Using a Novel 3D Perfusion Bioreactor to Culture ß-Actin-RFP Reporter Osteosarcoma

By: Carlos E. Caicedo-Carvajal, Ph.D.; Qing Liu, Ph.D., Biowire Spring 2012, 25–28

Biowire Spring 2012 — Live Cell Imaging of Signaling Pathways

Qing Liu, Ph.D., is an expert in biomaterials, tissue engineering, and in vitro tissue model for drug screening. Carlos E. Caicedo-Carvajal, Ph.D., has worked in several areas of Biomedical Engineering, including 3D tissue organization and engineering of 3D in vitro models for tissue regeneration and drug screening.


To efficiently expand cells and develop robust cell-based models for in vitro drug screening, in vivo-like cell culture conditions — such as dynamic perfusion and 3D growth — are required. Here, we show results on a study in which we cultured CompoZr® Zinc Finger Nuclease-modified U2OS osteosarcoma cells under 3D perfusion cell culture conditions. CompoZr Zinc Finger Nuclease (ZFN) technology was used to insert the gene encoding red fluorescent protein (RFP) into U2OS’s genome to report the presence of β-actin (U2OS RFPACTB). These cells regularly show an epithelial-like morphology when grown on 2-dimensional (2D) tissue culture plates. However, U2OS RFP-ACTB cells maintained in 3D perfusion bioreactor cell culture conditions showed different cell morphology, enhanced β-actin-RFP expression, cell proliferation, and collagen deposition when compared to U2OS RFP-ACTB growing in 3D static cell culture conditions. Therefore, the 3D perfusion bioreactor is a better platform for in vitro cell culture and cell-based models.


Current cell-based assays for drug screening are mostly based on oversimplified cell culture platforms. For instance, both standardized cell lines and primary cells are grown in 2D dishes under static conditions, creating a simplified cellular model without the complexity of target organs or tissue1. A recent study that compared a 3D in vitro model, a mouse model, and a traditional 2D model showed the 3D model has a better genotypic correlation to spontaneous tumor formation found in vivo2. Besides 3D cell culture conditions, perfusion is another factor of interest for drug screening models3. 3D static growth is sometimes not good enough because static cultures deliver suboptimal nutrition/waste exchange conditions that may limit cell growth. There is, therefore, a need to develop a 3D dynamic perfusion cell-based platform for growing optimal cell cultures.

As the first step towards developing an in vivo-relevant, cell-based, 3D perfusion cell culture platform, we have jointly developed a method to maintain and amplify ZFN-modified U2OS osteosarcoma reporter cell lines with 3D Biotek. Using ZFN technology, a reporter gene coding for red fluorescent protein (RFP) can be integrated into the locus encoding β-actin, a cytoskeletal protein expressed during cancer progression and metastasis4.

The purpose of this study is to show that 3D perfusion cell culture conditions provide an optimal cell culture platform for cell expansion and in vitro disease model creation.

Material and Methods

  1. Instrument: 3D Perfusion Bioreactor (Cat. No. Z687502), 3D Biotek, New Jersey, USA.
  2. Fluorescent Reporter Cell Line: ZFN-modified U2OS Osteosarcoma expressing RFP-labeled ß-actin (Cat. No. CLL1035).
  3. 24-well 3D Insert™ Poly-(ε)-caprolactone µ(PCL) (3030) (300 µm fiber diameter and 300 µm fiber-to-fiber spacing), 3D Biotek, New Jersey, USA.
  4. Cell proliferation was determined using the DNA Quantitation Kit, Fluorescent Assay (Cat. No. DNAQF).
  5. 3.8% formaldehyde-fixed 3D Insert™ PCL scaffolds were stained using 0.05% Direct Red 80 dissolved in picric acid to detect collagen deposition. Picric Acid (Cat. No. 239801) and Direct Red 80 (Cat. No. 365548).
  6. Fluorescent images of adherent U2OS RFP-ACTB were taken with a Nikon Eclipse TS100 adapted with an epi-illuminator (Lumen 200) and filter chromarhodamine (excitation D540/25x, emission D605/55m), NextDayScience, Inc., Finkburg, MD.
  7. Data processing was done using Prisms 4.0 (2005), GraphPad Software, Inc.
  8. Cell seeding in 3D Insert PCL scaffolds for static and perfusion cultures: 5 x 105 U2OS RFP-ACTB cells in 130 µl were seeded on PCL scaffolds (Figure 1A, 1B) for static and perfusion cultures, respectively. After seeding, the initial 130 µl cell suspension was allowed to infiltrate the scaffold in plates (3D Static) and inside the perfusion chambers (3D Dynamic) for 3 hours in the incubator (37 °C, 5% CO2, and 95% humidity). After 3 hours, 370 µl of medium were added to the static culture (37 °C, 5% CO2 , and 95% humidity). The bioreactor was set at 2.0 rpm (33 µm/sec) for 24 hours and 2.5 rpm (43 µm/sec) for long-term cell culture in an incubator (37 °C, 5% CO2, and < 20% humidity).
3D Perfusion and Static Cell Seeding

