opacity of beads) [19]. In a typical cytometric analysis of stained cells, these channels are used to specify a population of interest. In the current application we have made use of the high side scatter and low forward scatter of polystyrene beads relative to viable cells. These properties provide the basis for sensitive and precise determination of adhered populations (Fig. 1). Thus, no fluorescent labeling is required for the fundamental adhesion assay, allowing for the possibility of multiplexed experiments assessing adhesion as a function of other cellular properties such as surface expression. Polystyrene beads were either left uncoated or coated with BSA, rat IgG, type I collagen, or mAb 16B4, specific for the a2 integrin subunit (Fig. 2A). Human osteosarcoma cells transfected with the integrin a2 subunit (Saos a2) showed very low levels of interaction with BSA- and rat IgG-coated beads. Slightly higher interaction occurred with uncoated beads due to the hydrophobic nature of the polymer. However, cells showed high levels of binding to both the collagen- and the specific antibody-coated beads. CHO cells have no functional receptors for type I collagen [20]. Here, they were transfected with either the integrin a2 subunit (CHO a2) or a control vector (CHO pAW) and were exposed to uncoated, BSAcoated, or type I collagen-coated beads (Fig. 2B). Both cell types showed consistently low binding levels to negative control beads, and CHO pAW cells also showed negative control levels of interaction with type I collagen-coated beads. However, CHO a2 cells interacted strongly with type I collagen-coated beads. Furthermore, pretreatment of CHO a2 cells with the anti-a2 integrin function-blocking antibody P1H5 resulted in a return to background levels of binding to collagencoated beads (Fig. 2C), demonstrating that the cell–bead

250

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

accounts for its function-blocking ability [21]. Meanwhile, an anti-a2 mAb with an epitope outside the collagen binding domain, 16B4, is not function blocking [22]. Antibody P1H5 was not able to significantly decrease CHO a2 binding to beads coated with mAb 16B4, consistent with different epitopes recognized by the two antibodies. Evaluation of intracellular and extracellular inhibition of adhesion

Fig. 2. Specific interactions with protein-coated microspheres. (A) Saos-2 cells expressing the a2 integrin subunit were incubated for 3 h with uncoated beads, negative control beads (BSA or rat IgG), or beads coated with a2 integrin-specific proteins (type I collagen or mAb 16B4) and then trypsinized and assayed by flow cytometry (n = 3). (B) Interaction of CHO pAW negative control cells or CHO a2 cells with uncoated or protein coated beads (n = 3). Pretreatment (15 min) with the function-blocking a2-integrin-specific mAb P1H5 inhibits adhesion of CHO a2 cells to collagen-coated beads. (C) Polymer beads were coated with either type I collagen, the a2-integrin-specific mAb 16B4, or BSA (negative control). Untreated CHO a2 cells could adhere to beads coated with either collagen or 16B4. Pretreatment with the a2integrin-specific function-blocking antibody could inhibit adhesion to collagen-coated beads but not beads coated with 16B4, consistent with the differential specificity of these antibodies (n = 3).

interaction is mediated by integrin a2b1. The epitope of mAb P1H5 is located adjacent to the von Willebrand factor-like domain within the a2 I domain which

As a qualitative comparison and verification of results, side-by-side dose–response curves were constructed for anti-a2 integrin mAbs P1H5 and A2IIE10, using both the novel adhesion assay and a 96-well automated plate reader format with the colorimetric reagent MTS (Figs. 3A–D). These antibodies each inhibit a2b1 integrin-mediated adhesion to type I collagen, however, their characteristics with regard to effective concentrations and maximum inhibition are notably different. Onset of inhibition by mAb P1H5 occurs below 0.1 lg mL
1, and the maximum inhibition reached is approximately 77%. Meanwhile, mAb A2-IIE10 shows inhibitory effects only at concentrations above 1 lg mL
1; however it produces a maximum inhibition of greater than 90%. Curves obtained by each of the two alternative methods correlate quite well (r2 > 0.98) notwithstanding the relatively large average percentage deviations in the plate reader data (10.6%) versus that obtained by the bead-based adhesion assay (2%; calculated as SD · 100/mean, n = 3). These results demonstrate the utility of this assay in detecting variable inhibitory capabilities. Phosphatidylinositol 3-kinase (PI 3-kinase) activity is closely linked to integrin function [23–25]. However, the role of PI 3-kinase in adhesion to type I collagen through the a2b1 integrin is not clear. To evaluate the utility of the present method with intracellular inhibitors and its reproducibility, CHO a2b1 cells were treated with the PI 3-kinase blockade LY 294002 which effectively inhibited adhesion in a dose-dependent manner (Fig. 3E). Replicate experiments were performed with cells from nonconsecutive passages several days to several weeks apart, demonstrating the excellent reproducibility of the assay. Each data set was independently fit by iteration to the dose–response equation, R¼


