Biomedical Microdevices

, Volume 14, Issue 6, pp 1129–1140

Cell-based microfluidic device for screening anti-proliferative activity of drugs in vascular smooth muscle cells


  • R. Rodriguez-Rodriguez
    • School of Pharmacy, Department of PharmacologyUniversidad de Sevilla
    • Instituto Nacional de Microelectrònica (IMB-CNM, CSIC)
  • S. Demming
    • Institut für MikrotechnikTechnische Universität Braunschweig
  • S. Büttgenbach
    • Institut für MikrotechnikTechnische Universität Braunschweig
  • M. D. Herrera
    • School of Pharmacy, Department of PharmacologyUniversidad de Sevilla
  • A. Llobera
    • Instituto Nacional de Microelectrònica (IMB-CNM, CSIC)
    • Institut für MikrotechnikTechnische Universität Braunschweig

DOI: 10.1007/s10544-012-9679-y

Cite this article as:
Rodriguez-Rodriguez, R., Muñoz-Berbel, X., Demming, S. et al. Biomed Microdevices (2012) 14: 1129. doi:10.1007/s10544-012-9679-y


This paper presents a microfluidic device consisting of five parallel microchambers with integrated readout-grid for the screening of anti-proliferative activity of drugs in vascular smooth muscle cells (VSMC). A two-level SU-8 master was fabricated and replicated with poly(dimethylsiloxane), PDMS, using standard soft-lithographic methods. The relative small height (4–10 μm) of the integrated grid allowed the identification of single-cells or cell groups and the monitoring of their motility, morphology and size with time, without disturbing their proliferation pattern. This is of particular interest when considering VSMC which, apart of being crucial in the atherosclerotic process, do not proliferate in a single layer but in a non-homogenous hill and valley phenotype. The performance of the microfluidic device has been validated by comparison with conventional culturing methods, proving that the cell proliferation remains unaffected by the microchamber structure (with the integrated grid) and the experimental conditions. Finally, the microfluidic device was also used to evaluate the anti-proliferative activity of curcumin and colchicine in VSMC. With this cellular type, the anti-proliferative activity of curcumin (IC50 = 35 ± 5 μM) was found to be much lower than colchicine (IC50 = 3.2 ± 1.2 μM). These results demonstrate the good performance of the microfluidic device in the evaluation of the anti-proliferative activity (or cytotoxicity) of drugs.


Microfluidic deviceReadout grid-integrated parallel microbioreactorsCell proliferation assayVascular smooth muscle cellsAntiproliferative drug delivery

1 Introduction

The interest in developing microfluidic structures for biological applications has increased exponentially in the last few years, specially due to the expansion of soft lithographic methods based on poly(dimethylsiloxane) (PDMS) (Duffy et al. 1998). PDMS plays a significant role in that advance for being i) completely biocompatible and non-toxic, ii) flexible, iii) gas permeable, iv) low cost and v) for allowing the fabrication of structures with high aspect ratio (McDonald et al. 2000). Additionally, from an optical point of view, PDMS has a high transmittance from the visible to the near infra-red, making it an optimal candidate for optical microfluidic systems (Whitesides et al. 2001).

Interesting contributions combining PDMS-based microsystems and biological applications can be found in several areas, ranging from the measurement of small biological molecules, for instance the detection of DNA hybridization in automated, simple and cheap microstructures (Marasso et al. 2010), to whole cell analysis, such as monitoring single-cell tracking and proliferation using a hybrid PDMS-agar-glass microfluidic system (Wong et al. 2010). In mammalian cells analysis, the development of microfluidic structures for proliferation, functional, pharmacological or physiological studies has been summarized in relevant revisions principally based on neurons (Wang et al. 2009; Gross et al. 2007), stem cells (Gupta et al. 2010) or endothelial cells (Young and Simmons 2010).

