Margherita
Bernero
,
Doris
Zauchner
,
Ralph
Müller
and
Xiao-Hua
Qin
*
Institute for Biomechanics, ETH Zürich, Switzerland. E-mail: qinx@ethz.ch
First published on 10th January 2024
During bone formation, osteoblasts are embedded in a collagen-rich osteoid tissue and differentiate into an extensive 3D osteocyte network throughout the mineralizing matrix. However, how these cells dynamically remodel the matrix and undergo 3D morphogenesis remains poorly understood. Although previous reports investigated the impact of matrix stiffness in osteocyte morphogenesis, the role of matrix viscoelasticity is often overlooked. Here, we report a viscoelastic alginate–collagen interpenetrating network (IPN) hydrogel for 3D culture of murine osteocyte-like IDG-SW3 cells. The IPN hydrogels consist of an ionically crosslinked alginate network to tune stress relaxation as well as a permissive collagen network to promote cell adhesion and matrix remodeling. Two IPN hydrogels were developed with comparable stiffnesses (4.4–4.7 kPa) but varying stress relaxation times (t1/2, 1.5 s and 14.4 s). IDG-SW3 cells were pre-differentiated in 2D under osteogenic conditions for 14 days to drive osteoblast-to-osteocyte transition. Cellular mechanosensitivity to fluid shear stress (2 Pa) was confirmed by live-cell calcium imaging. After embedding in the IPN hydrogels, cells remained highly viable following 7 days of 3D culture. After 24 h, osteocytes in the fast-relaxing hydrogels showed the largest cell area and long dendritic processes. However, a significantly larger increase of some osteogenic markers (ALP, Dmp1, hydroxyapatite) as well as intercellular connections via gap junctions were observed in slow-relaxing hydrogels on day 14. Our results imply that fast-relaxing IPN hydrogels promote early cell spreading, whereas slow relaxation favors osteogenic differentiation. These findings may advance the development of 3D in vivo-like osteocyte models to better understand bone mechanobiology.
In the field of mechanobiology, which explores the effect of physical cues on cellular behavior, studying how matrix mechanics influence cell fates has raised increasing interest. Amongst these mechanical properties, matrix stress relaxation was shown to regulate various cellular processes such as cell spreading, proliferation and differentiation in mesenchymal stem cells.10 Specifically for osteocytic development, in vitro studies have focused on the effect of matrix stiffness, which changes drastically over time as the mineralization of the organic osteoid matrix proceeds. Lower stiffnesses promoted osteogenic differentiation of MC3T3 cells in 3D gelatin hydrogels (0.58 vs. 1.47 kPa) and increased mineral deposition for primary osteoprogenitor cells in gelatin methacryloyl (GelMA) gels (6.3 vs. 16.3 and 40.2 kPa).11,12 However, the viscoelastic nature of native osteoid tissue13 is often overlooked and thus, the effect of varying matrix stress relaxation on osteocyte differentiation has not been investigated to date.
In this work, we study how osteocyte-like IDG-SW3 cells respond to different stress relaxation speeds of their surrounding matrix. We hypothesized that a faster relaxing extracellular matrix (ECM) would facilitate cell spreading through matrix reorganization and that this might affect osteocytic differentiation. Thus, we developed fast- and slow-relaxing alginate–collagen interpenetrating network (IPN) hydrogels for 3D osteogenic culture. Alginate is a natural polysaccharide composed of β-D-mannuronic acid (M) and α-L-guluronic acid (G) that forms ionically crosslinked hydrogels in the presence of divalent cations such as Ca2+.14 The fact that these ionic crosslinks can be broken and reformed when experiencing strain, renders them stress relaxing.15 These stress relaxation properties can readily be tuned by lowering the molecular weight (Mw) of the alginate polymer, which decreases connectivity and thereby increases stress relaxation speed. Adjusting the stiffness is possible independently of the alginate Mw by increasing or decreasing the Ca2+ crosslinker concentration (Fig. 1A).16
Collagen type I, the main component of osteoid, has been shown to promote IDG-SW3 osteogenic differentiation and matrix mineralization in vitro. Yang et al. have investigated the incorporation of collagen I into a GelMA–hyaluronic acid methacrylate (HAMA) bioink for 3D bioprinting of osteocyte-laden constructs. Increased functionality, dendrite number and length of 3D embedded IDG-SW3 were observed with increasing concentrations of collagen I in the bioink.9 Therefore, collagen I was included in our IPN hydrogel compositions that combine the stress relaxation properties of alginate with the permissiveness of collagen.
