Acta Biomaterialia xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture James Jenkins a,1, Ruslan I. Dmitriev a,⇑,1, Karl Morten b, Kieran W. McDermott c, Dmitri B. Papkovsky a a
School of Biochemistry and Cell Biology, University College Cork, Ireland Nuffield Department of Obstetrics and Gynaecology, University of Oxford, United Kingdom c Department of Anatomy and Neuroscience, University College Cork, Ireland b
a r t i c l e
i n f o
Article history: Received 15 September 2014 Received in revised form 21 January 2015 Accepted 22 January 2015 Available online xxxx Keywords: 3D tissue model Oxygen Phosphorescence quenching microscopy Porous membrane scaffold PLIM
a b s t r a c t Porous membrane scaffolds are widely used materials for three-dimensional cell cultures and tissue models. Additional functional modification of such scaffolds can significantly extend their use and operational performance. Here we describe hybrid microporous polystyrene-based scaffolds impregnated with a phosphorescent O2-sensitive dye PtTFPP, optimized for live cell fluorescence microscopy and imaging of O2 distribution in cultured cells. Modified scaffolds possess high brightness, convenient spectral characteristics (534 nm excitation, 650 nm emission), stable and robust response to pO2 in phosphorescence intensity and lifetime imaging modes (>twofold response over 21/0% O2), such as confocal PLIM. They are suitable for prolonged use under standard culturing conditions without affecting cell viability, and for multi-parametric imaging analysis of cultured cells and tissue samples. We tested the O2 scaffolds with cultured cancer cells (HCT116), multicellular aggregates (PC12) and rat brain slices and showed that they can inform on tissue oxygenation at different depths and cell densities, changes in respiration activity, viability and responses to drug treatment. Using this method multiplexed with staining of dead cells (CellTox Green) and active mitochondria (TMRM), we demonstrated that decreased O2 (20–24 lM) in scaffold corresponds to highest expression of tyrosine hydroxylase in PC12 cells. Such hypoxia is also beneficial for action of hypoxia-specific anti-cancer drug tirapazamine (TPZ). Thus, O2 scaffolds allow for better control of conditions in 3D tissue cultures, and are useful for a broad range of biomaterials and physiological studies. Ó 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction 3D tissue models are increasingly used to study tissue function [1], behavior [2], morphology [3], (patho)physiological and disease states [4], responses to drugs and stimuli [5–7]. Compared to conventional 2D models they more closely mimic the microenvironment and cell–cell interactions in animal tissue in vivo, ensure normal signaling (regulated by cell adhesion molecules), gene Abbreviations: 3D, three dimensional; DIV, days in vitro; DMEM, Dulbecco’s modified Eagle’s medium; ECM, extracellular matrix; FBS, fetal bovine serum; HBSS, Hanks’ balanced salt solution; HIF-1a, hypoxia inducible factor-1a; HS, horse serum; PBS, phosphate buffered saline; PFA, paraformaldehyde; PLIM, phosphorescence lifetime imaging microscopy; PS, Polystyrene; PtTFPP, Pt(II)-tetrakis(pentafluorophenyl)-porphine; ROI, region of interest; ROS, reactive oxygen species; RT, room temperature; SFM, serum free media; STDEV, standard deviation; TBST, Tris buffered saline with Tween 20; TMRM, tetramethylrhodamine methyl ester; TPZ, Tirapazamine; TCSPC, time-correlated single photon counting. ⇑ Corresponding author at: Cavanagh Pharmacy Building, University College Cork, College Road, Cork, Ireland. Tel./fax: +353 21 4901339. E-mail address: [email protected] (R.I. Dmitriev). 1 These authors contributed equally to this work.
expression and post-translational protein modifications. In such systems cell function can be manipulated by the addition of physiologically relevant signaling molecules, growth factors and peptides, while internal gradients of nutrients, growth factors, drugs and O2, mediated by diffusion and cellular demand can be adjusted to resemble those found in vivo [8–10]. 3D tissue models can be categorized into scaffold-based and scaffold-free systems. Scaffold-free systems include spheroids [11], organoids and tissue slices and explants [12]. Scaffold-based tissue models allow cultivation of cells in the presence of extracellular matrix (collagen, Matrigels) [13,14], self-assembling peptides [15,16] or synthetic biocompatible polymers [17,18]. Scaffold materials are designed to be porous for efficient gas and mass exchange, possess mechanical strength and solid structures which encourage adhesion and growth of cells, proliferation, differentiation and migration properties, and normal function superior over 2D cell models [19]. One such example is Alvetex™ scaffolds comprising highly porous polystyrene-based membranes having thickness of 200 lm, pore size 36–40 lm and thin walls. These materials have been used in drug screening and toxicity studies
http://dx.doi.org/10.1016/j.actbio.2015.01.032 1742-7061/Ó 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
2
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
[20,21], culturing of brain and spinal cord tissue explants [22,23]. Apart from their mechanical and supporting function, 3D scaffolds may have additional functionalities providing controlled drug release, quantitative measurement of mechanical stress, specific analytes and others parameters [24], e.g., agarose gel with controlled pH affecting metabolic function [25] of chondrocytes or growth factor-release system for cardio-spheres [26]. On the other hand, high thickness and low transparency of such materials limit the use of conventional fluorescence microscopy and access to deep regions during culturing [10]. Use of many fluorescent probes, live cell markers and real-time analysis of viability and physiological processes within scaffolds is difficult. Molecular oxygen (O2) is a key metabolite affecting cell function and viability which must be carefully controlled in 3D tissue models [27,28]. O2 can affect cellular function via energy stress (ATP production by oxidative phosphorylation), generation of ROS (reactive oxygen species), altered activity of metabolic enzymes and signaling pathways [29], transcription, protein production or degradation [30,31]. Low oxygen availability (hypoxia state) can aid progression of cancer, through increased production of specific growth factors (VEGF, angiopoeitin 2), HIF-1a activity [32] and angiogenesis [32–35]. O2 fluctuations are an indicator of stroke, defective metabolism, neurological disorders such as Alzheimer’s disease [36,37], and can influence the efficiency of tumor-, chemoand radiation therapy. The phosphorescence quenching method allows quantitative measurement and imaging of O2 levels using solid-state sensor materials or probes introduced into the vasculature, tissue or cells [22,28,38]. Compared to probe-based methods [22,39], (pre-)staining of scaffold allows for controlled deposition of O2-sensitive indicator dye, higher brightness and provides opportunity to sense extracellular O2 gradients in simple manner. Using this approach O2 gradients in engineered tissue were measured using planar solid-state phosphorescent O2 sensors [40]. Fiber-optic O2 microsensors were used to investigate O2 tension in human dermal fibroblast or human bone marrow derived stromal cells grown on collagen type 1 scaffolds [41]. Scaffolds made of electrospun PLGA fibers with unevenly distributed nanosensors were used for O2 and pH detection [42]. Optical O2 sensors have advantages over the invasive and unstable electrode systems [43], however, the Fibreoptic probes [41], needle-type microsensors [44] and planar sensor membranes only allowing spot measurements or O2 visualization in 2D. 2D and 3D visualization of O2 distribution with sub-micron spatial resolution can be achieved by fluorescence/phosphorescence microscopy, which can also be combined with live cell imaging and multi-parametric analysis [45]. Fabrication of O2-sensitive materials from micro-porous polymeric substrates and their studies by phosphorescence lifetime imaging microscopy (PLIM) were recently demonstrated [46,47]. This approach can also be applied to scaffold materials based on polymers with moderate O2 permeability, such as polystyrene [48], thus enabling imaging analysis of O2 in 3D tissue models and scaffolding systems. Here, we applied the commercial polystyrene-based Alvetex™ membranes to prepare hybrid O2-sensitive scaffold materials to establish 3D tissue models with O2 levels determined using live cell imaging methods. We optimized staining of the scaffolds with the O2-sensitive phosphorescent dye, PtTFPP, evaluated their O2-sensitivity and general usability with several different cell and tissue models. We show that O2-sensitive 3D scaffold can be used in 3D cell cultures in a similar manner to ordinary scaffolds. They also allow measurement of O2 over large surface area without chemical or mechanical stress and the need to stain biological samples with O2 probes. Multi-parametric analysis of tissue samples in fluorescence/phosphorescence imaging modality and microsecond FLIM (PLIM) is also demonstrated, which provides reliable and
accurate quantification and mapping of O2 distribution and its correlation with cell/tissue function. 2. Experimental section 2.1. Materials Polystyrene scaffold membranes Alvetex™ (12-well inserts) were from Reinnervate (Amsbio, UK). PtTFPP dye was from Frontier Scientific (Inochem Ltd., Lancashire, UK, Cat. No. PtT975). ProLong Gold anti-fade reagent, Tethramethylrhodamine methyl ester (TMRM), Alexa Fluor-conjugated secondary antibodies, B27 serum-free supplement were from Invitrogen (Biosciences, Dublin Ireland). Epidermal growth factor (EGF), fibroblast growth factor (FGF), anti-b3-tubulin were from Millipore (Cork, Ireland). Anti-Tyrosine Hydroxylase antibody was from Abcam, Cambridge UK. CellTox Green assay kit was from Promega (MyBio, Ireland). Goat anti-Nestin antibody was from Santa Cruz biotechnology, Heidelberg Germany. Calcein Green AM probe, Tirapazamine, anti-GFAP antibody and all the other chemicals (HPLC or spectrophotometric grade) were from Sigma–Aldrich (Dublin, Ireland). Other plasticware (cell culture grade) was from Sarstedt (Wexford, Ireland). 2.2. Preparation of O2-sensitive 3D scaffolds PtTFPP dye was dissolved at 0.025–0.05 mg/ml in acetone:water (7:3) mixture at room temperature, and 1 ml aliquots of this solution were added to the Alvetex™ scaffold membranes placed individually in wells of a standard 24-well plate. The soaked membranes were incubated for 1 h at 60 °C, then washed sequentially (each time 5 min at 60 °C) with 1 ml of acetone:water (3.5:6.5), acetone:water (1.5:8.5) and lastly with sterile water. The stained ‘‘O2-scaffolds’’ were dried under sterile laminar air flow and stored at room temperature protected from light and contamination. 2.3. Growing cells in the O2 scaffolds Human colorectal carcinoma HCT116 were cultured in McCoy 5A medium supplemented with 10 mM HEPES pH 7.2, 2 mM L-glutamine, 10% FBS and 1% Penicillin–Streptomycin as described previously [22], and then seeded at a density of 100,000 cells per scaffold. Rat pheochromocytoma PC12 cells were cultured in Phenol Red free DMEM (Sigma D5030) supplemented with 10 mM glucose, 1 mM pyruvate, 2 mM L-glutamine, 10 mM HEPES, 1% Penicillin– Streptomycin, 2% B27, 20 ng/ml EGF, 10 ng/ml FGF and 100 ng/ml NGF to produce suspension cells and multi-cellular aggregates as described in [22] and then seeded at a density of 60–100 aggregates per O2 scaffold. Prior to cell seeding, dried O2 scaffolds were pre-soaked in small amount of 70% ethanol and then rinsed 3 times with 1 ml of sterile PBS. After seeding, cells were allowed to grow inside the scaffold for 1–5 days, then counter-stained with Calcein Green AM (1 lM, 30 min) and analyzed on a microscope in Phenol Red free DMEM. Cell viability within the O2 scaffold was assessed by analyzing the membrane integrity (CellTox Green assay, Promega) and number of viable (stained with TMRM) cells, in comparison with control samples (unstained scaffold). Cells were stained with TMRM (20 nM) and CellTox Green (0.1%) for 30 min, washed with fresh media and imaged on the confocal microscope. For immunofluorescence, cells grown in O2 scaffolds were fixed in paraformaldehyde (4%, 10 min, RT) and immunostained with anti-tyrosine hydroxylase antibody (Abcam, ab112) as described
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
3
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
Unstained
B
Stained
3D reconstruction
50
50
60 55 50 45 40 35 30 25 20 15
Intensity, a.u.
1100 Intensity
1000
Lifetime
900 800 700 600 500 400 0
50
100 150 O2 [mM]
D
200
t0/ t-1
1200
Lifetime [ms]
C
YZ
XZ
XY
A
2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0
Intensity Lifetime SternVolmer fit 0
50
100 O2 [mM]
150
200
Fig. 1. Optical properties of O2 scaffolds. (A) Transmission light images of unstained and PtTFPP-stained scaffolds. Scale bar unit is lm. (B) 3D PLIM images of O2 scaffold showing 3D (left) and 2D projections (right). Image was created using 25 Z planes of 2 lm step width (50 lm). (C) O2 calibrations produced on wide-field PLIM microscope in intensity and lifetime scales. (D) Stern–Volmer plots of O2 calibration and fitting of lifetime calibration using two-site model.
previously [39]. Brain slices were fixed in paraformaldehyde solution (4% , 40 min, RT), washed with PBS (0.5 h, 5 times), blocked in 10% FBS supplemented with 0.4% Triton X100 in PBS (2 h, room temperature), incubated with primary antibodies in 2.5% HS, 0.4% Triton X100 in PBS (48 h, 4 °C), washed with TBST (1 h, 5 times), incubated with appropriate Alexa Fluor-conjugated secondary antibodies 1:1000 dilution (18 h, 4 °C), washed with TBST (1 h, 5 times), counter-stained with DAPI (14 lM, 1 h). Samples were then mounted between coverslips in ProLong Gold antifade medium (Invitrogen) and subjected to fluorescence microscopy analysis. In preparation for drug treatment O2 scaffolds were seeded with 60–100 PC12 aggregates per scaffold in SFM and grown for 3 days. 1 mM TPZ (or 1% DMSO in controls) was added simultaneously with 0.1% CellTox green and TMRM (20 nM) to the membranes and incubated for 40 min. The membranes were washed once with 1 ml of Phenol Red free DMEM media containing TMRM. The numbers of dead cells and corresponding phosphorescence lifetimes were then determined and compared to the control samples.