One of the main interests in cell-based microsystems is the generation of suitable screening tools for massive drug analysis. Current drug screening processes rely on complex protocols with high reagent consumption, the use of expensive non-reusable sterile tools and the necessity of qualified personnel to avoid sample contamination. On the other hand, microfluidic devices, systems and platforms may represent a suitable alternative to these methods for being low cost, compact, integrated and flexible structures, with low reagent consumption, easy to automate and manipulate (Whitesides 2006). Drug screening microfluidic systems have been already reported. Accordingly, some interesting examples may be: i) flow-cytometer-based platforms for multiple single-cell screening of anti-cancer drugs(Wlodkowic et al. 2009), ii) a device trying to mimic the tight blood–brain barrier for testing particle systems carrying drugs (Genes et al. 2007), iii) integrated devices combining microfluidics and microelectrode arrays for the localized drug delivery to cellular networks (Kraus et al. 2006), iv) alginate-based systems for anti-cancer drug screening in a three dimensional cell culture (Chen et al. 2010), v) platforms for the analysis of the physiological response of hepatic cells (Novik et al. 2010) or vi) cell culture systems for monitoring drug delivery (Hsieh et al. 2009). Even though most of these examples are a significant step forward in drug screening tools, its massive implantation is hampered by either the technological complexity and/or the expensive peripheral equipment needed.

In this paper, readout grid-integrated parallel microbioreactors were developed for anti-proliferative drug screening in vascular smooth muscle cells (VSMC). The base of this approach relies on the fact that atherogenesis and atherosclerotic plaque formation, which are importantly implicated in cardiovascular complications, may involve an excessive proliferation and migration of VSMC (Ross 1999). For this reason, simple, cheap, automated and low reagent consumption microfluidic devices for the identification of good anti-proliferative candidates are of interest. It should be emphasized that the identification and monitoring of the same representative area with the grid is of particular interest when working with VSMC since this cellular type did not proliferate in a simple monolayer. Conversely, VSMC grows in a heterogeneous hill and valley pattern that may impede a suitable interpretation of the drug effect. Moreover, the micrometric size of the present grid pattern integrated on the reactor is advantageous for allowing the observation of both cells and grid patterns at the same plane without perturbing the flow path or cell proliferation. It is also important to remark that this microfluidic structure was designed for working under flow conditions reducing the manipulation of the culture (medium and drugs were automatically inoculated with a simple pump) thus reducing contamination problems.

In this study, VSMC proliferation in the PDMS microbioreactor have been investigated and the experimental conditions for the anti-proliferative assay have been optimized in terms of initial cell inoculum and fetal bovine serum (FBS) concentration in the culture medium. The anti-proliferative activity of colchicine and curcumin has been evaluated by injecting several drug concentrations in specific reactor chambers containing 24 h attached/proliferated VMSC. Colchicine and curcumin are natural drugs intensively investigated as anti-tumor substances in relation to their important effects in cell proliferation (Kunnumakkara et al. 2008). The mechanism of action of both substances is well described in the literature: colchicine interacts with tubulin altering the assembly dynamics of microtubules and consequently blocking cell division (Bhattacharyya et al. 2008), whereas curcumin acts at various stages of cell progression including cell cycle and apoptosis (Kuttan et al. 2007). Although the use of colchicine to block cell mitosis is very common, few applications of this anti-proliferative drug in vascular cells have been reported (Ivanov et al. 2004). On the contrary, numerous studies that focused on the inhibitory effects of curcumin on proliferation and migration of vascular cells have been published in the last few years (Pae et al. 2007; Qin et al. 2009; Yu and Lin 2010). Thus, in the present work, the anti-proliferative activity of these drugs was evaluated by monitoring the amount of area covered by VSMC with time at various drug concentrations and comparing with the controls (containing culture medium no supplemented with drugs). Also, single cells have been identified and analyzed in terms of morphology, motility and proliferation.

2 Material and methods

2.1 Parallel PDMS microbioreactor design and fabrication

The microfluidic device presented in this work is shown in Fig. 1. As it can be seen, five identical reactor chambers of 9 μL with independent fluidic inlet and outlet channels were integrated in a single microchip (21 × 39 mm2). Each chamber contained five grid arrays uniformly distributed along the reactor and numbered with Roman letter numbers from left to right. The grid was a 10 × 10 matrix (100 × 100 μm2) labeled with numbers (grid rows) and letters (grid columns) as illustrated in the figure.
Fig. 1

Images illustrating the microbioreactor used for monitoring VSMC proliferation, showing: five identical reactor chambers arranged in parallel (left), one single reactor chamber with five implemented grid arrays (top right) and one grid array with uniformly distributed grid columns and grid rows (bottom right)

