Contemporaneous cell spreading and phagocytosis: Magneto-resistive real-time monitoring of membrane competing processes

Biosensors and Bioelectronics 40 (2013) 82–88

Contents lists available at SciVerse ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

Contemporaneous cell spreading and phagocytosis: Magneto-resistive real-time monitoring of membrane competing processes A. Shoshi a,n, J. Schotter a, P. Schroeder a, M. Milnera a, P. Ertl a, R. Heer a, G. Reiss b, H. Brueckl a a b

AIT Austrian Institute of Technology, Molecular Diagnostics, Donau-City-Strasse 1, 1220 Vienna, Austria Thin Films & Physics of Nanostructures, University of Bielefeld, D2, Universitaetsstrasse 25, 33615 Bielefeld, Germany

a r t i c l e i n f o

a b s t r a c t

Available online 23 June 2012

Adhesion and spreading of cells strongly depend on the properties of the underlying surface, which has significant consequences in long-term cell behavior adaption. This relationship is important for the understanding of both biological functions and their bioactivity in disease-related applications. Employing our magnetic lab-on-a-chip system, we present magnetoresistive-based real-time and label-free detection of cellular phagocytosis behavior during their spreading process on particleimmobilized sensor surfaces. Cell spreading experiments carried out on particle-free and particle-modified surfaces reveal a delay in spreading rate after an elapsed time of about 2.2 h for particle-modified surfaces due to contemporaneous cell membrane loss by particle phagocytosis. Our associated magnetoresistive measurements show a high uptake rate at early stages of cell spreading, which decreases steadily until it reaches saturation after an average elapsed time of about 100 min. The corresponding cellular average uptake rate during the entire cell spreading process accounts for three particles per minute. This result represents a four times higher phagocytosis efficiency compared to uptake experiments carried out for confluently grown cells, in which case cell spreading is already finished and, thus, excluded. Furthermore, other dynamic cell-surface interactions at nano-scale level such as cell migration or the dynamics of cell attachment and detachment are also addressable by our magnetic lab-on-a-chip approach. & 2012 Elsevier B.V. All rights reserved.

Keywords: Lab-on-a-Chip Giant MagnetoResistance (GMR) biosensor Magnetic particles Normal Human Dermal Fibroblasts (NHDF) Phagocytosis Cell spreading

1. Introduction Cell-surface interactions such as cell spreading and phagocytosis represent important aspects in biology and are of special interest for biomedical applications. Adherent cells like fibroblasts continually probe their environment, and they need to attach to and spread on an underlying surface in order to perform numerous biological functions such as embryogenesis, maintenance of tissue structure, proliferation, differentiation, wound healing, metastasis or uptake of infectious agents as part of the immune response (Bardsley and Aplin, 1983; Cretel et al., 2010; Mrksich, 2000). A crucial parameter for immediate and long-term cell behavior is the surface characteristic of the adhesive substrate, comprising not only biomolecular and chemical features, but also physical properties such as stiffness, roughness and topography (Cavalcanti-Adam et al., 2007; Cretel et al., 2010). Thus, by designing nearly arbitrary surface characteristics using micrometer/ nanometer patterning techniques and biochemistry, the biocompatibility of biomaterials used for implants or rapid

n

Corresponding author. Tel.: þ43 50550 4311; fax: þ 43 50550 4399. E-mail address: [email protected] (A. Shoshi).