Rmax 1 þ ðC=EC50 Þ

b

þ Rmax ;

ð1Þ

where concentration C produces response R for a given effector, Rmax is the maximum attainable response, EC50 is the concentration eliciting half-maximal response, and b is a constant related to the slope of the curve [26]. The correlation coefficients for all data sets were greater than 0.99, and the EC50 value for inhibition of adhesion was determined to be 16.1 ± 3.5 lM.

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

251

Fig. 3. Dose–response curves and correlation plots showing inhibition of adhesion by function-blocking anti-a2 integrin mAbs P1H5 (A) or A2IIE10 (C). CHO a2 cells were incubated with antibodies for 15 min and evaluated by either the bead-based adhesion assay detected by flow cytometry (j) or a well plate assay (s) in which a colorimetric enzymatic reaction (MTS) proportional to the plated cell population is detected by an automated UV/vis plate reader. Data are averages of two triplicate experiments ± SD. Correlation plots (B and D) show good agreement between the two methods. Correlation coefficients were determined by linear regression analysis. (E) The PI 3-kinase inhibitor LY294002 inhibits integrin-mediated adhesion of CHO a2 cells in a dose-dependent manner. Plated CHO a2 cells were pretreated with LY294002 for 15 min at 37 C or left untreated before addition of beads, followed by trypsinization and cytometric analysis. Four triplicate experiments are shown ±SD. Averaged data were fit to the dose–response relationship (r2 = 0.9977). The EC50 was determined to be 16.1 ± 3.5 lM.

Evaluation of inside-out signaling-mediated increases in adhesion The utility of this system to measure effects of an agonist to adhesion was evaluated using the potent PKC activator and tumor-promoting agent, 12-O-tetradecanoylphorbol-acetate. Although chronic treatment with TPA can downregulate PKC expression [27], PKC activation mediated by TPA has been shown to play roles in both short-term modulation of integrin activity by altering receptor avidity [28] and long-term adhesive behavior through increases in receptor expression. Rosfjord

et al. [27] have shown that treatment with TPA causes an increase in the collagen binding avidity of a2b1 integrin within 1–4 h. CHO a2b1 cells were treated with a series of concentrations of TPA for durations of 45 min or 4 h. The shorter treatment resulted in a characteristic dose–response curve (Fig. 4A) which indicates a maximal effective concentration (ECmax) of 5 lg mL
1 and an EC50 of 0.4 lg mL
1. At concentrations higher than 5 lg mL
1, adhesion begins to decrease. After 4 h of TPA treatment, cellular adhesion is at basal levels (i.e., similar to untreated samples) for concentrations at or below

252

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

1 lg mL
1. This is consistent with the fact that prolonged exposure to TPA downregulates PKC expression [29]. Hence, PKC-dependent activation by TPA cannot

take place. At higher concentrations, a downward trend similar to that occurring after the 45-min treatment is apparent. The ability of the a2-integrin-specific function-blocking mAb P1H5 to substantially reduce both native and TPA-mediated adhesion indicates that this effect takes place through the a2b1 integrin (Fig. 4B). Because of the flexible nature of our adhesion assay, we were able to observe the time course of avidity modulation by TPA. Treatment of CHO a2b1 cells with TPA induced statistically significant increases in adhesive behavior within 10 min of addition (Fig. 4C). Untreated cells showed near linear increases in adhered populations for up to 50 min, subsequently leveling off. However, in TPA-treated cells, adhesion increased much more rapidly at time points below 40 min, leveling off somewhat sooner but attaining a final adhesion level almost 70% higher than that of untreated cells. These results verify that the functional response to TPA-mediated avidity modulation occurs in less than 10 min. Duplexed cellular adhesion/receptor expression assay