The PDMS microbioreactor was fabricated following a modification of a soft-lithographic protocol already reported by the group (Demming et al. 2009; Llobera et al. 2007). By completeness, it is briefly reported in this subsection. The system was obtained by defining a two-level master. The first level defined the reactor chamber by using the negative tone SU-8 polymer (SU-8 50 from MicroChem, Corp., Newton, MA, USA; total height = 230 μm). The second level was used for implementing the grid structure on top of the reactor chamber (SU-8 5 from MicroChem, Corp., Newton, MA, USA; total height between 5 and 10 μm). Propylene glycol methyl ether acetate (PGMEA, MicroChem Corporation, Newton, MA, USA) was used for developing the masters after the required exposure and bake. Then, the master was replicated with PDMS (Sylgard 184 elastomer kit, Dow Corning, Midland, MI, USA). After 30 min of curing at 80 °C, the structured PDMS was irreversibly bonded to a glass chip previously treated with oxygen plasma (plasma activate flecto 10USB, Plasma Technology, Germany). Finally, the needles (∅ of 0.8 × 40 mm, BRAUN Sterican, Germany) were assembled in the inlet and outlet channels of the chamber reactors and sealed with biocompatible silicone rubber glue (RS Components 692-542, Germany).

2.2 Cell culture

Primary cultures of VSMC were obtained from rat thoracic aorta by enzymatic dissociation with collagenase (Sigma, Spain). Male Wistar rats (Experimental Animal Service of the University of Seville, Spain) of 8 weeks age and 250 ± 20 g body weight were used. Cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Gibco®- Invitrogen, Spain) supplemented with 10 % fetal bovine serum (FBS) and 1 % antibiotic-antimytotic solution (containing: 10 kU mL−1 penicillin G sodium, 10 mg mL−1 streptomycin sulfate and 25 μg mL−1 amphotericin B) (Gibco®- Invitrogen, Spain). Experiments were performed with cells at passages 3-5. VSMC proliferated showing the typical hill-and-valley phenotype and spindle-like morphology.

The animal-related experimental procedures had the ethical approval and followed the guidelines issued by the Ethical Committee on Investigation in Animals of the University of Seville.

2.3 Proliferation protocol in microbioreactors

Microbioreactors were sterilized with UV light for 30 min. Then, after treatment with 0.05 % trypsin-EDTA (5 min, 37 °C) (Gibco®- Invitrogen, Spain) and cell counting using a Neubauer’s chamber, the suitable number of VSMC was manually inoculated with a 1 mL-syringe in the corresponding reactor chamber. After 24 h-cell adhesion and proliferation (37 °C, 5 % CO2), the inlet needle was suitably connected to a Minipuls peristaltic pump (Gilson Inc., USA) by using appropriate tubing, previously filled with DMEM (supplemented or not with several drugs depending on the experimental protocol). The outlet needle was additionally connected to a waste recipient flask thus closing the fluidic circuit. This configuration should avoid media evaporation and bubble formation for the duration of the assay. After 24 h adhesion, a 200 μL min−1 continuous flow was applied for 5 min to ensure optimal perfusion of the microbioreactor with the media under study. According to previous studies (Martinez-Gonzalez et al. 2008), media was changed every 48 h using the same flow conditions, thus providing a proper oxygen and nutrients supply. This automated configuration reduces human intervention facilitating manipulation and reducing contamination problems. Cell proliferation was daily monitored throughout the experiment.

2.4 Acquisition and analysis of optical microscopy images

Image acquisition was performed with a CKX41 inverted microscope (Olympus, Japan) (10x objective; NA 0.30) coupled to a SC30 digital camera (Olympus, Japan) using the Analysis GetIT v1.0 software (Olympus, Japan). After a general inspection of the reactor chamber, three representative proliferation areas (0.15 mm2 each) were selected inside the grid arrays region (commonly in those arrays labeled with numbers II, III and IV). Once identified, the same areas were inspected daily throughout the duration of the experiment. As previously commented, this was of particular interest when working with VSMC for their heterogeneous proliferation pattern. The amount of surface covered by VSMC was determined using the ImageJ 1.38x software (USA). Images were firstly calibrated to its real size by considering the readout grid dimensions. Next, the area of individual or groups of cells in the region under study was manually obtained. The total covered area was expressed as percentage of coverage (%Cov). In some cases, coverage data were rescaled by considering the 24 h-percentage coverage as 0 (rescaled %Cov). This modification facilitated the comparison between experiments made in different days or under different experimental conditions.