0956-5663/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bios.2012.06.028

wound healing applications can be improved by targeted tailoring of the cell-substrate interface bioactivity (Anselme, 2000; Jones, 2001; Pierres et al., 2003). The cellular decision making process to either spread on a surface or to remain rounded determines in most cases the fate of the cell, i.e. survival or initiation of programmed cell death (apoptosis). During the functional phases of cell spreading, thin lamellipodial protrusions creep onto the substrate surface, which lead to a shape-transformation from an initially spherical to a finally disk-shaped state with a steady increase of the surface-to¨ volume ratio (Dobereiner et al., 2004, 2005). Common for processes such as cell spreading and phagocytosis is their need for additional cell plasma membrane, the supply of which is limited. In fact, the entire cell spreading process can be considered as the attempt of a cell to internalize a particle that is too large for phagocytosis. Indeed, experiments on phagocytosis of particles by granulocytes showed that both processes obey similar characteristics (Evans et al., 1993; Herant et al., 2006; Stewart et al., 1989). Apart from conventional optical end-point detection methods, other optical techniques such as internal reflection microscopy (IRM) or total internal reflection fluorescence microscopy (TRIFM) are employed to follow the process of cell spreading with high spatial and temporal resolution (Burmeister et al., 1998; Cretel

A. Shoshi et al. / Biosensors and Bioelectronics 40 (2013) 82–88

et al., 2010; Ryzhkov et al., 2010). In addition to the visualized cell/surface contact area, also the separation distance between the cell and the substrate surface can be quantified. In this study we investigate the phagocytic behavior of human fibroblast cells during their spreading process on particle-immobilized sensor surfaces. Special focus is put on the susceptibility of cells during their adhesion process to the cell-membrane competitive mechanism of phagocytosis. To monitor the influence of particle uptake on the spreading characteristics in real-time, we employ our previously introduced magnetic approach based on magnetoresistive (GMR) sensors, magnetic particles and microfluidics (Shoshi et al., 2012). Research on real-time monitoring of cell-surface interactions at nano-scale level is of high importance in cellular biophysics, material science and the development of future biomaterials for biomedical applications (Gardel and Schwarz, 2010).

2. Magnetoresistive detection principle The concept of magnetic real-time monitoring of cell phagocytosis during their spreading processes is based on measuring changes of the local magnetic stray field of pre-immobilized superparamagnetic particles (beads) within embedded magnetoresistive sensors. These stray field variations are induced by distance changes of the beads relative to the sensor during cellbead interaction. The approach is sketched in Fig. 1. Initially, beads are immobilized onto various sensors of the biochip surface. The bead-induced response of the underlying sensor depends on the magnetic moment, the number of immobilized beads as well as the mean vertical separation distance r(t) to the sensor layer. When cells attach to and spread on the chip surface, they start to internalize the beads, which results in an increase of their mean vertical distance (Fig. 1(b)). With the increasing distance, the stray field strength of a bead at a point within the sensor region decreases approximately by r  3, which results in a lower sensor signal. As long as all other parameters are fixed, the sensor output decreases proportionally to the time dependent progress of cell spreading and phagocytosis, which allows continuous monitoring in real-time. Besides the general advantages of our magnetic approach (Shoshi et al., 2012), another unique feature is the capability of measuring the average post-phagocytosis bead-to-sensor separation distance within living cells. In view of drug delivery applications, this information could be used to study in addition to phagocytosis also possible subsequent exocytosis behavior. By tailoring the surface bio-chemistry and/or physical characteristics of the magnetic drug carriers, this

83

methodology might be used to analyze and tune the drug dwell time in cancerous cells/tissue, thus improving their efficiency.