Fig. 4. (A) Time-dependent stimulation of adhesion by TPA. CHO a2 cells were incubated with TPA concentrations and type I collagencoated beads for either 45 min (j) or 4 h (s), trypsinized, and assayed by flow cytometry. Data are normalized to untreated control samples within each experiment and consist of averages of two triplicate experiments ± SD. (B) Plated CHO a2 cells were either pretreated with mAb P1H5 (1 lg mL
1, 15 min at 37 C) or left untreated before addition of beads with or without TPA (1 lg mL
1, 45 min at 37 C), followed by trypsinization and cytometric analysis. Data are averages of three triplicate experiments ± SD. (C) Temporal profiles of normal and TPA-mediated adhesion for CHO a2 cells. Cells were added to 96well plates in DMEM + 10% FCS and allowed to attach for 30 min. Cells were then washed once, medium was replaced with SF-DMEM, and collagen-coated beads with TPA (1 lg mL
1; s) or without TPA (j) were added to triplicate wells at time intervals. All wells within an experiment were trypsinized simultaneously and analyzed by flow cytometry. Data are averages of two triplicate experiments ± SD.

The correlation of receptor expression and functional adhesion may be used as an objective determinant of functional avidity. However, immunostainings often reveal heterogeneous integrin expression, even within ‘‘clonal’’ populations, often including significant sub populations of low- or nonexpressing cells. In a typical assay, each subject is assumed to have the same response probability so that the presence of nonreactive subpopulations confound attempts at correlation. Only by selecting populations of specific expression levels can the relative avidity of a population be determined. To characterize differences in avidity between and within cell populations, a duplexed microsphere-based adhesion/expression method using an a2b1-specific antibody (12F1) was formulated (see Fig. 1C). By this method, adherent, nonadherent, and total cell populations can be selectively analyzed for their relative expression distributions. Also, difficulties in comparing adhesive function to surface expression in heterogeneous cell populations are ameliorated. TPA-induced avidity changes were evaluated by comparison of four clones of varying expression. Probability binning analysis (PBA) [30] was applied to fluorescence histograms of adherent cell and total cell populations (Fig. 5A). Our analysis revealed that with TPA treatment, lower-expressing cells make up a larger percentage of the total adherent population than untreated cells (Fig. 5B), indicating an increased capability of low-expressing cells to form functional adhesions. Furthermore, we found that TPA treatment resulted in a statistically significant (p < 0.01, PBA) downward shift in the fluorescence profile of total cell populations (Fig. 5C) as evidenced by the relatively higher percentages of population occurring in bins 1–3. This may be

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

253

Fig. 5. Comparison of receptor expression for total and adherent populations. (A) Integrin a2 expression profiles for untreated and TPA-treated cells with a schematic representation of probability binning analysis (PBA). Markers are used to partition fluorescence histogram data from a control sample into ‘‘bins,’’ each containing an equivalent percentage of the data. Here, 10 bins were constructed, each containing 10% of the fluorescent histogram data. These same ‘‘bins’’ are then applied to test data, and comparison of the percentage of data in each reveals shifts in the distribution relative to the control. In this case, an untreated total cell population is the control (thin line), and a TPA-treated total cell population is the test population (boldfaced line). (B) PBA comparison of the fluorescence distribution of total versus adherent populations for untreated and TPA-treated cells. Bins were constructed using total populations as controls (dashed line) and applied to adherent cells as the test populations. Four CHO a2 clones, both untreated (gray line) and TPA-treated (black line), were analyzed independently. Averaged data are presented. With TPA treatment, a higher percentage of adherent cells occur in low-expression partitions (bins 2–6) than untreated adherent cells. (C) PBA comparison of receptor expression profiles were performed independently for each of four CHO a2 clones without TPA (control; dashed line) or with TPA (tests; solid lines). A TPA-mediated shift to lower expression levels is evident from the higher percentages of events occurring in bins 1–3 for all clones. (D) Adhesion/ receptor expression correlation plots compare normal adhesion (j) to TPA-mediated adhesion (s) for CHO a2 cells. Four separate clones were analyzed using the duplexed adhesion/expression assay in which cells were incubated with collagen-coated beads with or without 1 lg mL
1 TPA for 45 min at 37 C followed by staining as described.