2.4. Brain slice culture All the procedures with animals were performed under a license issued by the Irish Government Department of Health and in accordance with the EU Directive 2010/63/EU. Newborn Sprague-Dawley rats (Biological Services Unit, University College Cork, Ireland) at postnatal day 4 (P4) were used. Brains slices were prepared as described before [22], but without embedding in agarose. Coronal sections (400 and 600 lm thick) were placed on O2 scaffolds in phenol red-free DMEM supplemented with 25% HBSS, 10% FBS, 1% penicillin–streptomycin, 10 mM Glc, 20 mM HEPES-Na, pH 7.2 (‘‘Slice medium’’) and cultured in a humidified incubator at 20% O2, 5% CO2, 37 °C for 1–5 days. Then, they were stained with CellTox Green (0.05%, 3 h), washed in fresh medium and imaged on confocal PLIM microscope as described below.
2.5. Widefield PLIM microscopy The O2 calibration and photobleaching experiments were performed on a wide-field inverted microscope Axiovert 200 (Carl Zeiss) equipped with and oil-immersion objective 40/1.3 EC Plan Neofluar, 390 nm LED excitation module, time-gated CCD camera and emission filter 635–675 nm, ImSpector software (LaVision BioTec, Germany) and integrated CO2/O2 climate control chamber (PeCon) as described before [39]. In PLIM mode the parameters were: snapshot mode, 50 ms exposure, devices delay T. For the intensity imaging the parameters were: exposure time – 2 ms, devices – none, DC mode – ON, no binning. Photobleaching experiments were performed in DC mode with the following settings: exposure time – 2 ms, devices – time–time, numbers of steps – 12, wait time – 10 s. 2.6. Measurement of dye leaching using microplate reader Dye leaching experiments were carried out on a high-sensitivity time-resolved fluorescence reader Victor2 (PerkinElmer). Supernatants from samples with O2 scaffolds were taken periodically and measured 5 times over 10 min in the wells of 96-well microplate under the following settings: mode – time-resolved fluorescence; filters – D340 excitation and D642 emission, delay time – 30 ls, gate time – 100 ls, cycle time – 1000 ls. Mean signal values were then calculated. 2.7. Confocal PLIM microscopy analysis Confocal fluorescence microscopy imaging and PLIM were carried out on a system based on an upright fluorescence microscope Axio Examiner Z1 (Carl Zeiss) with 20/1.0 W-Plan-Apochromat and 40/1.1 W-LD-C-Apochromat objectives, heated stage (37 °C) with motorized Z-axis control, DCS-120 confocal scanner (Becker & Hickl GmbH), R10467U-40 photon counting detector (Hamama-
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
4
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
A
B
C
D
E
F
Fig. 2. Cell viability and growth properties in O2 scaffolds. (A) HCT116 cells in the original (unstained) and O2 scaffolds stained with TMRM (20 nM, Red) and CellTox Green (Green). (B, C) Numbers of dead (CellTox Green) and viable (Calcein/TMRM staining) cells in the scaffolds at different depths. (D) 3D reconstruction of PC12 cells in O2 scaffold. (E, F) Viable (TMRM) and dead (CellTox Green) PC12 cells at different times in culture (2, 5 days). Scale bar unit is lm. Asterisks indicate significance level of P = 0.05.
tsu Photonics K.K.) and TCSPC hardware (Becker & Hickl GmbH) [39]. The probes were excited with tunable picosecond supercontinuum laser SC400-4 (Fianium, UK). The phosphorescence of O2 scaffolds was excited at 534 nm and collected with 635–675 nm bandpass filter, and lifetimes were calculated by mono-exponential fitting of decay curves in SPCImage software (Becker & Hickl GmbH) using the following settings: shift: 0–10, binning: 1–3, T1: 120–140, T2: 230–250. Alexa Fluor 488, CellTox Green and Calcein Green probes were excited at 488 nm with emission collected at 512–536 nm. TMRM – at 540 nm excitation, 565–605 nm emission; Alexa Fluor 594 at 594 nm excitation, 635–675 nm emission. Transmission light images (Fig. 1A) were collected with a D5100 digital SLR camera (Nikon) attached to the microscope. Drug treatments (FCCP, AntA, Tirapazamine) were performed using addition of 10 concentrated stocks in Phenol Red free DMEM at 1/10 of total volume with live samples. For 3D reconstruction samples
were fixed with PFA, immunostained (see Section 2.3) and scanned in Z direction (20–25 optical sections, 2 lm thickness, 40–50 lm total depth). 2.8. Data analysis Fitting of phosphorescence decays was performed either in ImSpector software (widefield PLIM, La vision) or SPCImage software (Becker & Hickl) using single-exponential decay function and pixel binning as appropriate. The resulting 2D matrices with lifetime data were converted to ASCII format and processed in Microsoft Excel to produce O2 concentration values using calibration function. 3D projections of O2 scaffolds were produced from intensity images representing individual optical sections using Volume Viewer plugin (Fiji software (http://fiji.sc/Fiji)), cells stained with CellTox Green and other dyes were counted
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
A
B
C
D
5
Fig. 3. Changes in oxygenation of HCT116 and PC12 cells in O2 scaffolds measured by PLIM. (A) HCT116 cells seeded at low and high densities: images of false-color PLIM superimposed with fluorescence (red). (B) Lifetime distribution at various depths. (C) PLIM and TMRM images of PC12 cells. (D) Lifetimes and oxygen concentrations in the vicinity of cell aggregates at different depths. Scale bar unit is lm. Asterisks indicate significance levels of difference with P = 0.05 (⁄) and P = 0.005 (⁄⁄).
using ’3D objects counter’ plugin in Fiji software (http://fiji.sc/ Fiji). Data are shown as average values with standard deviation as error bars. To ensure consistency of results, all the experiments were performed in triplicate. For statistical significance, data were evaluated using independent t-test, with confidence levels P = 0.05 and P = 0.005, indicated with asterisks in figures. 3. Results and discussion 3.1. Preparation of O2 scaffolds and evaluation of their optical sensor properties Alvetex scaffolds are a convenient material for the fabrication of O2 sensors. Polystyrene (PS) from which they are made provides moderate quenching for Pt(II)-porphyrins and is often used in O2 sensors [48]. The dense network of pores with thin walls allows impregnation by simple swelling in low-polarity solvents and diffusion of dissolved dye molecules into the PS [49]. For the impregnation we choose PtTFPP dye which possesses high photo-stability, brightness and hydrophobicity. We optimized the solvent, dye concentration and temperature, aiming at efficient, fast and uniform staining with the dye and minimal damage of scaffold structure (organic solvent may dissolve PS). Acetone:water mixture (7:3) was found to provide sufficient solubility for PtTFPP and impregnation at 60 °C, while tetrahydrofurane, acetonitrile or ethanol either did damage the polystyrene or did not provide sufficient solubility
for PtTFPP. The 2–3 washing steps at 60 °C with decreased acetone concentration in water successfully removed excess of dye and reinforced uniform staining throughout the volume of scaffolds. Transmission light (Fig. 1A) and confocal phosphorescence (Fig. 1B) images show that the impregnation did not alter significantly the structure of the O2 scaffold and minimally affected its roughness and shape. The size and shape of the pores remained consistent in both stained and unstained samples, and that the dye was distributed evenly across the scaffold and in depth (e.g., 50 lm optical section). The O2 sensitivity of scaffolds impregnated with 0.05 mg/ml PtTFPP was analyzed on a wide-field PLIM microscope with incubator chamber, over the range of air saturation from 21% (equivalent to 209 lM of dissolved O2) to 0%, at 37 °C, in solution. O2 calibration showed typical hyperbolic dependence of the intensity and lifetime signals (Fig. 1C), with approximately twofold reduction at 209 lM, and curved Stern–Volmer plots (Fig. 1D) reflecting non-ideal dye distribution [50]. The following equation to convert the lifetime values in oxygen concentration was obtained, according to two-site model [50]: O2 = (0.71108/( 1 + 0.71108 + s/ 56.5) 1)/0.06414 (R2 = 0.987). Fitting of the lifetime calibration with modified two-site model [50,51] produced Ksv = 0.06414 lM 1 and s0 = 56.5 ls. Confocal images on the PLIM–TCSPC microscope produced similar lifetime values under excitation at 405 nm and 534 nm. Bright phosphorescent signals allow excitation of O2 scaffolds with green light (500–535 nm), which is less damaging to the cells than 405 nm.