3. Materials and methods The most relevant constituent parts of our magnetic lab-on-achip (MAGLab) setup as well as a detailed description regarding the sensor-chip fabrication and characteristics, bead properties and micromagnetic simulations can be found in Shoshi et al. (2012). Therefore, only a brief summary is given here. 3.1. MAGLab setup Two decoupled pairs of Helmholtz (HH) coils generate homogeneous magnetic fields parallel (in-plane, max. 723 kA/m) and perpendicular (out-of-plane, max. 736 kA/m) to the chip plane. These fields are required for sensor characterization and magnetizing the beads during the sensor read-out process. In the center of both HH-coils a chip holder is positioned which is also equipped with an external temperature control. The fluidical and electrical connections are established by a connector lid. In addition, we use a long-range microscope with a CCD camera for on-chip optical observations. The magnetic fields as well as the read-out of the sensors are computer controlled (LabViewTM, www.ni.com). 3.2. Sensor-chip characteristics A continuous stack of ten Ni80Fe20/Cu-double layers in the second antiferromagnetic coupling maximum (AFCM) is sputterdeposited on a 20 mm  20 mm silicon substrate and patterned into meander-shaped sensors. Each chip consists of 16 sensors of four different sizes arranged in two identical rows, representing reference and magnetic particle detection sensors (ref- and biosensors). The resistance of each sensor is about 6–8 kO and its GMR amplitude is around 12%. All sensors are protected by a 230 nm thick Si3N4 passivation layer from interactions with cells and/or fluids. The surface of the passivation layer is functionalized by a 2% APTES solution to ensure cell growth to the chip-surface and immobilization of surface-modified beads. For all cell experiments, a fluidic channel system made of cross-linked polydimethylsiloxane (PDMS) is mounted on the chip which enables direct access solely to the biosensors for magnetic particle immobilization and cell incubation. After surface functionalization and PDMS mounting, the chip is assembled into the chip holder of the MAGLab setup.

Fig. 1. (a) Sketch of spread cells on top of a magnetoresistive sensor surface covered by pre-immobilized magnetic particles. (b) Cross-section sketch of phagocytosis stages: After particle recognition, the cell starts engulfing and finally internalizing the particle in phagosomes, leading to an increase of the particle-to-sensor distance r(t).

84

A. Shoshi et al. / Biosensors and Bioelectronics 40 (2013) 82–88

3.3. Magnetic beads We employ streptavidin-coated superparamagnetic Magnetite (Fe3O4) particles of 1.2 mm (nominally 1.0 mm) in diameter from MagSense Life Science. Further physiochemical bead properties are evaluated by electron microscopy, zeta potential and dynamic light scattering measurements. 3.4. Cell culture Normal Human Dermal Fibroblast (NHDF) cells from PromoCell are cultured at 37 1C under humidified atmosphere of 5% CO2 in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). For on-chip cell experiments, additionally 10 mM HEPES buffer solution and 1% gentamicin are added (all components from PAA Laboratories GmbH). 3.5. Micromagnetic simulations The GMR-sensor is modeled by a Ni80Fe20/Cu/Ni80Fe20trilayer in the second AFCM, where each layer has a thickness of 2 nm. The lateral sensor area of 2.26 mm  2.26 mm is subdivided into cells of 20 nm  20 nm. The bilinear (JL) and biquadratic (JQ) exchange coupling constants and the associated effective exchange coupling Jn ¼2JQ þJL¼7.74 mJ/m2 are obtained from Downhill Simplex Simulations. The maximum bead-magnetizing field of 36 kA/m applicable in our MAGLab setup results in a maximum bead moment of M¼33 kA/m, which is another input factor of the simulation. 3.6. Cell adhesion/spreading Cell spreading is studied on sensor-like plain (Si/SiO2/Si3N4/ APTES) and bead modified (Si/SiO2/Si3N4/APTES-Beads) surfaces. For the bead-modified surfaces, streptavidin-coated 1.2 mm MagSense beads are immobilized additionally onto the APTES surfaces, resulting in a mean bead surface coverage of  23%. To enhance contrast in the microscopic imaging, we stain the NHDF cells with 4 mM Calcein–AM (Calbiochem and Merck-Chemicals) by a 30 min incubation accompanied with two times PBS washing steps before and after staining. To evaluate the spreading progress of the cells, we use the software ImageJ 1.41o.