accounted for by increased PKC-mediated internalization of receptors [31]. Thus, selective analysis on the basis of adhesion shows that, even with fewer collagen receptors, TPA-treated cells formed more functional adhesions than untreated cells. When percentage of adherent cells was plotted versus the geometric mean of fluorescence value for the isolated adhered population (Fig. 5D), they were found to be strongly correlated. Here, the slopes of the resultant lines provide a relative index for the assessment of collagen binding function. After treatment with TPA, an increase in slope of approximately threefold was accompanied by the downward trend in the geometric mean fluorescence noted above (p = 0.031, StudentÕs

paired t test). These data may reflect a fundamental change in the affinity state of the collagen receptor due to TPA treatment. The novel assay presented here, founded on the concept of microcarrier cell culture and utilizing cytometric detection, has enabled improved sensitivity and reproducibility in the collection and evaluation of cellular adhesion data, thus facilitating the study of receptor–ligand interaction. Using the interaction between cell surface integrins and collagen-coated microspheres, the extent of cellular adhesion can be precisely determined. Although simple in concept, the format of this assay presents several advantages over previously used methods. Typical adhesion assays require multiple time-con-

254

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

suming intermediate steps for agonist/antagonist exposure, for harvest, resuspension, and incubation of the cells, and finally for addition to a substrate and attachment of cells. Many protocols call for several hours of exposure to chemical effectors while cells are maintained in suspension, thus calling into question the viability of cells in subsequent steps. Here, elimination of these steps makes accurate measurements of adhesion agonism possible within 10 min of stimulant addition. Specificity is high since cells will interact with collagen-coated beads only through collagen receptors, and the sensitivity of cytometric detection facilitates assessment of very small differences in adhesive behavior in the samples. Furthermore, the results obtained by this assay are not relative as in well plate assays but give the actual percentage of the population that is responding, since nonadherent cells within each well act as an internal standard. Interestingly, the use of cytometry also enables a simultaneous indication of cytotoxicity based on the cellular forward scatter profile. Most importantly, though, this assay enables correlation of ligand binding function to cellular properties such as receptor expression. In combination with immunostaining, this method has permitted us to, very specifically, isolate adherent cell populations and assess their receptor distributions. This provided a means to establish expression versus adhesive function relationships, a capability which can in the future be applied to the determination of relative functional efficiencies for individual receptor types, for instance in mutational analysis of receptor function. Moreover, using probability binning analysis, we were able to extract important information on the functional dynamics of receptor activation which would otherwise have been inaccessible. This assay is also compatible with intracellular staining (our unpublished results), which may be useful in revealing functionally mediated activation of signaling molecules. These results demonstrate the utility of the microsphere-based adhesion assay in detecting variable effects of both intra- and extracellular agents on integrin-mediated adhesion. More generally, this assay provides a new tool to elucidate the molecular mechanisms by which ligand binding to cell surface receptors participates in bidirectional cellular signaling. Acknowledgment The authors thank Maria Tuominen for expert technical assistance. References [1] E. Ruoslahti, Integrins, J. Clin. Invest. 87 (1991) 1–5. [2] R.O. Hynes, Integrins: versatility, modulation, and signaling in cell adhesion, Cell 69 (1992) 11–25.

