Scaffold-free and label-free biofabrication technology using levitational assembly in a high magnetic field

The human chondrosarcoma cell line SW1353 was purchased from the American Type Culture Collection (ATCC) (Cat. # HTB-94) and cultured in DMEM/F12 medium with 10% fetal bovine serum (FBS) in humidified atmosphere 5% CO₂ at 37 °C. Cells were routinely split at 85%–95% confluence using mild enzymatic dissociation with a 0.25% trypsin/0.53 mM EDTA solution (all reagents from Gibco, USA).

Tissue spheroids were prepared using 81-well MicroTissues 3D Petri dish micromolds (Sigma-Aldrich, USA) according to the manufacturer's protocol. Briefly, 190 μl of SW1353 cell suspension with concentration 3.4 × 10⁶ cells per ml was placed into each agarose mold, incubated for 1 h and then covered with complete growth media. The resultant tissue spheroids consisted of 8 × 10³ cells.

The viability of tissue spheroids in the presence of paramagnetic gadobutrol ('Gadovist', Bayer Pharma AG, Germany) was assessed using the CellTiter-Glo 3D kit (Promega, USA). Thus, four-day-old spheroids were placed in 0, 0.8, 50 and 100 mM gadobutrol for 24 h. The reagent was added for 30 min, and then luminescence was recorded using a VICTOR X3 multilabel plate reader (Perkin Elmer, USA). The influence of a magnetic field with 19 T intensity in the presence of an effective gadobutrol concentration (0.8 mM) on cell viability was analyzed using the Live/Dead Cell Double Staining Kit (Sigma-Aldrich, USA). After the exposure in an experimental setup, the spheroids were incubated with calcein AM and propidium iodide at 37 °C for 1 h. After washing with PBS, the spheroids were imaged by fluorescent microscopy (Nikon Eclipse Ti-E, Japan).

A spheroid fusion assay was performed using ultralow-adhesive spheroid microplates (Corning, USA) (Corning, cat.# 4520). Pairs of tissue spheroids were placed in contact in the presence of 0, 0.8 and 100 mM gadobutrol concentrations and incubated for five days. Bright-field images of spheroid doublets were obtained at the time points 0, 1, 2, 3, 4, 6, 8, 12 and 24 h. Contact intersphere angles were measured using Image J 1.48v software (NIH, Bethesda, MD, USA) and plotted as a function of time using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA).

Samples were fixed in PBS-buffered 4% paraformaldehyde solution, then put in melted 2% agarose gel and finally embedded in paraffin (Merck, Germany). After dewaxing, serial sections with a thickness of 4 μm were cut and routinely stained with hematoxylin-eosin (Sigma-Aldrich, Germany). For transmission electron microscopy (TEM) examination samples were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer, and post-fixed with 1% osmium tetroxide in the buffer containing 1.5% potassium ferricyanide. Then, the samples were dehydrated in ethanol, and embedded in the Epon resin. Ultrathin sections were obtained using a Leica Ultracut UCT ultramicrotome and mounting on Formvar-coated copper grids. The sections were stained with 2% uranyl acetate in water and lead citrate. The sections were observed in a Tecnai T12 electron microscope equipped with an Eagle 4k × 4k CCD camera (Thermo Fisher Scientific, The Netherlands).

To create a magnetic field, a 50 mm-bore, 31 Tesla Bitter magnet was used (figures 1(a), (b)). To ensure the temperature regime of 37 °C, a water thermostat was connected to the working area of the magnet. To place the cuvette with the material inside the magnet, a custom-designed holder was used (figure 1(c)), which fixed the cuvette at a certain height and used a mirrors and lenses system to allow the observation of the levitation and assembly process. LEDs were fixed to the holder to illuminate the working area. A glass flask with a plastic screw cap was used as a cuvette and filled with culture medium containing polystyrene beads or tissue spheroids with different concentrations of the gadobutrol. A camera was used for registration of the assembly process.