measurement procedure. The beads are magnetized by an external homogeneous field perpendicular to the sensor plane and their magnetic stray field is detected by the embedded sensors. GMR-sensors basically respond to the in-plane component of the bead’s dipolar magnetic field. The remaining minimal signal arising from the perpendicular component is denoted as blank level in the calibration plot. A quantitative analysis of the signal is achieved by successively increasing the concentration of beads immobilized on the same chip. The GMR-sensors are calibrated in an aqueous environment, representing similar sensor conditions encountered during cell monitoring experiments. Basically, the output signal increases linearly with increasing density of immobilized beads and eventually goes into saturation due to the overlap of the dipolar stray fields of densely packed beads. The best operating point of our sensor is at half the saturation level, which is reached at about 25% coverage. Therefore, we aim at an initial 25%-bead surface coverage for the uptake experiments during cell spreading (Section 4.3). 4.2. Micromagnetic simulations We perform micromagnetic simulations (OOMMF, http:// math.nist.gov/oommf) to model the change of the GMR-sensor signal on increasing bead distance induced by cellular uptake. The total lateral sensor area of 2.26 mm  2.26 mm included in the simulation equals the average area per bead at the mean surface coverage of 22.3% encountered during the uptake experiments. In addition to the stray field originating from the bead centered within this unit cell, also the stray fields of the eight nearest neighboring beads are considered in the simulation (rectangular arrangement, see inset of Fig. 2). The uptake process is simulated by a steady increase of the vertical bead-array distance to the sensor surface, which leads to a decrease of the sensor signal due to the diminishing effect of the beads’ stray fields (Fig. 2). If the separation distance exceeds a threshold value of  600 nm, the stray field magnitude of the bead array within the simulated sensor region can be neglected. The total vertical sensor to beadcenter distance is given by adding the passivation layer thickness (230 nm) and the bead radius to the x-values denoted in Fig. 2. The results obtained by this sensor model are used to assess the average bead distance increase from the magnetoresistive measurements observed during the spreading/uptake experiments.

3.7. Real-time phagocytosis monitoring during cell spreading

4. Results and Discussions 4.1. Sensor calibration The dependency of the sensor signal on the sensor surface coverage by MagSense beads of 1.2 mm in diameter is investigated in Shoshi et al. (2012). Here, we give a brief description of the

0.10

0.08

0.06 GMR [%]

After chip preparation and assembling into the MAGLab setup, magnetic particles of 1.2 mm in diameter are immobilized on the biosensor surfaces by adding 400 ml of a 25 mg/ml particle-dH2O solution. After replacing the water by DMEM, approximately 6000 NHDF cells are seeded onto the bead immobilized sensor-area and the magnetoresistive response of particle-covered biosensors and non-covered ref-sensors are measured in appropriate time intervals. At every stage of each experiment, both the ref- and biosensors response is recorded. In addition, optical microscopy images are taken at different stages of the experiment. At the end, the cells are completely removed by trypsin-EDTA (0.25% trypsin and 0.02% EDTA) followed by several PBS and dH2O washing steps.

0.04

0.02

0.00 0

300

600 900 1200 bead distance [nm]

1500

1800

Fig. 2. Dependence of the magnetoresistance on the vertical bead-to-sensor separation. The inset shows the 9-bead array configuration together with the simulated sensor unit area of side length a ¼2.26 mm.

A. Shoshi et al. / Biosensors and Bioelectronics 40 (2013) 82–88

4.3. Real-time monitoring of bead phagocytosis during cell spreading Interactions between cells and immobilized beads are investigated by real-time monitoring the time evolution of the GMR-signal. After attachment to the sensor surface, cells undergo an adhesion/spreading process and the associated GMR-response is highly dependent on its stage of progress. This shape transformation from an initial spherical to a final flattened state implies two relevant morphological changes. First, the cell surface contact area increases with time and consequently the cell-bead interaction range, which leads to a time dependent alteration of the overall distance between the immobilized beads due to phagocytosis. Second, with proceeding spreading stage also the surface-to-volume ratio increases, which rivals the surplus membrane area required for vesicle formation during bead phagocytosis. Because of these competing events, one could expect a delayed cell spreading progress on bead-immobilized surfaces in comparison to plain APTES control surfaces. Therefore, in addition to the online GMR-monitoring, we also analyze cell spreading on bead-covered and bead-free surfaces by optical microscopy. Moreover, in preliminary experiments we also determined phagocytosis-relevant cell and bead specific properties and a comprehensive overview is given in Shoshi et al. (2012). Briefly, this includes the uptake capacity of NHDF cells, the characterization of significant bead parameters, i.e. size, surface chemistry and charge, as well as their long-term stability within phagolysosomes. In this study, all experiments are carried out with identical cells and beads.