We established fast-relaxing (t1/2, 1.5 s) and slow-relaxing (t1/2, 14.4 s) IPN hydrogels as an in vitro platform to address the lack of osteocyte mechanobiology studies investigating matrix viscoelasticity. For this 3D culture system, pre-differentiated IDG-SW3 cells were characterized and used as an osteocyte model. For the first time, Ca2+ imaging was reported in this cell line as a tool for real-time visualization of cellular mechanotransduction. We then characterized the cell and tissue development under osteogenic conditions to identify the effects of different matrix stress relaxation speeds. This study provides new insights into osteocyte morphogenesis and differentiation in viscoelastic 3D microenvironments.
The initial elastic moduli of acellular hydrogels were computed by unconfined compression tests and resulted in a range of 4.4–4.7 kPa for the selected conditions (Fig. 1C and Table 1). Stress relaxation to compressive strain was also evaluated and a significant difference in relaxation half-time t1/2 was found between the fast- and slow-relaxing gels (1.5 s and 14.4 s, respectively, Fig. 1D). It should however be noted that overall, the stress relaxation half-times are on a smaller magnitude range than similar studies performed with mesenchymal stem cells (∼1 min to ∼1 h by Chaudhuri et al. and ∼30 s to ∼13 min by Darnell et al.).10,18 Thus, the slow-relaxing composition still exhibits substantial stress relaxation, which is also shown in Fig. 1E. This can mainly be attributed to the inclusion of collagen, a viscoelastic material itself that further increases the IPN stress relaxation speed.
Unconfined compression | Rheology | |||
---|---|---|---|---|
Init. elastic modulus [kPa] | Relaxation time t1/2 [s] | Storage modulus G′ [kPa] | Relaxation time t1/2 [s] | |
Fast | 4.429 ± 0.253 | 1.5 ± 0.2 | 0.789 ± 0.009 | 17.8 ± 3.8 |
Slow | 4.684 ± 1.526 | 14.4 ± 2.7 | 0.972 ± 0.077 | 32.35 ± 4.4 |
The second method used to mechanically characterize the casted hydrogels was in situ rheology. Using amplitude sweeps to determine the linear viscoelastic range and storage moduli as a measure of stiffness, the relative stiffnesses found in compression tests could be confirmed (0.79–0.97 kPa, Fig. 1F and ESI Fig. S1†). The difference in shear stress relaxation was also significant, although less pronounced with stress relaxation half-times t1/2 of 17.8 s for fast-relaxing gels and 32.4 s for the slow-relaxing condition (Fig. 1G). A difficulty for the analysis of the rheological stress relaxation tests was the relatively large noise at the beginning of the measurements caused by a high sampling rate to capture the initial relaxation kinetics. Such noise has also been reported previously19 but filtering it out prior to evaluation allowed the generation of relatively smooth stress relaxation curves (Fig. 1H).
Multiple assays including this marker were used to assess the successful pre-differentiation of IDG-SW3 after 14 days in 2D culture. Quantification of Dmp1-GFP-expressing cells by confocal fluorescence microscopy showed that around 34% of cells had reached an osteocytic phenotype by day 14 and nearly 50% of all cells were GPF-positive by day 28 (Fig. 2A and B). These results fail to reach the levels reported in the original IDG-SW3 publication6 and could indicate a slower differentiation progress in our culture. Apart from the continuous increase in the fraction of Dmp1-expressing cells, the fluorescent intensity of GFP-positive cells also significantly increased until day 28 (Fig. 2C). This implies that gradually more and more cells differentiate, reach an early osteocyte stage, and continue to mature during the 28 days of osteogenic culture. Thus, a similar continued development can also be expected for 14-days pre-differentiated IDG-SW3 used for 3D cell embedding.