[3] K.M. Yamada, B. Geiger, Molecular interactions in cell adhesion complexes, Curr. Opin. Cell Biol. 9 (1997) 76–85. [4] J.L. Mobley, E. Ennis, Y. Shimizu, Differential activationdependent regulation of integrin function in cultured human Tleukemic cell lines, Blood 83 (1994) 1039–1050. [5] J.L. Mobley, E. Ennis, Y. Shimizu, Isolation and characterization of cell lines with genetically distinct mutations downstream of protein kinase C that result in defective activation-dependent regulation of T cell integrin function, J. Immunol. 156 (1996) 948– 956. [6] V. Pichard, S. Honore, H. Kovacic, C. Li, C. Prevot, C. Briand, J.B. Rognoni, Adhesion, actin cytoskeleton organisation and the spreading of colon adenocarcinoma cells induced by EGF are mediated by alpha2beta1 integrin low clustering through focal adhesion kinase, Histochem. Cell Biol. 116 (2001) 337–348. [7] N. Choe, J. Zhang, A. Iwagaki, S. Tanaka, D.R. Hemenway, E. Kagan, Asbestos exposure upregulates the adhesion of pleural leukocytes to pleural mesothelial cells via VCAM-1, Am. J. Physiol. Lung Cell. Mol. Physiol. 277 (1999) L292–L300. [8] S.A. Santoro, M.M. Zutter, J.E. Wu, W.D. Staatz, E.U. Saelman, P.J. Keely, Analysis of collagen receptors, Methods Enzymol. 245 (1994) 147–183. [9] R.S. Sawhney, B. Sharma, L.E. Humphrey, M.G. Brattain, Integrin alpha2 and extracellular signal-regulated kinase are functionally linked in highly malignant autocrine transforming growth factor-alpha-driven colon cancer cells, J. Biol. Chem. 278 (2003) 19861–19869. [10] D.D. McAbee, F. Grinnell, Fibronectin-mediated binding and phagocytosis of polystyrene latex beads by baby hamster kidney cells, J. Cell. Biol. 97 (1983) 1515–1523. [11] W. Lee, J. Sodek, C.A. McCulloch, Role of integrins in regulation of collagen phagocytosis by human fibroblasts, J. Cell. Physiol. 168 (1996) 695–704. [12] J.T. Soini, M.E. Waris, P.E. Hanninen, Detection methods of microsphere based single-step bioaffinity and in vitro diagnostics assays, J. Pharm. Biomed. Anal. 34 (2004) 753–760. [13] T. Riikonen, L. Koivisto, P. Vihinen, J. Heino, Transforming growth factor-beta regulates collagen gel contraction by increasing alpha 2 beta 1 integrin expression in osteogenic cells, J. Biol. Chem. 270 (1995) 376–382. [14] P. Nykvist, K. Tasanen, T. Viitasalo, J. Kapyla, J. Jokinen, L. Bruckner-Tuderman, J. Heino, The cell adhesion domain of type XVII collagen promotes integrin-mediated cell spreading by a novel mechanism, J. Biol. Chem. 276 (2001) 38673–38679. [15] P.S. Ohashi, T.W. Mak, P. Van den Elsen, Y. Yanagi, Y. Yoshikai, A.F. Calman, C. Terhorst, J.D. Stobo, A. Weiss, Reconstitution of an active surface T3/T-cell antigen receptor by DNA transfer, Nature 316 (1985) 606–609. [16] R.T. Lee, F. Berditchevski, G.C. Cheng, M.E. Hemler, Integrinmediated collagen matrix reorganization by cultured human vascular smooth muscle cells, Circ. Res. 76 (1995) 209–214. [17] E. Fingerman, M.E. Hemler, Regulation of proteins in the VLA cell substrate adhesion family: influence of cell growth conditions on VLA-1, VLA-2, and VLA-3 expression, Exp. Cell Res. 177 (1988) 132–142. [18] L.A. Cantarero, J.E. Butler, J.W. Osborne, The adsorptive characteristics of proteins for polystyrene and their significance in solid-phase imunnoassays, Anal. Biochem. 105 (1980) 375–382. [19] BD Biosciences, Introduction to Flow Cytometry: A Learning Guide, in Manual Part Number: 11-11032-01, 2000, pp. 13–14, Becton–Dickinson and Company, San Jose, CA. [20] P. Nykvist, H. Tu, J. Ivaska, J. Kapyla, T. Pihlajaniemi, J. Heino, Distinct recognition of collagen subtypes by alpha(1)beta(1) and alpha(2)beta(1) integrins. Alpha(1)beta(1) mediates cell adhesion to type XIII collagen, J. Biol. Chem. 275 (2000) 8255–8261. [21] T. Kamata, W. Puzon, Y. Takada, Identification of putative ligand binding sites within I domain of integrin alpha 2 beta 1