85

4.3.1. Cell spreading Immediately after seeding, the cells start to sediment and attach to the surface. The attachment process is followed by a lag time of about 10 min during which period no change in cell morphology is observed. Afterwards, the cells spread at a certain rate until a final disk-like cell state with maximum spreading area is reached. A crucial parameter for the kinetics of cell spreading is the characteristics of the substrate, which is controlled via specific surface biochemistry, topography and its mechanical properties (Cretel et al., 2010). Here, as a reference experiment for the real-time magnetoresistive phagocytosis monitoring, we investigate cell spreading by phase contrast (fluorescence) microscopy both on plain APTES surfaces and on surfaces with beads immobilized onto the APTES layer. The average bead surface coverage employed for the cell spreading experiments is  23%, which is identical to those used during the magnetic monitoring experiments. Fig. 3(a) and (b) shows the time evolution of the cellular top view projection area along with the corresponding phase contrast and fluorescent time-lapse images. In Fig. 3(a), each data point represents the average value of about 30 unsynchronized cells. The first 2.2 h of cell spreading can be considered as quasi-linear with similar spreading rates of 700 mm2/h for both surfaces. For later times, the spreading process in both cases shows an asymptotic behavior. However, there is a divergence observed with a clear inhibition in the spreading rate for the bead-modified surface. At equilibrium state, an average maximum projection area of 3900 mm2 is observed for the reference sample, which is comparable to literature values

5000

APTES reference bead modified

projection area [µm2]

4000

3000

2000

1000

0 0

2

4

6

8

10 12 14 16 18 20 22 24

time [h]

Fig. 3. (a) Kinetics of cell spreading on plain APTES and bead-modified surfaces and (b) time-lapse of cell spreading on plain APTES (left) and bead modified surfaces (right) with the corresponding phase contrast, fluorescent and overlay images; scale bar 20 mm.

86

A. Shoshi et al. / Biosensors and Bioelectronics 40 (2013) 82–88

(e.g. 4170 mm2, Woods et al., 1986). For the bead-modified sample, the average maximum projection area amounts to 3250 mm2, which is 17% less than observed in the reference case. The expected delay in spreading rate due to the contemporaneous plasma membrane loss and volume increase as a consequence of phagocytosis becomes important only after an elapsed time of around 2.2 h. Within this time interval, no noticeable delay is observed between the investigated surfaces, thus implying that the additional cell membrane area utilized for bead engulfment is supplied by rapid membrane recycling. At times beyond 2.2 h, the projection area of cells spreading on top of bead-modified surfaces only increases very slowly, while the projection area of the reference cells continues to grow. Thus, a time r2.2 h can be considered as spreading saturation time for the case of bead-modified surfaces. The maximum difference in projection area is reached after about 6 h (  900 mm2), which corresponds to an additional membrane area utilized for vesicle formation sufficient for about 400 phagocytosed beads (projection area difference taken twice due to dorsal and ventral membrane). At time 6 h, the cell projection area on bead-modified surfaces is about 2440 mm2, which leads to around 500 beads covered by each spread cell (taking into account the known bead surface coverage of  23%). The approximate agreement of these two values suggests that spreading and phagocytosis are competing processes limited by the capacity of endomembranes. As reviewed in Desjardins and Griffiths (2003), in addition to

the plasma membrane also endomembranes contribute to the phagosome formation. These recycling intracellular membrane compartments are limited and continual cell spreading accompanied by phagocytosis eventually leads to their successive depletion (Cannon and Swanson, 1992). The total plasma membrane area required to internalize 500 beads is 2260 mm2, which is equivalent to 163% of the total cell plasma membrane area of a rounded NHDF cell. In comparison, macrophages are capable of ingesting the equivalent of 48%–145% of their macroscopic surface area within 30 min (Cox et al., 2000). The average projection areas obtained from the last data set at time 23 h still suggests a slight increase in projection area, which could be attributed to further endomembrane recruitment originating from different sources (e.g. endosomes, lysosomes or endoplasmic reticulum) at different stages of phagocytosis (Booth et al., 2001; Braun and Niedergang, 2006).