Fig. 2 Characterization of 2D pre-differentiating IDG-SW3 cells. (A) Representative fluorescence microscopy maximum intensity projections on days 1, 7, 14 and 28 of 2D culture under osteogenic conditions used for quantitative analysis of Dmp1-GFP expression. (B and C) The fraction of Dmp1-GFP-positive cells at all time points and each fluorescent cell's mean intensity were quantified using the CellProfiler software. (D) ALP activity quantified by a colorimetric assay and normalized to the DNA content. Welch's one-way ANOVA showed a significant difference among means (p = 0.05 to 0.01), however, pairwise comparisons were non-significant. (E) DNA content of initially 0.5 × 106 seeded cells quantified by Quant-iT PicoGreen dsDNA assay. (F–H) Mechanoresponse of pre-differentiated IDG-SW3 visualized by live cell Ca2+ imaging under fluid flow in 2D on day 14. Cells were stained with Fluo-4 to emit a Ca2+-concentration dependent fluorescent signal when a fluid shear stress of 2 Pa was applied. See ESI Fig. S2† for a schematic depiction of the experimental setup. (F) Maximum intensity projections of the baseline under static conditions and at three time points during fluid flow application are shown visualizing fluorescence intensity (Dmp1-GFP and Ca2+ signal overlapping). Selected responsive cells are highlighted and green boxes denote a Ca2+ spike. See also ESI Video S1.† (G) Normalized intensity signals of selected responsive cells are shown over time. Each line represents an individual cell and most cells respond immediately after fluid flow onset at 1 minute. (H) Mean intensities of all cells were analyzed individually over time and the number of spikes was recorded. (I) Gene expression of markers for osteogenic differentiation (Alpl: osteoblast, Pdpn and Dmp1: early osteocyte, B2m as reference gene) was quantified by RT-qPCR before and after the two-week pre-differentiation period. (B, D, E and I) Data is presented as mean ± SD, n = 3. Welch's one-way ANOVA with Dunnett's T3 multiple comparisons test or two-way ANOVA with Šídák's multiple comparisons test were performed. (C) Data is shown with median (solid line) and quartiles (dotted line), n = 75–2631. A Kruskal–Wallis test followed by Dunn's multiple comparisons test was performed. *p = 0.05 to 0.01, ****p ≤ 0.0001. |
Dmp1 gene expression during culture in osteogenic conditions was subsequently confirmed using real-time quantitative PCR (RT-qPCR). After 14 days, Dmp1 expression showed a significant increase with a log2-fold change above 10 compared to the expansion cells (Fig. 2I). Additionally, the osteoblast differentiation marker alkaline phosphatase (Alpl) and early osteocyte marker podoplanin (Pdpn, also known as E11) were analyzed by RT-qPCR. Both genes showed significant upregulation after differentiation was induced.
Alkaline phosphatase enzyme (ALP) activity can also be measured on the protein level and has been used to characterize IDG-SW3 during osteoblast-to-osteocyte transition when the cell line was first reported, peaking on day 14.6 ALP activity was evaluated for our 2D cell cultures up to day 14 using a colorimetric assay and showed a strong increasing trend (Fig. 2D). DNA contents of the samples were quantified for ALP assay normalization and revealed a significant decrease (Fig. 2E). We assume that some cells are lost during the differentiation process, either by failing to attach properly to the culture flask surface or by apoptosis.