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
6
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
A
B
C
D E
F
Fig. 4. Oxygenation and viability of cultured brain tissue slices in O2 scaffolds. Coronal 400–600 lm thick sections of newborn rat brains (P4) were adhered to O2 scaffolds, cultured for 1 and 5 days in vitro (DIV) and analyzed for viability and oxygenation by PLIM. (A) Transmission light images of brain slice with focal points adjusted to the top and bottom (membrane). (B) Fluorescent images of neural cells in O2 scaffolds stained with cholera toxin. (C) PLIM image at the edge of brain slice. (D–F) Distribution of dead cells (CellTox Green) in 400 and 600 lm thick brain slices cultured for 1 and 5 DIV and average lifetime values for corresponding regions (N = 4). Scale bar unit is lm.
We next tested photo-stability, which is important for intensity based and long-term imaging experiments [52,53], by illuminating the O2 scaffolds on the wide-field microscope continuously for 2 min at maximal power of 390 nm LED. Similar to the other PtTFPP based materials [54], practically no photobleaching was observed. To evaluate stability of impregnation, we incubated O2 scaffolds in PBS containing 5% FBS for 7 days at 37 °C measuring leakage of PtTFPP in the supernatant. Scaffold stained with 0.05 mg/ml PtTFPP showed intensity signals of 140 106 cps (counts per second), but only trace amounts (5–7 days). Their brightness, preserved pore architecture and optimal sensitivity to O2 are well suited for quantitative imaging of O2 in phosphorescence lifetime or intensity mode. 3.2. Cell growth and viability in O2 scaffolds We studied whether the phosphorescent scaffold had any effect on cell viability or growth, using two cell models with different growth properties but with strong activity of mitochondria (oxida-
tive respiration) [55]: adherent HCT116 cells forming monolayers and PC12 cells forming multi-cellular aggregates. Since the scaffolds are not transparent and not suitable for transmission light microscopy, we visualized cells by staining with fluorescent dyes Calcein Green (cytoplasm) and TMRM (polarized mitochondria). In unstained and O2 scaffolds, we observed similar numbers and morphology of HCT116 cells, both at the surface and deeper inside (Fig. 2C, Fig. S1). PC12 cell aggregates also penetrated deep into the scaffold (Fig. 2D) and retained viability for at least 5 days (Fig. 2E). The number of dead cells (CellTox Green) was also similar, without statistically significant difference between unstained and stained scaffolds (Fig. 2A, B, E and F). PC12 aggregates showed a significantly (confidence level P = 0.05) higher number of dead cells, residing closer to the surface (10 lm depth) when compared to 40 lm depth, which can be explained by the higher density of cells in these surface regions. Gradual reduction in numbers of dead cells with the depth could also be due to the more optimal physiological conditions, physical and chemical micro-environment inside the scaffold. These results confirm that O2 scaffolds retain 3D architecture and promote cell growth as per unstained scaffolds.
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
7
A
B
C
D
Fig. 5. Stimulation of cells with drugs affecting cellular respiration in O2 scaffolds. (A) Effects of FCCP and Antimycin A. (B) Lifetime changes for selected ROIs close to cell aggregate. (C, D) Changes of lifetime values with increasing distance from cells. Scale bar unit is lm. Asterisks indicate significance levels of difference with P = 0.05 (⁄) and P = 0.005 (⁄⁄).
3.3. Monitoring of oxygenation in different cell cultures Planar O2 sensitive coatings allow monitoring of local O2 levels [56]. Local micro-gradients of O2 formed by actively respiring cells [57] can be also present within O2 scaffold. Using the O2 scaffolds, we tested 3 different models: individual cells (5–10 lm), multicellular aggregates (50–100 lm) and live brain slices (400– 600 lm). In absence of cells the scaffolds showed lifetime values 21.5–23 ls at 21% O2 and minimal (1 ls) variation across the 20 lm depth (Fig. 3A and B). When HCT116 cells were seeded at low density (monolayer cultures), the oxygen distribution remained constant, but at higher numbers (60,000–100,000 cells seeded) following cell penetration into the scaffold lifetime values increased by 2–8 ls, displaying significant lifetime differences when compared to unseeded O2 scaffolds (Fig. 3A and B). No indepth O2 gradients were observed. For multi-cellular PC12 aggregates containing 10–50 respiring cells more profound O2 gradients were expected [22]. In order to facilitate the cell invasion into the O2 scaffold, serum-free media supplemented with neural growth factor (NGF) was used. Indeed, the magnitude of lifetime increase and variation at different depths were higher than for HCT116 cells (Fig. 3C and D). Lifetimes in vicinity of PC12 aggregates ranged from approximately 26 ls at the surface to 36 ls deeper into the scaffold. This corresponds to O2 concentrations 49–16 lM (Fig. 3D). Highest lifetimes were seen at depths of 10–20 lm, and lowest – at the edges of the aggregate (0 and 40 lm). Deeper into the O2 scaffold (20 lm) O2 was significantly (P = 0.005) lower than on the surface (Fig. 3D). This is due to increased cell density and respiration activity in the aggregate core
and limitations for glucose and oxygen diffusion. For control, we stained PC12 cells with intracellular O2 probe PA2 [58] and measured aggregates in unstained Alvetex scaffolds. Similarly, higher O2 levels (lower values of phosphorescence lifetimes) were observed close to the surface, and lowest levels at 10–20 lm depths (Fig. S2). Thus, O2 scaffolds provide similar (or complementary) information about local cell oxygenation as the cell permeable icO2 probes. In cell culture brain tissue slices cells remain under native microenvironment, cytoarchitecture and are also accessible for microscopy imaging [45,59]. As brain tissue is metabolically active and sensitive to O2 fluctuations so to model in vivo physiology experiments O2 levels must be carefully controlled. We prepared 400 and 600 lm-thick brain slices from newborn rats and cultured them on O2 scaffolds for 1–5 days in vitro (DIV). Slices adhered strongly to the scaffolds and some cells migrated into it. Staining with Cholera toxin (neuronal cell tracer) was limited to surface layers
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture James Jenkins a,1, Ruslan I. Dmitriev a,⇑,1, Karl Morten b, Kieran W. McDermott c, Dmitri B. Papkovsky a a
School of Biochemistry and Cell Biology, University College Cork, Ireland Nuffield Department of Obstetrics and Gynaecology, University of Oxford, United Kingdom c Department of Anatomy and Neuroscience, University College Cork, Ireland b
a r t i c l e
i n f o
Article history: Received 15 September 2014 Received in revised form 21 January 2015 Accepted 22 January 2015 Available online xxxx Keywords: 3D tissue model Oxygen Phosphorescence quenching microscopy Porous membrane scaffold PLIM