It should be noted although the grid and the bottom part of the reactor could not be simultaneously focused with a 10x objective (Fig. 2), the obtained image provides enough information so as to determine the relative position of the cells at the grid, as well as to monitor their proliferation patterns. Therefore, the configuration proposed in this work is validated.
Fig. 2

Images of a reactor chamber just after VSMC inoculation (0 h, top panels) or after 70 h of cell proliferation (bottom panels) focusing on either the grid (left) or the cells (right). White arrows indicate cells that appear at the same time in the grid and the cell focus

2.5 Drugs preparation

Colchicine (AppliChem, Germany) was dissolved in phosphate buffered saline (PBS) to obtain 250 μM. A stock solution of curcumin 104 μM (Sigma, Spain) was prepared in dimethyl sulfoxide (DMSO). Dilutions were made by adding the suitable amount of the stock drug solution in 1.5 mL of cell culture medium (DMEM).

2.6 Statistical analysis

Values were expressed as average standard deviation for n analyzed areas from N reactor chambers. Significant differences between data were determined with analysis of variance or Student’s t-test, as appropriated, using GraphPad Prism software (v. 5.01, USA). Differences were considered significant when p < 0.05.

3 Results and discussion

3.1 Study of the VSMC proliferation: comparison with standard methods

In odert to determine whether the presence of the integrated grid in the microchamber structure or the fluidic conditions affect the optimal proliferation pattern of the cells, VSMC proliferation in the microbioreactor was compared with that obtained using conventional culture flasks. Experimentally, 1·106 cell mL−1 were inoculated in both structures and the proliferation was monitored in terms of %Cov according to previous sections. After inoculation, a total number of (4.8 ± 0.8)·102 cells (n = 3, N = 2) were found suspended in each grid reactor area (without considering the files containing numbers and the columns containing letters). After 16 h proliferation, a 40 ± 11 % of cells were found attached on bottom glass slice [(1.5 ± 0.3)·102 cells] resulting in a 21 ± 6%Cov. The variation of the %Cov with time for cells proliferating in both the microbioreactor structure and the flask is illustrated in Fig. 3(a).
Fig. 3

Comparison of VSMC proliferation in the microbioreactor and in a conventional culture flask after an initial inoculation of 1·106 cell mL−1. (a): Graph showing percentages of coverage (%Cov) of VSMC with time in both culture systems. Error bars represent standard deviation (n = 3, N = 2). (b): Images of cultured VSMC at three different times (in hours) in a conventional flask (top) and in the microbioreactor (bottom)

In both cases, attached VSMC proliferated increasing the %Cov until stabilization between 100 and 120 h after inoculation (90 % confluence). From that point, the %Cov increased slowly to almost 100 % confluence. It is important to notice that no significant differences were found along the experiment in the %Cov of VSMCs growing in microbioreactors compared with conventional flasks. Additionally, the typical spindle-like morphology and the hill-and-valley phenotype characteristic from cultured VSMCs was observed in both systems proving that cells were not under stress conditions (Fig. 3(b)). The growth rate was evaluated throughout the doubling coverage time (DCT) (in hours), a variation of the conventional doubling time commonly found in the literature (Loukotova et al. 1998) but with covered area instead of cell number thus considering the fact that VSMC proliferate attached. Hence, the DCT, defined as the time necessary to double the covered area percentage, was calculated according to Eq. 1,
$$ DCT = \frac{{\left( {{t_2} - {t_1}} \right)\cdot { \log }(2)}}{{\left[ {\log \left( {\% Co{v_2}} \right) - { \log }\left( {\% Co{v_1}} \right)} \right]}} $$
where %Cov1 and %Cov2 were the percentage of cell coverage at t1 and t2 (in hours). Regarding the first three days of analysis (that corresponds with the exponential proliferation region), the DCT was found to be 33 ± 4 h and 32 ± 4 h for VSMC growing in the microbioreactor and in the flask, respectively, with no significant differences between them. It also is important to note that both values are in concordance with those previously reported (Loukotova et al. 1998).