Duplexed microsphere-based cellular adhesion assay / W.L. Connors, J. Heino / Anal. Biochem. 337 (2005) 246–255

[22]

[23]

[24]

[25]

(VLA-2, CD49b/CD29), J. Biol. Chem. 269 (1994) 9659– 9663. D.S. Tuckwell, L. Smith, M. Korda, J.A. Askari, S. Santoso, M.J. Barnes, R.W. Farndale, M.J. Humphries, Monoclonal antibodies identify residues 199-216 of the integrin alpha2 vWFA domain as a functionally important region within alpha2beta1, Biochem. J. 350 (Pt 2) (2000) 485–493. A.L. Berrier, A.M. Mastrangelo, J. Downward, M. Ginsberg, S.E. LaFlamme, Activated R-ras, Rac1, PI 3-kinase and PKCepsilon can each restore cell spreading inhibited by isolated integrin beta1 cytoplasmic domains, J. Cell. Biol. 151 (2000) 1549–1560. H. Wang, Q. Gai, X. Yang, Z. Li, B. Linders, S.A. Santoro, M.M. Zutter, Role of the alpha1 and alpha2 integrin cytoplasmic domains in cell morphology, motility and responsiveness to stimulation by the protein kinase C pathway, Cell Adhes. Commun. 7 (2000) 281–297. K. Podar, Y.T. Tai, B.K. Lin, R.P. Narsimhan, M. Sattler, T. Kijima, R. Salgia, D. Gupta, D. Chauhan, K.C. Anderson, Vascular endothelial growth factor-induced migration of multiple myeloma cells is associated with beta 1 integrin- and phosphatidylinositol 3-kinase-dependent PKC alpha activation, J. Biol. Chem. 277 (2002) 7875–7881.

255

[26] R.J. Tallarida, L.S. Jacob, The Dose–Response Relationship in Pharmacology, Springer-Verlag, New York, 1999. [27] E.C. Rosfjord, M. Maemura, M.D. Johnson, J.A. Torri, S.K. Akiyama, V.L. Woods Jr., R.B. Dickson, Activation of protein kinase C by phorbol esters modulates alpha2beta1 integrin on MCF-7 breast cancer cells, Exp. Cell Res. 248 (1999) 260– 271. [28] K. Vuori, E. Ruoslahti, Activation of protein kinase C precedes alpha 5 beta 1 integrin-mediated cell spreading on fibronectin, J. Biol. Chem. 268 (1993) 21459–21462. [29] S.L. Hsu, Y.H. Chou, S.C. Yin, J.Y. Liu, Differential effects of phorbol ester on growth and protein kinase C isoenzyme regulation in human hepatoma Hep3B cells, Biochem. J. 333 (1) (1998) 57–64. [30] M. Roederer, A. Treister, W. Moore, L.A. Herzenberg, Probability binning comparison: a metric for quantitating univariate distribution differences, Cytometry 45 (2001) 37–46. [31] P. Upla, V. Marjomaki, P. Kankaanpaa, J. Ivaska, T. Hyypia, F.G. Van Der Goot, J. Heino, Clustering induces a lateral redistribution of alpha 2 beta 1 integrin from membrane rafts to caveolae and subsequent protein kinase C-dependent internalization, Mol. Biol. Cell. 15 (2004) 625–636.