4.3.2. Real-time monitoring of cell phagocytosis during cell spreading Exemplarily, the signal trace of a single biosensor over the entire cell spreading and uptake experiment is shown in Fig. 4(a) and (b), while the analyzed data of a set of biosensors is specified in Table 1. Simultaneously recorded time-lapse optical microscope images are depicted in Fig. 4(c). Each experiment starts with the read-out of ref- and biosensors at every stage indicated

0.14

0.12

0.10

0.11

0.08 GMR [%]

GMR [%]

cell phagocytosis exponential fit

0.13

cell phagocytosis reference

0.12

0.06 0.04

0.10 0.09

0.02 0.08 0.00 -40

-20

0

20 time [h]

40

60

80

0

1

2

3

4

time [h]

Fig. 4. (a) Time-lapse of the entire bead uptake experiment during adhesion and (b) zoom in of the early stage of bead uptake kinetics. (c) Time-lapse optical microscope images: in air/dH2O (I/II), with immobilized particles (III), 20 min (V1) and 1100 min after cell incubation (V2), during trypsin-EDTA detachment (VI1) and complete cell replacement (VI2).

A. Shoshi et al. / Biosensors and Bioelectronics 40 (2013) 82–88

Table 1 Real-time monitoring data analysis of bead uptake during cell spreading. Sensor

B1 C1 D1 A3 B3 C3 Average values

Cell

GMR drop (%) due to

Number Surface cov. (%)

Uptake

Con. Replacement growth

26 77 178 12 27 88 –

57 23 21 52 37 45 397 15

1 23 12 18 30 16 16 7 10

14 15 14 19 15 17 16.1 7 2.1

52 54 67 30 33 39 467 14

Saturation time (min)

55 81 70 115 135 125 97 732

by the letters I until VI. In air and dH2O (I, II) both sensor responses are equal and correspond to the blank signal. After bead immobilization in dH2O, about 23% of the sensor surface is covered (III). Accordingly, the signals of the biosensors increase, while the ref-sensors remain at blank level. To provide a suitable environment for cell growth, we exchange the water by DMEM without significant distortion of bead positions, thus retaining the biosensors’ previous GMR-responses (IV). At time zero in stage V, about 6000 cells are added to the chip, which in a spread-out state can form a confluent layer across the exposed chip-area defined by the fluidic chamber. Table 1 shows the counted number of cells on each sensor just after the cells complete sedimentation ( 5–10 min). At this time, about 16% of the sensor surface area is covered by the as-attached cells. At the end of each experiment the cells are detached by applying trypsin-EDTA and replaced by several washing steps (VI). The first rapid decrease of the GMR-signal after cell incubation is analyzed by an exponential fit function as shown in Fig. 4(b). The time required for the fit function to decay to 5% of its initial value is denoted as saturation time during the spreading process (Table 1), where the 5% threshold value is the mean standard deviation of the GMR-signals within the monitoring time of 2–4 h. In the following, the average values of the set of biosensors are presented. Within the average saturation time of 97 min, a mean signal drop of 39% relative to the full signal above blank occurs. This signal drop reflects cell spreading and the corresponding decreasing stray field of the magnetized beads due to their phagocytosis and agrees to the optically determined spreading saturation time within the standard deviation of 32 min. The number of beads internalized until saturation is around 320 per cell, which is calculated on the basis of the initial bead surface coverage and the cell spreading analysis data (Section 4.3.1). Thus, within this quasi-linear cell spreading regime, on average 3 beads per minute are phagocyted by each cell, which is four times higher compared to the uptake rate of already confluently grown cells (Shoshi et al., 2012). Following the average saturation time, cell spreading and bead uptake continue till a confluent layer is reached, resulting in another slower average GMR-signal drop by about 16% (relative signal change from level after uptake time according to the fit curve to level before replacement, last three data points averaged to obtain pre-replacement signal level). Typically, this drop is reached well before the time at which the cells are detached by Trypsin-EDTA (73 h), which is coherent to the usual time it takes NHDF cells to spread-out and grow confluently (Fig. 3). Statistical variations in sensor signals observed at times beyond formation of a confluent layer can be attributed to migration of cells with different bead loadings in and out of the sensor region and/or intracellular bead movements. In addition to determining time scales, the magnetic monitoring approach also allows to estimate the medium vertical distance the magnetic particles are lifted from the surface during phagocytosis.