Zoom In Zoom Out Reset image size

The molecular dynamics (MD) modeling was carried out as described previously [11, 21–23]. We assumed that all particles in MD calculations were spherical, with the same size and mass m_p. For modeling dynamics of the diamagnetic polystyrene particles in the magnetic field of the Bitter electromagnet, we have numerically solved the system of the Newton equations for all the particles (1 < k < N):

$$\begin{eqnarray}&&{m}_{p}\displaystyle \frac{{d}^{2}{{\bf{r}}}_{k}}{d{t}^{2}}=\displaystyle \displaystyle \sum _{l}F\left(\left|{{\bf{r}}}_{kl}\right|\right)\displaystyle \frac{{{\bf{r}}}_{kl}}{\left|{{\bf{r}}}_{kl}\right|}+{{\bf{F}}}_{kB}+{{\bf{f}}}_{k}+{{\bf{f}}}_{g}+{{\bf{f}}}_{b}\end{eqnarray} \tag{ 1 }$$

where N is the number of particles for which the simulation was performed, k is the particle serial number in the ensemble, l is the serial numbers of surrounding particles and rk, rl is the position of the center of a particle k and l, respectively, ${{\bf{r}}}_{kl}={{\bf{r}}}_{k}-{{\bf{r}}}_{l}.$ The first term in the right-hand part of equation (1) is the Lennard-Jones interaction with other particles, the second term is the interaction with the magnetic field of the trap $\left|{{\bf{F}}}_{kB}\right|=\left({\chi }_{p}-{\chi }_{s}\right)B{\rm{\nabla }}B$ (figure 1) and ${\chi }_{p},\,{\chi }_{s}$– is the magnetic susceptibility of the particle and solution, respectively. The third term is the force of viscous friction against the continuum liquid which is determined by the Stokes formula ${{\bf{f}}}_{k}=-3\pi d\eta {{\bf{u}}}_{k},$ η is the viscosity of the gadobutrol suspension, d is the particle diameter and u_k = $d{{\bf{r}}}_{k}/dt$ is the particle velocity relative to the stationary liquid, the fourth and the fifth terms are a force of gravity ${{\bf{f}}}_{g}$ and buoyancy force ${{\bf{f}}}_{b}$ acting on the particle. Numerical simulation was performed for the number of particles N = 400.

The simulation was performed for polystyrene beads and tissue spheroids with densities $\rho =1.0405$ g cm⁻³ and $\rho =1.05$ g cm⁻³, respectively. At the first step, particles were distributed randomly within the computational domain with initial zero velocity. The computational domain corresponds to the internal volume of the experimental cylindrical cell—R7 × 25 mm.

To simulate a levitation assembly, we modelled the magnetic field generated by a Bitter magnet (figure 1(a)). The optimal position along the vertical axis of the Bitter magnet was calculated from the measured vertical field profile, following the methodology described by Berry and Geim [4]. The stable zone was determined to be 88 ± 6 mm above the field center (figure 2).

Zoom In Zoom Out Reset image size

For modeling of the assembly process, we carried out MD simulation for the polystyrene beads. MD simulation showed that the polystyrene beads start to cluster after 4 min, achieving a stable assembled construct in 10 min (figures 3(a), (d)). The experimental and simulated shapes of the assembled constructs have a good similarity in the final ellipsoid volume (figures 3(b), (c)). The small difference between experimental and simulated kinetics can be attributed to the initial parameters of the modelling. In simulation, the initial speed of the particles was equal to 0, and the particles were equally distributed throughout. In the experiment, firstly, the particles lay at the bottom of the cuvette; secondly, it took 5 min to get the desired magnetic field intensity. Because of this we moved the cuvette up and down to stir up the particles. As a result, the particles peaked up the initial speed in the different directions, which contributed to the difference between the simulation and experimental results.

Zoom In Zoom Out Reset image size

The polystyrene beads were used as physical analogs of living tissue spheroids with similar sizes and shapes. The levitation assembly of the polystyrene bead construct occurred at a magnetic field intensity of 22 T in the presence of 0.5 mM gadobutrol. The assembly occurred in 10 min (figure 5(a)). The results of the magnetic levitation assembly of the construct from polystyrene beads confirmed developed predictive mathematical models.

In this study we selected the human chondrosarcoma cell line SW1353 as a reliable well-characterized human cell source for fabrication of tissue spheroids [24–26]. The employed fabrication method enabled the formation of tissue spheroids with regular sizes and shapes, which is essential for the standardization of magnetic levitation assembly experiments, as well as for the estimation of the possible toxicity of a high magnetic field and paramagnetic medium (figure 4(a)). The spheroid fusion assay showed that at a low concentration of gadolinium salt (0.8 mM gadobutrol) tissue spheroids can fuse, which confirmed their viability and capacity to migrate from the surface during the tissue fusion process (figure 4(b)). As shown in figure 4(b), the intersphere angle increased as a function of time, and in 20 h, after the beginning of incubation, the intersphere angle increased up to 179° indicating complete spheroid fusion.