a b s t r a c t Porous membrane scaffolds are widely used materials for three-dimensional cell cultures and tissue models. Additional functional modification of such scaffolds can significantly extend their use and operational performance. Here we describe hybrid microporous polystyrene-based scaffolds impregnated with a phosphorescent O2-sensitive dye PtTFPP, optimized for live cell fluorescence microscopy and imaging of O2 distribution in cultured cells. Modified scaffolds possess high brightness, convenient spectral characteristics (534 nm excitation, 650 nm emission), stable and robust response to pO2 in phosphorescence intensity and lifetime imaging modes (>twofold response over 21/0% O2), such as confocal PLIM. They are suitable for prolonged use under standard culturing conditions without affecting cell viability, and for multi-parametric imaging analysis of cultured cells and tissue samples. We tested the O2 scaffolds with cultured cancer cells (HCT116), multicellular aggregates (PC12) and rat brain slices and showed that they can inform on tissue oxygenation at different depths and cell densities, changes in respiration activity, viability and responses to drug treatment. Using this method multiplexed with staining of dead cells (CellTox Green) and active mitochondria (TMRM), we demonstrated that decreased O2 (20–24 lM) in scaffold corresponds to highest expression of tyrosine hydroxylase in PC12 cells. Such hypoxia is also beneficial for action of hypoxia-specific anti-cancer drug tirapazamine (TPZ). Thus, O2 scaffolds allow for better control of conditions in 3D tissue cultures, and are useful for a broad range of biomaterials and physiological studies. Ó 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction 3D tissue models are increasingly used to study tissue function [1], behavior [2], morphology [3], (patho)physiological and disease states [4], responses to drugs and stimuli [5–7]. Compared to conventional 2D models they more closely mimic the microenvironment and cell–cell interactions in animal tissue in vivo, ensure normal signaling (regulated by cell adhesion molecules), gene Abbreviations: 3D, three dimensional; DIV, days in vitro; DMEM, Dulbecco’s modified Eagle’s medium; ECM, extracellular matrix; FBS, fetal bovine serum; HBSS, Hanks’ balanced salt solution; HIF-1a, hypoxia inducible factor-1a; HS, horse serum; PBS, phosphate buffered saline; PFA, paraformaldehyde; PLIM, phosphorescence lifetime imaging microscopy; PS, Polystyrene; PtTFPP, Pt(II)-tetrakis(pentafluorophenyl)-porphine; ROI, region of interest; ROS, reactive oxygen species; RT, room temperature; SFM, serum free media; STDEV, standard deviation; TBST, Tris buffered saline with Tween 20; TMRM, tetramethylrhodamine methyl ester; TPZ, Tirapazamine; TCSPC, time-correlated single photon counting. ⇑ Corresponding author at: Cavanagh Pharmacy Building, University College Cork, College Road, Cork, Ireland. Tel./fax: +353 21 4901339. E-mail address: [email protected] (R.I. Dmitriev). 1 These authors contributed equally to this work.
expression and post-translational protein modifications. In such systems cell function can be manipulated by the addition of physiologically relevant signaling molecules, growth factors and peptides, while internal gradients of nutrients, growth factors, drugs and O2, mediated by diffusion and cellular demand can be adjusted to resemble those found in vivo [8–10]. 3D tissue models can be categorized into scaffold-based and scaffold-free systems. Scaffold-free systems include spheroids [11], organoids and tissue slices and explants [12]. Scaffold-based tissue models allow cultivation of cells in the presence of extracellular matrix (collagen, Matrigels) [13,14], self-assembling peptides [15,16] or synthetic biocompatible polymers [17,18]. Scaffold materials are designed to be porous for efficient gas and mass exchange, possess mechanical strength and solid structures which encourage adhesion and growth of cells, proliferation, differentiation and migration properties, and normal function superior over 2D cell models [19]. One such example is Alvetex™ scaffolds comprising highly porous polystyrene-based membranes having thickness of 200 lm, pore size 36–40 lm and thin walls. These materials have been used in drug screening and toxicity studies
http://dx.doi.org/10.1016/j.actbio.2015.01.032 1742-7061/Ó 2015 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
2
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
[20,21], culturing of brain and spinal cord tissue explants [22,23]. Apart from their mechanical and supporting function, 3D scaffolds may have additional functionalities providing controlled drug release, quantitative measurement of mechanical stress, specific analytes and others parameters [24], e.g., agarose gel with controlled pH affecting metabolic function [25] of chondrocytes or growth factor-release system for cardio-spheres [26]. On the other hand, high thickness and low transparency of such materials limit the use of conventional fluorescence microscopy and access to deep regions during culturing [10]. Use of many fluorescent probes, live cell markers and real-time analysis of viability and physiological processes within scaffolds is difficult. Molecular oxygen (O2) is a key metabolite affecting cell function and viability which must be carefully controlled in 3D tissue models [27,28]. O2 can affect cellular function via energy stress (ATP production by oxidative phosphorylation), generation of ROS (reactive oxygen species), altered activity of metabolic enzymes and signaling pathways [29], transcription, protein production or degradation [30,31]. Low oxygen availability (hypoxia state) can aid progression of cancer, through increased production of specific growth factors (VEGF, angiopoeitin 2), HIF-1a activity [32] and angiogenesis [32–35]. O2 fluctuations are an indicator of stroke, defective metabolism, neurological disorders such as Alzheimer’s disease [36,37], and can influence the efficiency of tumor-, chemoand radiation therapy. The phosphorescence quenching method allows quantitative measurement and imaging of O2 levels using solid-state sensor materials or probes introduced into the vasculature, tissue or cells [22,28,38]. Compared to probe-based methods [22,39], (pre-)staining of scaffold allows for controlled deposition of O2-sensitive indicator dye, higher brightness and provides opportunity to sense extracellular O2 gradients in simple manner. Using this approach O2 gradients in engineered tissue were measured using planar solid-state phosphorescent O2 sensors [40]. Fiber-optic O2 microsensors were used to investigate O2 tension in human dermal fibroblast or human bone marrow derived stromal cells grown on collagen type 1 scaffolds [41]. Scaffolds made of electrospun PLGA fibers with unevenly distributed nanosensors were used for O2 and pH detection [42]. Optical O2 sensors have advantages over the invasive and unstable electrode systems [43], however, the Fibreoptic probes [41], needle-type microsensors [44] and planar sensor membranes only allowing spot measurements or O2 visualization in 2D. 2D and 3D visualization of O2 distribution with sub-micron spatial resolution can be achieved by fluorescence/phosphorescence microscopy, which can also be combined with live cell imaging and multi-parametric analysis [45]. Fabrication of O2-sensitive materials from micro-porous polymeric substrates and their studies by phosphorescence lifetime imaging microscopy (PLIM) were recently demonstrated [46,47]. This approach can also be applied to scaffold materials based on polymers with moderate O2 permeability, such as polystyrene [48], thus enabling imaging analysis of O2 in 3D tissue models and scaffolding systems. Here, we applied the commercial polystyrene-based Alvetex™ membranes to prepare hybrid O2-sensitive scaffold materials to establish 3D tissue models with O2 levels determined using live cell imaging methods. We optimized staining of the scaffolds with the O2-sensitive phosphorescent dye, PtTFPP, evaluated their O2-sensitivity and general usability with several different cell and tissue models. We show that O2-sensitive 3D scaffold can be used in 3D cell cultures in a similar manner to ordinary scaffolds. They also allow measurement of O2 over large surface area without chemical or mechanical stress and the need to stain biological samples with O2 probes. Multi-parametric analysis of tissue samples in fluorescence/phosphorescence imaging modality and microsecond FLIM (PLIM) is also demonstrated, which provides reliable and
accurate quantification and mapping of O2 distribution and its correlation with cell/tissue function. 2. Experimental section 2.1. Materials Polystyrene scaffold membranes Alvetex™ (12-well inserts) were from Reinnervate (Amsbio, UK). PtTFPP dye was from Frontier Scientific (Inochem Ltd., Lancashire, UK, Cat. No. PtT975). ProLong Gold anti-fade reagent, Tethramethylrhodamine methyl ester (TMRM), Alexa Fluor-conjugated secondary antibodies, B27 serum-free supplement were from Invitrogen (Biosciences, Dublin Ireland). Epidermal growth factor (EGF), fibroblast growth factor (FGF), anti-b3-tubulin were from Millipore (Cork, Ireland). Anti-Tyrosine Hydroxylase antibody was from Abcam, Cambridge UK. CellTox Green assay kit was from Promega (MyBio, Ireland). Goat anti-Nestin antibody was from Santa Cruz biotechnology, Heidelberg Germany. Calcein Green AM probe, Tirapazamine, anti-GFAP antibody and all the other chemicals (HPLC or spectrophotometric grade) were from Sigma–Aldrich (Dublin, Ireland). Other plasticware (cell culture grade) was from Sarstedt (Wexford, Ireland). 2.2. Preparation of O2-sensitive 3D scaffolds PtTFPP dye was dissolved at 0.025–0.05 mg/ml in acetone:water (7:3) mixture at room temperature, and 1 ml aliquots of this solution were added to the Alvetex™ scaffold membranes placed individually in wells of a standard 24-well plate. The soaked membranes were incubated for 1 h at 60 °C, then washed sequentially (each time 5 min at 60 °C) with 1 ml of acetone:water (3.5:6.5), acetone:water (1.5:8.5) and lastly with sterile water. The stained ‘‘O2-scaffolds’’ were dried under sterile laminar air flow and stored at room temperature protected from light and contamination. 2.3. Growing cells in the O2 scaffolds Human colorectal carcinoma HCT116 were cultured in McCoy 5A medium supplemented with 10 mM HEPES pH 7.2, 2 mM L-glutamine, 10% FBS and 1% Penicillin–Streptomycin as described previously [22], and then seeded at a density of 100,000 cells per scaffold. Rat pheochromocytoma PC12 cells were cultured in Phenol Red free DMEM (Sigma D5030) supplemented with 10 mM glucose, 1 mM pyruvate, 2 mM L-glutamine, 10 mM HEPES, 1% Penicillin– Streptomycin, 2% B27, 20 ng/ml EGF, 10 ng/ml FGF and 100 ng/ml NGF to produce suspension cells and multi-cellular aggregates as described in [22] and then seeded at a density of 60–100 aggregates per O2 scaffold. Prior to cell seeding, dried O2 scaffolds were pre-soaked in small amount of 70% ethanol and then rinsed 3 times with 1 ml of sterile PBS. After seeding, cells were allowed to grow inside the scaffold for 1–5 days, then counter-stained with Calcein Green AM (1 lM, 30 min) and analyzed on a microscope in Phenol Red free DMEM. Cell viability within the O2 scaffold was assessed by analyzing the membrane integrity (CellTox Green assay, Promega) and number of viable (stained with TMRM) cells, in comparison with control samples (unstained scaffold). Cells were stained with TMRM (20 nM) and CellTox Green (0.1%) for 30 min, washed with fresh media and imaged on the confocal microscope. For immunofluorescence, cells grown in O2 scaffolds were fixed in paraformaldehyde (4%, 10 min, RT) and immunostained with anti-tyrosine hydroxylase antibody (Abcam, ab112) as described
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
3
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
Unstained
B
Stained
3D reconstruction
50
50
60 55 50 45 40 35 30 25 20 15
Intensity, a.u.
1100 Intensity
1000
Lifetime
900 800 700 600 500 400 0
50
100 150 O2 [mM]
D
200
t0/ t-1
1200
Lifetime [ms]
C
YZ
XZ
XY
A
2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0
Intensity Lifetime SternVolmer fit 0
50
100 O2 [mM]
150
200
Fig. 1. Optical properties of O2 scaffolds. (A) Transmission light images of unstained and PtTFPP-stained scaffolds. Scale bar unit is lm. (B) 3D PLIM images of O2 scaffold showing 3D (left) and 2D projections (right). Image was created using 25 Z planes of 2 lm step width (50 lm). (C) O2 calibrations produced on wide-field PLIM microscope in intensity and lifetime scales. (D) Stern–Volmer plots of O2 calibration and fitting of lifetime calibration using two-site model.
previously [39]. Brain slices were fixed in paraformaldehyde solution (4% , 40 min, RT), washed with PBS (0.5 h, 5 times), blocked in 10% FBS supplemented with 0.4% Triton X100 in PBS (2 h, room temperature), incubated with primary antibodies in 2.5% HS, 0.4% Triton X100 in PBS (48 h, 4 °C), washed with TBST (1 h, 5 times), incubated with appropriate Alexa Fluor-conjugated secondary antibodies 1:1000 dilution (18 h, 4 °C), washed with TBST (1 h, 5 times), counter-stained with DAPI (14 lM, 1 h). Samples were then mounted between coverslips in ProLong Gold antifade medium (Invitrogen) and subjected to fluorescence microscopy analysis. In preparation for drug treatment O2 scaffolds were seeded with 60–100 PC12 aggregates per scaffold in SFM and grown for 3 days. 1 mM TPZ (or 1% DMSO in controls) was added simultaneously with 0.1% CellTox green and TMRM (20 nM) to the membranes and incubated for 40 min. The membranes were washed once with 1 ml of Phenol Red free DMEM media containing TMRM. The numbers of dead cells and corresponding phosphorescence lifetimes were then determined and compared to the control samples.