3.2 Optimization of the proliferation protocol

The proliferation protocol was optimized for drug screening considering the initial inoculum concentration and FBS concentration to ensure the suitable monitoring of cell proliferation in terms of %Cov.
  1. a)

    Determination of the initial inoculum concentration

Microbioreactors were inoculated with 5·105, 1·105, 5·104, 1·104 or 5·103 cells mL−1 and VSMC proliferation was monitored following the protocol previously detailed for around 120 h. Concentrations between 5·105 and 1·104 cells mL−1 showed similar attachment efficiency (58 ± 11 %, in average) defined as the relationship between the number of attached (day 1) and suspended cells (day 0) in percentage terms. Below 1·104 cells mL−1, the large variability between grid reactors impede obtaining reliable attachment efficiency values. The 24 h coverage values kept concordance with the inoculated and attached cell concentration, varying from 18 ± 4%Cov to 1 ± 2%Cov for a 5·105 cells mL−1 and a 5·103 cells mL−1 concentration, respectively. Figure 4 shows the variation of rescaled %Cov, calculated as indicated above, with time, for the 5 inoculated cell concentrations.
Fig. 4

(a) Representation of the variation of the rescaled percentage of coverage (%Cov) with time for five different initial VSMC inoculum concentrations. (b) Plot comparing the %Cov for the five inoculum concentrations after 116 h of inoculation. Error bars represent standard deviation. *p < 0.05, ***p < 0.001 compared to 5·105 cell mL−1 at 116 h

In all cases, the rescaled %Cov increased with time (Fig. 4(a)) showing a positive correlation with the inoculated concentration value. At each time, significant differences in the coverage magnitude were found between the cell concentrations analyzed, as exemplified in Fig. 4(b) (116 h after inoculation). However, the proliferation rate, analyzed in terms of DCT, did not change with the initial concentration (no significant differences were obtained) with an average magnitude of 32 ± 6 h. These DCT values were in concordance with those previously described in our study. According to the experimental data, only the %Cov magnitude at any time was found to depend on the initial inoculum concentration. Anti-proliferative drugs should stop cell growth (and consequently the surface coverage). Therefore, 5·105 cell mL−1 concentration was selected as the optimal initial seeding concentration for showing the maximum %Cov variation in the lapse time selected for the anti-proliferative- response analysis. From this point, all experiments were performed inoculating 5·105 cell mL−1.
  1. b)

    Evaluation of the FBS concentration

It is well known that FBS supplies to cell cultures the growth factors necessary for the appropriate proliferation. In the case of VSMC, most of studies indicated 10 % FBS as a standard concentration for developing culture in optimal conditions (Martinez-Gonzalez et al. 2008; Redondo et al. 2007). Additionally, some investigations have reported a decreased or increased VSMC proliferation rate in cell media supplemented with 0 or 20 % of FBS, respectively (Stepien et al. 1998). In this section, VSMC growing in grid reactors containing culture media supplemented with 0, 10 or 20 % FBS were investigated following the protocols previously detailed. Figure 5(a) shows the variation of the rescaled %Cov with the time for the three FBS concentrations.
Fig. 5

Effects of FBS on VSMC proliferation in the microbioreactor. (a) Graph showing the variation in rescaled percentage of coverage (%Cov) with time. (b) Representation of the %Cov when the cell media was deprived of FBS (0 % FBS) or supplemented with 10 % and 20 % FBS after 122 h of cells inoculation. Error bars represent standard deviation. ***p < 0.001 compared to 10 % FBS (considered as normal growing conditions)

In comparison with 10 % FBS (considered as normal growing conditions), cell proliferation was found extremely limited when the medium was deprived of FBS (0 % FBS). Statistically, these conditions showed significant differences in both: (1) the rescaled %Cov magnitude at any time (as illustrated in Fig. 5(b)) and (2) the DCT value, that varied from 35 ± 4 h for 10 % FBS to (1.1 ± 0.3)·102 h for 0 % FBS (p < 0.001). This long DCT discarded the use of arrested medium. Analyzing 20 % FBS supplemented cells proliferation, no significant differences were obtained when comparing with 10 % FBS either in the DCT magnitude (31 ± 3 h for 20 % FBS) or in the rescaled %Cov values before 95 h. Although significant differences were observed after 95 h (Fig. 5), the use of 20 % FBS was discarded by considering its effect insufficient since the double FBS concentration only produced a 13 % increase in %Cov magnitude after 122 h (Fig. 5(b)).