Zoom In Zoom Out Reset image size

To quantitatively evaluate the toxicity of gadobutrol on SW1353 tissue spheroids, we used a CellTiter-Glo 3D kit. At a low concentration of gadobutrol (0.8 mM) the tissue spheroids were 100% viable, whereas at a high concentration of gadobutrol (100 mM) viability decreased down to 70% in 24 h (figure 4(c)). Having taken into account the fact that the complete spheroid fusion takes at least 20 h, such cytotoxicity is crucial.

We also performed TEM studies of cells and tissue spheroids incubated at a high concentration of gadobutrol (50 mM). As shown in figure 4(d), in the absence of gadobutrol (control) no ultrastructural changes were observed. TEM showed that 50 mM gadobutrol caused cell injury (figure 4(e)) accompanied by intracellular dystrophy, which correlated with the previous reports [3, 27, 28]. The toxic effect of gadolinium(III)-based gadobutrol at high concentration (>50 mM) on cell ultrastructure usually includes: (i) margination of nuclear chromatin (figure 4(f)), (ii) swelling and reduction in the number of cristae in mitochondria (figure 4(g)), (iii) enlargement of the endoplasmic reticulum (figure 4(h)), as well as (iv) formation of phagosomes (figure 4(i)) and (v) severe vacuolization of the cytoplasm (figure 4(j)). It is worth mentioning that this set of ultrastructural changes is irreversible as it does not improve after 1–2 day incubation in gadolinium-free medium. In the case of a low concentration of gadolinium the ultrastructure remains intact. Taken together the reported observations strongly indicate cell and tissue viability and the absence of ultrastructural changes in cells of tissue spheroids after their incubation with low doses of gadobutrol.

The assembly during levitation of tissue spheroids was carried out at a magnetic field intensity of 19 T and a concentration of 0.8 mM gadobutrol. In contrast to polystyrene beads (figure 5(a)), the process of tissue spheroid assembly (figure 5(b)) took 40 min, which results from some differences in magnetic properties between polystyrene beads and tissue spheroids and the difference in intensities of the magnetic field used. After the spheroids clustered together, the construct was kept under the same levitation conditions for 3 h to ensure fusion of tissue spheroids into a single construct (figure 5(c)). The histology of the construct fragment (figure 5(c), insert circle) demonstrates intact morphology of the tissue spheroids. A series of experiments were performed to show the interdependence between the paramagnetic salt concentration in the medium, the tissue spheroids' or beads' properties and the intensity of the magnetic field to maintain a stable levitation for construct formation. The results show that these parameters can be turned, while the concentration of gadobutrol and the magnetic field applied are directly related despite some small variations between the polystyrene beads' and tissue spheroids' levitation characteristics (figure 5(d)). The beads or spheroids are initially located in the bottom of the cuvette when no magnetic field is applied independently of the gadobutrol used in this study. Upon screening of different magnetic fields and gadobutrol concentrations it was possible to define trend curves that were used to define the threshold values to achieve a levitation state. Figure 5(d) shows the threshold values required to promote polystyrene bead or tissue spheroid levitation, where areas above the curve mean that objects levitate in the vertical direction upwards; the area below the curve means that the objects are not levitating. These data demonstrate that magnetic levitation of both polystyrene beads and tissue spheroids is possible in a high magnetic field at a low non-toxic concentration of gadolinium salt (figure 5(d)).

Zoom In Zoom Out Reset image size

We performed a Live/Dead assay for tissue spheroids after incubation in 0.8 mM gadobutrol in the high magnetic field (19 T). In the low concentration of gadobutrol the high magnetic field provides levitation of tissue spheroids without affecting their viability (figure 5(e)). It was also confirmed by TEM, which did not identify strong ultrastructural changes (figure 5(f)). Taken together, these data strongly indicate the maintenance of the tissue spheroids' viability exposed to the combination of a high magnetic field (19 T) and a low concentration of gadobutrol (0.8 mM). Thus, our theoretical and experimental data demonstrate, for the first time, that magnetic levitation assembly based on tissue spheroid fusion at a non-toxic concentration of gadolinium in a high magnetic field is technologically feasible and can be implemented.