87

Our micromagnetic simulations in Section 4.2 reveal that lifting beads vertically from the sensor surface leads to a corresponding GMRsignal decrease. In our measurements (Fig. 4), we observe a GMRsignal drop due to an increase of the vertical bead-to-sensor distance following their phagocytosis. In order to compare our measurement data (Fig. 4) and simulations (Fig. 2), both results are normalized, thus removing bead concentration dependent effects. By evaluating the normalized plots, it is possible to determine the average vertical distance increase of the beads due to their phagocytosis. To that end, the experimentally obtained relative signal drops between cell seeding (t¼0 h) and just before cell replacement (t70 h) are averaged, resulting in a relative signal level change of 54714% from the highest level at t¼0 (excluding the blank level). Concerning the normalized simulated distance-dependence, the same relative GMR signal decrease is obtained if the beads are lifted vertically from their initial (0 nm) to their final bead-to-sensor surface position of 120745 nm. In comparison, the lipid double layer of the plasma membrane surrounding the phagocytosed beads is about tl ¼10 nm thin (Alberts et al., 2008). Furthermore, based on the cell size in suspension ( 21 mm) and the final spreading area on bead-immobilized surfaces (3250 mm2), the estimated average height of a confluent cell body under the assumption of constant cell volume is about tc ¼1.5 mm (simplified view: uniform average height). Thus, the approximate range of possible distances of a phagocytosed bead with db ¼1.2 mm diameter to the sensor surface is 20–280 nm (lower value¼2  tl; upper value¼tc  db  2  tl). The average value of this range of 150 nm agrees well to both our real-time monitoring sensor results and the phagocytic membrane thickness during bead engulfment and vesicle formation previously reported by electron microscopy (Korn and Weisman, 1967). Each monitoring experiment ends with complete cell replacement. As can be seen from the corresponding optical microscope images (Fig. 4 VI1 and VI2), no beads remain at the sensor surface following replacement, suggesting that all beads are internalized by the cells during the preceding phagocytosis monitoring experiment, which is consistent to the maximum cell uptake capacity of 1270 beads of 1.2 mm in diameter determined previously (Shoshi et al., 2012). Following replacement, the GMR-response drops back to the blank level, which demonstrates that our method is reproducible and background-free.

5. Conclusion The phagocytic behavior of human fibroblast cells during their spreading process on bead-immobilized sensor surfaces is investigated by real-time monitoring of the magnetoresistive sensor signal evolution. Reference cell spreading measurements on beadmodified and bead-free APTES surfaces show that cell spreading and phagocytosis are competing events, which rival the surplus plasma membrane area required for both processes. As a result, the difference in saturation spreading area corresponds to the total membrane area required to engulf the respective number of beads. With the presented magnetic cell monitoring platform, we can also determine the average vertical separation distance of the phagocytosed beads from the surface, which is around 120 nm. Real-time magnetoresistive measurements reveal that the uptake rate is not a linear function with time. It is higher at early stages and it decreases steadily until it reaches saturation after an average saturation time of 97 min. The associated bead uptake rate accounts for three beads per minute, which shows a four times higher uptake efficiency compared to the uptake rate of already confluently grown cells. Thus, for applications based on particle internalization into cells (e.g. magnetic cell therapy), it is advantageous to seed and spread cells on bead-modified surfaces for most efficient labeling. Future on-chip cell investigations will

88

A. Shoshi et al. / Biosensors and Bioelectronics 40 (2013) 82–88

also focus on cell migration assays as well as monitoring of dynamic processes such as cell attachment and detachment.