To further verify the mechanosensitivity of the pre-differentiated IDG-SW3 cells, a protocol for live cell calcium (Ca2+) imaging under fluid flow stimulation was developed. Osteocytes have been shown to be especially responsive to fluid flow. Such fluid shear stresses can induce the opening of stretch-activated ion channels in the cell membrane, causing a Ca2+ influx from the extracellular fluid and the endoplasmatic reticulum that initiates an intracellular mechanotransduction cascade.20–22 Visualizing changes in Ca2+ concentration therefore allows the tracking of mechanosensation in real-time. Applying a shear stress of 2 Pa in a commercial ibidi μ-Slide, we demonstrate the successful Ca2+ visualization in pre-differentiated IDG-SW3 using the Ca2+-indicator Fluo-4 (Fig. 2F). The inclusion of Pluronic F-127, a dispersing agent increasing Fluo-4 AM solubility and uptake into the cells, was found to be crucial to ensure sufficient dye loading for Ca2+ imaging. Despite the emission spectrum overlaying with Dmp1-GFP, it was possible to detect Ca2+ spikes upon fluid flow stimulation (Fig. 2G and ESI Video S1†). Almost 60% of the analyzed IDG-SW3 responded to the 2 Pa shear stress and most of them spiked immediately after fluid flow onset. Additionally, more than half of the responsive cells were able to produce multiple spikes over the course of 4 minutes (Fig. 2H). Although calcium spikes had previously been recorded in response to fluid shear stress for MLO-Y4 cells, another murine osteocyte-like cell line,22 this is the first account of live cell calcium imaging with IDG-SW3 cells. Having confirmed the mechanosensitivity of 14-days pre-differentiated IDG-SW3 will allow us to extend our studies of osteocyte mechanobiology to 3D settings in vitro.
Pre-differentiated IDG-SW3 were directly suspended in the hydrogel precursor solution at a concentration of 4.5 × 106 cells per mL and embedded in the IPN hydrogels analogously to acellular gel casting (Fig. 1B). To validate the system for 3D culture compatibility, the cell viability was quantified over 7 days after embedding. Viabilities did not differ significantly between fast- and slow-relaxing hydrogels but were overall lower on day 1 with around 75% (Fig. 3B and C). However, viability increases until day 7 where it reaches around 90% for both groups, indicating that the IPN hydrogels support IDG-SW3 culture in vitro.
Further insights can also be gained from the DNA quantification that was performed in combination with the ALP assay for its normalization. Both conditions showed a comparable decrease from day 1 to days 7 and 14 (Fig. 3D). This may partly be explained by the loss of cells migrating out of the gels into the culture medium, as cells had been observed growing on the culture well bottoms. Osteogenic differentiation followed by apoptosis may be a second factor causing a reduction in DNA, which is in line with DNA decreases during 2D pre-differentiation and was also noted in previous studies working with mesenchymal stem cells under osteogenic conditions in 3D.24,25 However, quantification of pro-apoptotic (Bax) and anti-apoptotic (Bcl-2) marker gene expression indicates that compared to 2D pre-differentiation, there is still a stronger apoptotic stress on cells in osteogenic 3D culture that decreases until day 14. This may be related to the cellular stress during embedding (ESI Fig. S3†).
Finally, the sum of dendrite lengths was also quantified from the actin-nuclei z-stacks on day 1 using an automated dendrite tracking tool. The longest dendrite lengths were measured in the fast-relaxing composition which also showed a slightly higher median length (Fig. 4C). The difference, however, was not statistically significant and future more highly resolved studies of individual dendrite lengths might be of interest to capture the visible difference in the microscopy images.