2.4. Brain slice culture All the procedures with animals were performed under a license issued by the Irish Government Department of Health and in accordance with the EU Directive 2010/63/EU. Newborn Sprague-Dawley rats (Biological Services Unit, University College Cork, Ireland) at postnatal day 4 (P4) were used. Brains slices were prepared as described before [22], but without embedding in agarose. Coronal sections (400 and 600 lm thick) were placed on O2 scaffolds in phenol red-free DMEM supplemented with 25% HBSS, 10% FBS, 1% penicillin–streptomycin, 10 mM Glc, 20 mM HEPES-Na, pH 7.2 (‘‘Slice medium’’) and cultured in a humidified incubator at 20% O2, 5% CO2, 37 °C for 1–5 days. Then, they were stained with CellTox Green (0.05%, 3 h), washed in fresh medium and imaged on confocal PLIM microscope as described below.
2.5. Widefield PLIM microscopy The O2 calibration and photobleaching experiments were performed on a wide-field inverted microscope Axiovert 200 (Carl Zeiss) equipped with and oil-immersion objective 40/1.3 EC Plan Neofluar, 390 nm LED excitation module, time-gated CCD camera and emission filter 635–675 nm, ImSpector software (LaVision BioTec, Germany) and integrated CO2/O2 climate control chamber (PeCon) as described before [39]. In PLIM mode the parameters were: snapshot mode, 50 ms exposure, devices delay T. For the intensity imaging the parameters were: exposure time – 2 ms, devices – none, DC mode – ON, no binning. Photobleaching experiments were performed in DC mode with the following settings: exposure time – 2 ms, devices – time–time, numbers of steps – 12, wait time – 10 s. 2.6. Measurement of dye leaching using microplate reader Dye leaching experiments were carried out on a high-sensitivity time-resolved fluorescence reader Victor2 (PerkinElmer). Supernatants from samples with O2 scaffolds were taken periodically and measured 5 times over 10 min in the wells of 96-well microplate under the following settings: mode – time-resolved fluorescence; filters – D340 excitation and D642 emission, delay time – 30 ls, gate time – 100 ls, cycle time – 1000 ls. Mean signal values were then calculated. 2.7. Confocal PLIM microscopy analysis Confocal fluorescence microscopy imaging and PLIM were carried out on a system based on an upright fluorescence microscope Axio Examiner Z1 (Carl Zeiss) with 20/1.0 W-Plan-Apochromat and 40/1.1 W-LD-C-Apochromat objectives, heated stage (37 °C) with motorized Z-axis control, DCS-120 confocal scanner (Becker & Hickl GmbH), R10467U-40 photon counting detector (Hamama-
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
4
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
A
B
C
D
E
F
Fig. 2. Cell viability and growth properties in O2 scaffolds. (A) HCT116 cells in the original (unstained) and O2 scaffolds stained with TMRM (20 nM, Red) and CellTox Green (Green). (B, C) Numbers of dead (CellTox Green) and viable (Calcein/TMRM staining) cells in the scaffolds at different depths. (D) 3D reconstruction of PC12 cells in O2 scaffold. (E, F) Viable (TMRM) and dead (CellTox Green) PC12 cells at different times in culture (2, 5 days). Scale bar unit is lm. Asterisks indicate significance level of P = 0.05.
tsu Photonics K.K.) and TCSPC hardware (Becker & Hickl GmbH) [39]. The probes were excited with tunable picosecond supercontinuum laser SC400-4 (Fianium, UK). The phosphorescence of O2 scaffolds was excited at 534 nm and collected with 635–675 nm bandpass filter, and lifetimes were calculated by mono-exponential fitting of decay curves in SPCImage software (Becker & Hickl GmbH) using the following settings: shift: 0–10, binning: 1–3, T1: 120–140, T2: 230–250. Alexa Fluor 488, CellTox Green and Calcein Green probes were excited at 488 nm with emission collected at 512–536 nm. TMRM – at 540 nm excitation, 565–605 nm emission; Alexa Fluor 594 at 594 nm excitation, 635–675 nm emission. Transmission light images (Fig. 1A) were collected with a D5100 digital SLR camera (Nikon) attached to the microscope. Drug treatments (FCCP, AntA, Tirapazamine) were performed using addition of 10 concentrated stocks in Phenol Red free DMEM at 1/10 of total volume with live samples. For 3D reconstruction samples
were fixed with PFA, immunostained (see Section 2.3) and scanned in Z direction (20–25 optical sections, 2 lm thickness, 40–50 lm total depth). 2.8. Data analysis Fitting of phosphorescence decays was performed either in ImSpector software (widefield PLIM, La vision) or SPCImage software (Becker & Hickl) using single-exponential decay function and pixel binning as appropriate. The resulting 2D matrices with lifetime data were converted to ASCII format and processed in Microsoft Excel to produce O2 concentration values using calibration function. 3D projections of O2 scaffolds were produced from intensity images representing individual optical sections using Volume Viewer plugin (Fiji software (http://fiji.sc/Fiji)), cells stained with CellTox Green and other dyes were counted
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
A
B
C
D
5
Fig. 3. Changes in oxygenation of HCT116 and PC12 cells in O2 scaffolds measured by PLIM. (A) HCT116 cells seeded at low and high densities: images of false-color PLIM superimposed with fluorescence (red). (B) Lifetime distribution at various depths. (C) PLIM and TMRM images of PC12 cells. (D) Lifetimes and oxygen concentrations in the vicinity of cell aggregates at different depths. Scale bar unit is lm. Asterisks indicate significance levels of difference with P = 0.05 (⁄) and P = 0.005 (⁄⁄).
using ’3D objects counter’ plugin in Fiji software (http://fiji.sc/ Fiji). Data are shown as average values with standard deviation as error bars. To ensure consistency of results, all the experiments were performed in triplicate. For statistical significance, data were evaluated using independent t-test, with confidence levels P = 0.05 and P = 0.005, indicated with asterisks in figures. 3. Results and discussion 3.1. Preparation of O2 scaffolds and evaluation of their optical sensor properties Alvetex scaffolds are a convenient material for the fabrication of O2 sensors. Polystyrene (PS) from which they are made provides moderate quenching for Pt(II)-porphyrins and is often used in O2 sensors [48]. The dense network of pores with thin walls allows impregnation by simple swelling in low-polarity solvents and diffusion of dissolved dye molecules into the PS [49]. For the impregnation we choose PtTFPP dye which possesses high photo-stability, brightness and hydrophobicity. We optimized the solvent, dye concentration and temperature, aiming at efficient, fast and uniform staining with the dye and minimal damage of scaffold structure (organic solvent may dissolve PS). Acetone:water mixture (7:3) was found to provide sufficient solubility for PtTFPP and impregnation at 60 °C, while tetrahydrofurane, acetonitrile or ethanol either did damage the polystyrene or did not provide sufficient solubility
for PtTFPP. The 2–3 washing steps at 60 °C with decreased acetone concentration in water successfully removed excess of dye and reinforced uniform staining throughout the volume of scaffolds. Transmission light (Fig. 1A) and confocal phosphorescence (Fig. 1B) images show that the impregnation did not alter significantly the structure of the O2 scaffold and minimally affected its roughness and shape. The size and shape of the pores remained consistent in both stained and unstained samples, and that the dye was distributed evenly across the scaffold and in depth (e.g., 50 lm optical section). The O2 sensitivity of scaffolds impregnated with 0.05 mg/ml PtTFPP was analyzed on a wide-field PLIM microscope with incubator chamber, over the range of air saturation from 21% (equivalent to 209 lM of dissolved O2) to 0%, at 37 °C, in solution. O2 calibration showed typical hyperbolic dependence of the intensity and lifetime signals (Fig. 1C), with approximately twofold reduction at 209 lM, and curved Stern–Volmer plots (Fig. 1D) reflecting non-ideal dye distribution [50]. The following equation to convert the lifetime values in oxygen concentration was obtained, according to two-site model [50]: O2 = (0.71108/( 1 + 0.71108 + s/ 56.5) 1)/0.06414 (R2 = 0.987). Fitting of the lifetime calibration with modified two-site model [50,51] produced Ksv = 0.06414 lM 1 and s0 = 56.5 ls. Confocal images on the PLIM–TCSPC microscope produced similar lifetime values under excitation at 405 nm and 534 nm. Bright phosphorescent signals allow excitation of O2 scaffolds with green light (500–535 nm), which is less damaging to the cells than 405 nm.