3.3 Screening of the anti-proliferative activity of drugs: colchicine and curcumin

Colchicine was selected for being a well known anti-proliferative drug widely used in cell cultures, especially in cancer cell lines (Bhattacharyya et al. 2008). In human carcinoma cell lines, colchicine was found to effectively block cell mitosis at concentrations higher than 0.1 μM (Taylor 1965). Other authors demonstrated that vascular smooth muscle (R22) cells exposed to 10 μM colchicine for 4 h produced a reduction of the cellular footprint area of 35 % (Samarakoon and Higgins 2002). Nowadays short exposure times to colchicine concentrations above 50 μM are commonly used for synchronization of cell cultures (Ivanov et al. 2004). Accordingly, in this study, colchicine concentrations ranging from 0.05 to 50 μM were analyzed in VSMC. The variation of the rescaled %Cov with the incubation time for each colchicine concentration is represented in Fig. 6(a).
Fig. 6

Screening of the effect of colchicine in VSMC proliferation in the microbioreactor. (a) Graph showing rescaled percentages of coverage (%Cov) of VSMC with the time in control conditions (untreated media), in the presence of the vehicle of the drug (20 % PBS) and in the presence of different concentrations of colchicine (0.05–50 μM). Error bars represent standard deviation. (b) Images of cultured VSMC at three different times (in hours) in control conditions and in the presence of colchicine (0.05 μM and 25 μM)

Considering %Cov, colchicine activity was found to depend on the concentration. Three different activities were identified depending on the colchicine concentration. Below 0.05 μM, only a moderate reduction of cell proliferation was obtained, even after long term treatments [see 0.05 μM concentration; significant differences with the control were only found at 102 h (p < 0.05) and at 120 h (p < 0.01) after inoculation]. Between 0.5 and 25 μM, cell proliferation was completely blocked and the %Cov magnitude was always close to 0. Finally, the treatment with high doses (above 50 μM) resulted in data points that felt below 0%Cov, indicating cell detachment and death (cytotoxicity). Similar results were found when analyzing the DCT: the DCT magnitude varied from 47 ± 2 h to (0.9 ± 0.3)·102 h when increasing the colchicine concentration from 0.05 to 25 μM, being significantly different to the control in all the cases. The vehicle solution was also tested and no significant differences were observed along the experiment regarding to %Cov or DCT. Colchicine also acted at morphological level: although VSMC kept the characteristic spindle-like cell structure and the hill and valley phenotype at low colchicine concentrations (0.05 μM), increasing drug concentrations progressively deteriorated the normal cellular shape reducing the spread cell area (Fig. 6(b)). That observation was also in concordance with previous results (Samarakoon and Higgins 2002).

The anti-proliferative activity of curcumin in VSMC was also evaluated. A concentration range from 10 to 75 μM was investigated, according to the literature (Pae et al. 2007; Qin et al. 2009; Yu and Lin 2010). Figure 7 illustrates the effect of different curcumin doses on the rescaled %Cov of VSMC with time.
Fig. 7

Screening of the effect of curcumin in VSMC proliferation in the microbioreactor. (a) Graph showing rescaled percentages of coverage (%Cov) of VSMC with the time in control conditions (untreated media), in the presence of the vehicle of the drug (0.75 % DMSO) and in the presence of different concentrations of curcumin (10–75 μM). Error bars represent standard deviation. (b) Images of cultured VSMC at three different times (in hours) in control conditions and in the presence of curcumin (10 μM and 50 μM)

As previously reported (Yu and Lin 2010), curcumin was found to have a concentration-dependent anti-proliferative activity in VSMC: although 10 μM curcumin did not modify the growth pattern of VSMC under control circumstances (no significant differences were observed either in the rescaled %Cov or in the DCT value), above 25 μM, curcumin evoked a significant and progressive reduction of the proliferation until 75 μM, where a deep cytotoxic effect with cell detachment and death was observed (Fig. 7(a)). Significant differences were observed in the DCT magnitude for 50 μM (p < 0.05) and for 75 μM (p < 0.001). The vehicle solution, 0.75 % DMSO, did not show significant differences with the control in terms of either %Cov or DCT. Morphological differences considering cellular shape, spread cell area and growth phenotype were also observed when increasing curcumin concentration (Fig. 7(b)).