Figure 1. 3D Perfusion and Static Cell Seeding. In static seeding, a cell suspension aliquot is placed carefully in the center of the scaffold, followed by 3 hours of incubation. In dynamic seeding, a cell suspension aliquot is placed in the center of the scaffold, and each scaffold is stacked with an O-ring separating each scaffold. The loaded chamber requires 3 hours of static incubation to allow the cell suspension to infiltrate the scaffold (same condition as 3D static). After 3 hours, the 3D perfusion system is assembled, and medium perfusion is set at 1.9 rpm for 24 hours. For regular static, a volume of medium is added for regular culture into each well containing a scaffold.

Results and Discussion

U2OS RFP-ACTB cells cultured as a monolayer on tissue culture plates have a cobblestone-like morphology with significant basal cytoskeletal stress fiber formation (Figure 2A). It has been extensively documented that in vivo-like conditions such as 3D geometry, surface roughness, and medium perfusion can exert significant morphological changes on cells, giving rise to more relevant in vivo characteristics. U2OS RFP-ACTB cells growing under 3D perfusion conditions showed a more spindle-like morphology and recurrent cytoplasmic projections (Figure 2B, 2C, white arrows). In addition, osteosarcoma cells grew away from the PCL fiber surface as adherent cell agglomerates (Figure 2D, white arrows). The complex cell morphologies seen in 3D Perfusion are the direct effect of 3D perfusion culture conditions. Besides cell morphology differences, cells grown in 3D static and 3D perfusion culture conditions showed marked differences in proliferation, ß-actin RFP expression, and collagen deposition. Even though cell proliferation plateaued for both 3D static and 3D perfusion at later times, there was a significant difference in cell proliferation as early as day 5. At day 5, the 3D perfusion culture was 2.8-fold higher than the 3D static culture (Figure 3, p < 0.0001). The enhanced proliferation seen in 3D perfusion conditions at days 5 and 10 may be explained as the result of enhanced nutrient/waste transport.

U2OS Osteosarcoma Growing on 2D and 3D Perfusion

Figure 2. U2OS Osteosarcoma Growing on 2D and 3D Perfusion. U2OS RFP-ACTB adopts a cobblestone morphology when grown in regular 2D plates under static conditions A). The cells have extensive stress fibers as a result of their basal adhesion to the culture plate. U2OS RFP-ACTB adopts different cell morphologies when grown under 3D perfusion cell culture conditions. Cells are more spindly, as seen by the RFP outlined extending cell processes (B, C, white arrows). In addition, aggregation or cell masses are seen forming away from the surface of the polymer struts (D, white arrows) (bar = 150 mm).

Proliferation of Osteosarcoma in 3D Static and Perfusion Conditions

Figure 3. Proliferation of Osteosarcoma in 3D Static and Perfusion Conditions. U2OS RFP-ACTB DNA content was measured in scaffolds maintained in both 3D static and 3D perfusion scaffolds. The 3D perfusion scaffolds caused significant differences in cell proliferation as early as day 5 in culture, with a DNA content of 12,760 +/-217.2 ng (+/-SEM, n = 4), as opposed to 3D static with a value of 4,517 +/-162.8 ng (+/-SEM, n = 4) unpaired t-Test (p < 0.0001). There was no significant difference in proliferation at later times for each cell culture condition.