Acknowledgments We gratefully acknowledge M. Purtscher and F. Bellutti for NHDF cell supply as well as T. Uhrmann for LabView programming. The research leading to these results has received financial ¨ support by the Oesterreichische Forschungsforderungsgesellschaft (FFG) under Grant no. 810985. References Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., Walter, P., 2008, Molecular Biology of the Cell, 5th Edition, Library of Congress Cataloging-in-Publication Data. Anselme, K., 2000. Biomaterials 21, 667–681. Bardsley, W.G., Aplin, J.D., 1983. Journal of Cell Science 61, 365–373. Booth, J.W., Trimble, W.S., Grinstein, S., 2001. Seminars in Immunology 13, 357–364. Braun, V., Niedergang, F., 2006. Biology of the Cell 98, 195–201. Burmeister, J.S., Olivier, L.A., Reichert, W.M., Truskey, G.A., 1998. Biomaterials 19, 307–325. Cannon, G.J., Swanson, J.A., 1992. Journal of Cell Science 101, 907–913.

Cavalcanti-Adam, E.A., Volberg, T., Micoulet, A., Kessler, H., Geiger, B., Spatz, J.P., 2007. Biophysical Journal 92, 2964–2974. Cox, D., Lee, D.J., Dale, B.M., Calafat, J., Greenberg, S., 2000. Proceedings of the National Academy of Sciences 97 (2), 680–685. Cretel, E., Touchard, D., Benoliel, A.M., Bongrand, P., Pierres, A., 2010. Journal of Physics: Condensed Matter 22, 194107. Desjardins, M., Griffiths, G., 2003. Current Opinion in Cell Biology 15, 498–503. ¨ Dobereiner, H.-G., Dubin-Thaler, B., Giannone, G., Xenias, H.S., Sheetz, M.P., 2004. Physical Review Letters 93 (10), 108105–108111. ¨ Dobereiner, H.-G., Dubin-Thaler, B.J., Giannone, G., Sheetz, M.P., 2005. Journal of Applied Physiology 98, 1542–1546. Evans, E., Leung, A., Zhelev, D., 1993. Journal of Cell Biology 122 (6), 1295–1300. Gardel, M., Schwarz, U., 2010. Journal of Physics: Condensed Matter 22, 190301. Herant, M., Heinrich, V., Dembo, M., 2006. Journal of Cell Science 119 (9), 1903–1913. Jones, F.H., 2001. Surface Science Reports 42, 75–205. Korn, E.D., Weisman, R.A., 1967. Journal of Cell Biology 34, 219–227. Mrksich, M., 2000. Chemical Society Reviews 29, 267–273. Pierres, A., Eymeric, P., Baloche, E., Touchard, D., Benoliel, A.-M., Bongrand, P., 2003. Biophysical Journal 84, 2058–2070. ¨ ¨ Ryzhkov, P, Prass, M, Gummich, M, Kuhn, J-S, Oettmeier, C, Dobereiner, H-G, 2010. Journal of Physics: Condensed Matter 22, 194106. Shoshi, A.I., Schotter, J., Schroeder, P., Milnera, M., Ertl, P., Charwat, V., Purtscher, M., Heer, R., Eggeling, M., Reiss, G., Brueckl, H., 2012. Biosensors and Bioelectronics 36 (1), 116–122. Stewart, M.G., Moy, E., Chang, G., Zingg, W., Neumann, A.W., 1989. Colloids and Surfaces 42, 215–232. Woods, A., Couchman, J.R., Johansson, S., Hook, M., 1986. The EMBO Journal 5 (4), 665–670.