The anti-proliferative activity of both drugs was compared. For this, the inhibitory concentration 50 (IC50), corresponding to the concentration capable of reducing 50 % of the cell proliferation, and the minimum toxic concentration (MTC), in this case the minimum dose that reduced the rescaled %Cov below 0, were determined. Figure 8 illustrates the variation of the rescaled %Cov with the drug dose after 116 h of treatment.
Fig. 8

Representation of the variation of the rescaled %Cov with the colchicine and curcumin concentration after 116 h of treatment. Inset, a magnification of the 0 to 0.6 μM colchicine concentration range is included. Also the IC50 and the MTC values for both drugs are shown

According to the data, colchicine presented a much higher anti-proliferative (IC50 = 3.2 ± 1.2 μM) and cytotoxic activity (MTC = 39.5 ± 0.2 μM) than curcumin (IC50 = 35 ± 5 μM, MTC = 71 ± 2 μM).

3.4 Single cell analysis

When considering the case of single cells, the readout-grid structure integrated in the system allowed the identification, localization and analysis of individual cells at different experimental conditions or under the effect of drugs. Figure 9 illustrates the morphology and proliferation pattern of individual cells growing under optimal experimental conditions, serum deprivation and in presence of colchicine (25 μM) or curcumin (50 μM).
Fig. 9

Monitoring of a single VSMC proliferation in the microfluidic device during three days from initial inoculums under control conditions, serum deprived or in the presence of colchicine (25 μM) or curcumin (50 μM)

VSMCs growing under optimal conditions (control) showed the typical spindle-like morphology, high motility and duplication times around 48 h. In serum deprived culture mediums, although the size and the morphology did not vary from the ideal situation, cell motility and proliferation were critically limited. For instance, the duplication time was almost doubled. Finally, the presence of well-known anti-proliferative drugs (colchicine or curcumin) in a cytotoxic dose, stopped cell proliferation also producing morphological changes. Size reduction, motility decrease and the lost of the typical spindle-like morphology were the most relevant changes observed.

4 Conclusions

This paper describes the development of cell-based readout grid-integrated parallel microbioreactors for the screening of the anti-proliferative activity of drugs in VSMC. The microfluidic device, fabricated using a two-level SU-8 mask, did not modify the proliferation of VSMC in a hill and valley phenotype. The integrated grid array was probed capable of identifying and monitoring the proliferation pattern (motility, size and morphology) of individual (single cell) or groups of cells. In relation to the key role of VSMC proliferation in the atherosclerotic process, the microfluidic device was found extremely useful in the evaluation of the anti-proliferative activity and cytotoxicity of drugs. In this case, colchicine resulted much more anti-proliferative and cytotoxic than curcumin to VSMC. Therefore, this microbioreactor represents a simple, automated and reliable alternative with little manipulation requirements for the anti-proliferative and cytotoxicity analysis of drugs in mammalian cell cultures.

Future work focused on integrating optical transduction in the microfluidic device for the real time monitoring of cell proliferation will be performed.


The research leading to these results has received funding from the European Research Council under the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC grant agreement n° 209243. The authors would like to acknowledge the Spanish Ministry of Economy and Competitivity for the award of a Ramón y Cajal contract and the German Research Foundation (DFG) for supporting this work in the framework of the Collaborative Research Group mikroPART FOR 856 (Microsystems for particulate life-science products). This work was also supported by funds from the Ministerio de Ciencia e Innovación (AGL2009-11559) and Consejería de Innovación, Ciencia y Empresa de la Junta de Andalucía (PAIDI, CTS178), Spain. Cell culture was performed in the Biology Service of the Centro de Investigación, Tecnología e Innovación of the University of Seville (CITIUS). Authors want to acknowledge Dr. Modesto Carballo (CITIUS) for scientific support and facilities.

Copyright information

© Springer Science+Business Media, LLC 2012