Fig. 5 Molecular cell markers of differentiation and matrix mineralization under osteogenic 3D culture. (A–C) Gene expression quantification of osteogenic markers (Alpl: osteoblast, Pdpn and Dmp1: early osteocyte, B2m as reference gene) by RT-qPCR for 3D cultures over 14 days. The fold change is shown relative to a single reference point for all genes. (D) ALP activity in 3D samples quantified on days 1, 7 and 14 of osteogenic culture by a colorimetric assay and normalized to the DNA content (see Fig. 3D). (E) To visualize internal matrix mineralization over the course of 14 days, maximum intensity projections of 20 μm sections stained with OsteoImage reagent and Hoechst for hydroxyapatite and nuclei, respectively, are shown. A negative control without OsteoImage staining was included to verify a negligible signal contribution by Dmp1-GFP. (F) The mean OsteoImage signal area normalized by the number of cells was quantified form the fluorescence images. Data points with the same symbol (circles, triangles) represent different positions within the same sample. (A–D and F) Data is presented as mean ± SD, (A–D) n = 3, (F) n = 2. Two-way ANOVA with Šídák's multiple comparisons test was performed. *p = 0.05 to 0.01, **p = 0.01 to 0.001, ****p ≤ 0.0001. |
Furthermore, ALP enzyme activity was assayed on days 1, 7 and 14 in 3D culture. In agreement with the RT-qPCR data, ALP activity also increased on the protein level and was significantly higher in the slow-relaxing condition on day 14 (Fig. 5D). This disagreement with reports in literature might be due to differentiation proceeding more slowly in our cultures, which had already been indicated by the Dmp1-GFP-positive cell fraction during osteogenic culture in 2D. Additional stresses may be put on the cells during dissociation after pre-differentiation, the cell embedding procedure and adaptation to a 3D culture environment.
As the embedded cells mature, we also expect changes in the surrounding matrix due to proteolytic remodeling, secretion of matrix proteins and induced mineral deposition. Thus, matrix mineralization was investigated using OsteoImage staining for hydroxyapatite. Despite a spectral overlap with the Dmp1-GFP signal, OsteoImage was detectable in mineralized nodules on day 14 (Fig. 5E). At the selected image contrast settings, barely any Dmp1-GFP signal can be seen in the negative control without OsteoImage staining. The signal from the fluorescence microscopy images was quantified and normalized to the cell number (Fig. 5F). A significant difference was found for slow-relaxing gels on day 14 both compared to the baseline sample on day 1 and compared to the fast-relaxing gels on day 14. This enhanced mineralization can also be observed in histological sections stained with Alizarin Red S (ESI Fig. S4†) and confirms our findings of increased differentiation markers in slow-relaxing hydrogels at the end timepoint day 14. To verify whether the cell-induced matrix remodeling and mineralization had an effect on bulk gel mechanical properties, unconfined compression tests were performed on days 1, 7 and 14. These did not reveal any evident changes of elastic modulus over time and showed that the hydrogels retained their initial stress relaxation properties (ESI Fig. S5†). Nevertheless, the local mechanical environment may undergo changes undetectable by bulk measurements. Future investigations into the dynamics of cellular matrix remodeling are therefore warranted. Newly deposited matrix greatly influences how cells sense their local environment, which could for example be visualized by nascent protein labelling.26
By administering the gap junction inhibitor 18α-glycyrrhetinic acid (AGA), a negative control condition with inhibited dye transfer over gap junction could be established (Fig. 6C). This enabled us to confirm gap junctional activity at the observed cell–cell-contacts and exclude major contributions from other potential mechanisms of fluorescent dye uptake. In untreated samples, the connected cells within the slow-relaxing gels were able to regain up to >60% of their initial fluorescent intensity. The large variation within the treatment groups could be attributed to varying numbers of connecting gap junctions between the cells. Seeding cells at a higher density in future experiments could increase the number of successfully connected cells and further allow for a greater sampling number and more in-depth investigation.
Besides cell connectivity, osteocyte functionality also depends on processing and responding to soluble signals. In vivo, osteocytes and the coordination of bone homeostasis are regulated by parathyroid hormone (PTH). For studies of osteocyte models in vitro, treatment with PTH fragment 1–34 and the appropriate response in the gene expression levels of osteocyte-specific genes has been used to demonstrate osteocyte-like functionality.5,6,9 Generally, a decrease in osteocyte-specific markers including Dmp1 was observed, but Pdpn and RANK ligand tend to increase. We evaluated cellular responsiveness to PTH treatment (50 nM) for 24 h in both fast- and slow-relaxing hydrogels after 14 days of 3D culture. Gene expression levels were analyzed in comparison with untreated control samples (Fig. 6D). As expected, Dmp1 expression was downregulated in both fast- and slow-relaxing compositions, however only showed a significant difference to the untreated control samples in slow-relaxing gels. A trend for upregulation of Pdpn could be seen for the slow-relaxing gels that showed strongest osteocyte-like responses on day 14 so far, but not for the fast-relaxing gels.