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
6
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
A
B
C
D E
F
Fig. 4. Oxygenation and viability of cultured brain tissue slices in O2 scaffolds. Coronal 400–600 lm thick sections of newborn rat brains (P4) were adhered to O2 scaffolds, cultured for 1 and 5 days in vitro (DIV) and analyzed for viability and oxygenation by PLIM. (A) Transmission light images of brain slice with focal points adjusted to the top and bottom (membrane). (B) Fluorescent images of neural cells in O2 scaffolds stained with cholera toxin. (C) PLIM image at the edge of brain slice. (D–F) Distribution of dead cells (CellTox Green) in 400 and 600 lm thick brain slices cultured for 1 and 5 DIV and average lifetime values for corresponding regions (N = 4). Scale bar unit is lm.
We next tested photo-stability, which is important for intensity based and long-term imaging experiments [52,53], by illuminating the O2 scaffolds on the wide-field microscope continuously for 2 min at maximal power of 390 nm LED. Similar to the other PtTFPP based materials [54], practically no photobleaching was observed. To evaluate stability of impregnation, we incubated O2 scaffolds in PBS containing 5% FBS for 7 days at 37 °C measuring leakage of PtTFPP in the supernatant. Scaffold stained with 0.05 mg/ml PtTFPP showed intensity signals of 140 106 cps (counts per second), but only trace amounts (5–7 days). Their brightness, preserved pore architecture and optimal sensitivity to O2 are well suited for quantitative imaging of O2 in phosphorescence lifetime or intensity mode. 3.2. Cell growth and viability in O2 scaffolds We studied whether the phosphorescent scaffold had any effect on cell viability or growth, using two cell models with different growth properties but with strong activity of mitochondria (oxida-
tive respiration) [55]: adherent HCT116 cells forming monolayers and PC12 cells forming multi-cellular aggregates. Since the scaffolds are not transparent and not suitable for transmission light microscopy, we visualized cells by staining with fluorescent dyes Calcein Green (cytoplasm) and TMRM (polarized mitochondria). In unstained and O2 scaffolds, we observed similar numbers and morphology of HCT116 cells, both at the surface and deeper inside (Fig. 2C, Fig. S1). PC12 cell aggregates also penetrated deep into the scaffold (Fig. 2D) and retained viability for at least 5 days (Fig. 2E). The number of dead cells (CellTox Green) was also similar, without statistically significant difference between unstained and stained scaffolds (Fig. 2A, B, E and F). PC12 aggregates showed a significantly (confidence level P = 0.05) higher number of dead cells, residing closer to the surface (10 lm depth) when compared to 40 lm depth, which can be explained by the higher density of cells in these surface regions. Gradual reduction in numbers of dead cells with the depth could also be due to the more optimal physiological conditions, physical and chemical micro-environment inside the scaffold. These results confirm that O2 scaffolds retain 3D architecture and promote cell growth as per unstained scaffolds.
Please cite this article in press as: Jenkins J et al. Oxygen-sensing scaffolds for 3-dimensional cell and tissue culture. Acta Biomater (2015), http:// dx.doi.org/10.1016/j.actbio.2015.01.032
J. Jenkins et al. / Acta Biomaterialia xxx (2015) xxx–xxx
7
A
B
C
D
Fig. 5. Stimulation of cells with drugs affecting cellular respiration in O2 scaffolds. (A) Effects of FCCP and Antimycin A. (B) Lifetime changes for selected ROIs close to cell aggregate. (C, D) Changes of lifetime values with increasing distance from cells. Scale bar unit is lm. Asterisks indicate significance levels of difference with P = 0.05 (⁄) and P = 0.005 (⁄⁄).
3.3. Monitoring of oxygenation in different cell cultures Planar O2 sensitive coatings allow monitoring of local O2 levels [56]. Local micro-gradients of O2 formed by actively respiring cells [57] can be also present within O2 scaffold. Using the O2 scaffolds, we tested 3 different models: individual cells (5–10 lm), multicellular aggregates (50–100 lm) and live brain slices (400– 600 lm). In absence of cells the scaffolds showed lifetime values 21.5–23 ls at 21% O2 and minimal (1 ls) variation across the 20 lm depth (Fig. 3A and B). When HCT116 cells were seeded at low density (monolayer cultures), the oxygen distribution remained constant, but at higher numbers (60,000–100,000 cells seeded) following cell penetration into the scaffold lifetime values increased by 2–8 ls, displaying significant lifetime differences when compared to unseeded O2 scaffolds (Fig. 3A and B). No indepth O2 gradients were observed. For multi-cellular PC12 aggregates containing 10–50 respiring cells more profound O2 gradients were expected [22]. In order to facilitate the cell invasion into the O2 scaffold, serum-free media supplemented with neural growth factor (NGF) was used. Indeed, the magnitude of lifetime increase and variation at different depths were higher than for HCT116 cells (Fig. 3C and D). Lifetimes in vicinity of PC12 aggregates ranged from approximately 26 ls at the surface to 36 ls deeper into the scaffold. This corresponds to O2 concentrations 49–16 lM (Fig. 3D). Highest lifetimes were seen at depths of 10–20 lm, and lowest – at the edges of the aggregate (0 and 40 lm). Deeper into the O2 scaffold (20 lm) O2 was significantly (P = 0.005) lower than on the surface (Fig. 3D). This is due to increased cell density and respiration activity in the aggregate core
and limitations for glucose and oxygen diffusion. For control, we stained PC12 cells with intracellular O2 probe PA2 [58] and measured aggregates in unstained Alvetex scaffolds. Similarly, higher O2 levels (lower values of phosphorescence lifetimes) were observed close to the surface, and lowest levels at 10–20 lm depths (Fig. S2). Thus, O2 scaffolds provide similar (or complementary) information about local cell oxygenation as the cell permeable icO2 probes. In cell culture brain tissue slices cells remain under native microenvironment, cytoarchitecture and are also accessible for microscopy imaging [45,59]. As brain tissue is metabolically active and sensitive to O2 fluctuations so to model in vivo physiology experiments O2 levels must be carefully controlled. We prepared 400 and 600 lm-thick brain slices from newborn rats and cultured them on O2 scaffolds for 1–5 days in vitro (DIV). Slices adhered strongly to the scaffolds and some cells migrated into it. Staining with Cholera toxin (neuronal cell tracer) was limited to surface layers