One of the advantages of U2OS RFP-ACTB cells is their adhesion may be monitored using the RFP-tagged β-actin signal in real time. Quantification of the normalized cumulative RFP fluorescence was done in the center of the 3D Insert PCL scaffolds. RFP signal was better in the center of the scaffolds than at the periphery under static culture conditions. However, this was not the case for cells cultured in scaffolds maintained under perfusion conditions, as RFP fluorescence was homogeneous throughout the scaffolds. In terms of RFP expression, there was a statistically significant difference of about 2.7-fold between the 3D perfusion culture and the 3D static culture as early as day 5 (Figure 4, p < 0.0001). Control PCL Insert™ cultured in medium was used to normalize for background noise. The fold difference was calculated from the cumulative fluorescence of 4% of the insert’s area. Thus, the fold differences in β-actin expression between the 3D perfusion culture and 3D static culture could be far greater if the total surface area of the 3D insert were taken into account. Thus, enhanced β-actin RFP may be representative of enhanced cell proliferation and increased adhesion of U2OS towards deposited extracellular matrix (ECM).

The Effect of 3D Perfusion on beta-Actin RFP Expression

Figure 4. The Effect of 3D Perfusion on β-Actin RFP Expression. U2OS RFPACTB β-actin RFP expression was quantified from fluorescent images of scaffolds maintained under 3D static and 3D perfusion cell culture conditions. The 3D perfusion culture caused significant differences in β-actin RFP expression as early as day 5 in culture, with an RFU value of 163,200 +/-10,570 RFU (+/-SEM, n = 7), as opposed to the 3D static culture with a value of 62,020 +/-2450 RFU (+/-SEM, n = 7) unpaired t-Test (p < 0.0001). There was no significant difference in RFP RFU at later times for each cell culture condition. The top inserts are representative fluorescent photographs matched to the average RFU value for their respective cell culture conditions at days 5 and 10.

Even though osteosarcoma is an aberrant form of normal bone cells, they still secrete collagen5. In this study, it was found that U2OS RFP-ACTB growing in 3D Insert PCL scaffolds under both static and perfusion conditions showed a positive stain for collagen deposition. However, there were striking visual differences in the total area of staining (Figure 5). At day 5, the collagen deposition in the static culture was restricted to the center of the scaffold, correlating with the localized/minimal expression of the β-actin RFP signal. In contrast, 3D-perfusion-cultured U2OS RFP-ACTB showed homogenous collagen distribution throughout the scaffolds, correlating with the homogenous β-actin RFP fluorescence seen throughout the 3D Insert.

Based on the results obtained in the presented study, we concluded that U2OS RFP ACTB osteosarcoma cells grew more efficiently under 3D perfusion cell culture conditions.

3D Perfusion Enhances Cell Adhesion and Collagen Deposition

Figure 5. 3D Perfusion Enhances Cell Adhesion and Collagen Deposition. U2OS RFP-ACTB had differences in collagen matrix deposition. 3D PCL scaffolds were stained with Direct Red 80 to determine the presence of collagen. The stain coverage was complete in scaffolds cultured under 3D perfusion as early as 5 days in culture. However, scaffolds cultured under 3D static conditions had less staining coverage than 3D perfusion. The pattern of collagen deposition correlated visually with the pattern of RFP fluorescence seen in the scaffolds — i.e., at the center for the 3D static culture, and at the center and the periphery for the 3D perfusion culture. (bar = 150 mm; 3D PCL scaffold radius = 7.4 mm).




  1. Nyga A, et al. 3D tumour models: novel in vitro approaches to cancer studies. J Cell Commun Signal. 2011;5(3):239–48.
  2. Ridky TW , et al. Invasive three-dimensional organotypic neoplasia from multiple normal human epithelia. Nat Med. 2010;16(12):1450–5.
  3. Tostoes RM, et al. Perfusion of 3D encapsulated hepatocytes — a synergistic effect enhancing long-term functionality in bioreactors. Biotechnol Bioeng. 2011;108(1):41–9.
  4. Gu W, et al. Regulation of local expression of cell adhesion and motility-related mRNAs in breast cancer cells by IMP1/ZBP1. J Cell Sci. 2012;125:81–91.
  5. Kostenuik PJ, et al. Bone cell matrix promotes the adhesion of human prostatic carcinoma cells via the alpha 2 beta 1 integrin. Clin Exp Metastasis. 1996;14(1):19–26.


Related Links