Finally, to evaluate whether the embedded cells retained their mechanosensing abilities, we performed live-cell 3D Ca2+ imaging in slow-relaxing IPN hydrogels on idenTx 3 chips. The Fluo-4 staining method was validated for the 3D setup by administering GSK1016790A as TRPV4 agonist, which induced prolonged Ca2+ signals in selected cells. In a second experiment, much shorter calcium spikes indicative of mechanotransduction were occasionally detected when applying a fluid flow through the IPN hydrogel (ESI Fig. S6 and Videos S2, S3†). The interpretation of this preliminary data is however still limited by the missing knowledge of the IPN hydrogel porosity, affecting both molecular diffusion and the effective shear stress on the cells.
Cell morphological analysis showed enhanced cell spreading in the fast-relaxing composition at early time points, but differences in mean cell area decreased over time. At the same time, however, biomolecular and functional osteocyte markers as well as matrix mineralization appeared to increase with slow matrix relaxation speeds at the end timepoint after 14 days. We believe that although faster stress relaxation can accelerate early cell spreading, it may be unable to supply sufficient mechanical support for long-term osteogenic differentiation. Instead, a minimal ECM elasticity may be required for continued cell–matrix interaction and the successful extension of stable dendrites into the extracellular space. It is noteworthy that both our fast- and slow-relaxing compositions exhibit substantial stress relaxation. Thus, future work is needed to extend the investigation to hydrogels with a wider range of stress relaxation speeds.
Our viscoelastic IPN hydrogels prove to be suitable for the investigation of osteogenic cultures in vitro and our findings provide further insights into the influence of matrix mechanics on osteocyte morphogenesis. The next goal would be to improve cell networking in order to better model the osteocyte networks found in bone. An avenue to be investigated is the use of more defined, synthetic hydrogels that allow the decoupling of viscous and elastic properties.28,29 This would further allow the combined study of multiple matrix properties, such as stiffness and stress relaxation, and potentially reveal synergistic effects of matrix mechanics on osteocyte morphogenesis. Still, additional mechanical cues by compressive or fluid shear stimulation may be needed to mimic the dynamic nature of the bone microenvironment and enable long-term, more in vivo-like tissue cultures. In combination with the herein established Ca2+ imaging method, this would open new possibilities to study the mechanosensation and cell–cell communication of osteocytes in vitro.
Amplitude sweeps were carried out at 1 Hz oscillating shear strain of 0.01–6.85%. Storage moduli were calculated from the average value in the range of 0.101–0.685% strain, which was linear for all gels (ESI Fig. S1†). After 5 minutes of recovery, stress relaxation tests were performed by preconditioning the samples at a 1 Hz oscillating strain of 0.5% for 3 minutes, then suddenly increasing the strain to a constant 10% while monitoring the shear stress.31 Measurement time points were logarithmically ramped from 0.01 s to increase the resolution of the initial relaxation. The thereby generated noise was removed using a moving median filter with a window size of 30 points, followed by a low-pass filter with a constant cut-off frequency of 0.5 on all data points after 10% strain is reached. Then, the measured shear stress was normalized to the initial value defined as the first value at 10% shear strain. The relaxation half-time was calculated based on the time needed to reach half of the initial shear stress.
The calcium dye staining solution was prepared as per the manufacturer's recommendations, mixing the 1 mM Fluo-4 AM stock (Invitrogen, F14201) 1:1 with 20% Pluronic® F-127 (Sigma-Aldrich, P2443) in DMSO and diluting this 1:100 in phenol-red-free alpha MEM to obtain a final concentration of 5 μM Fluo-4 AM and 0.1% Pluronic® F-127. On day 14 of pre-differentiation, cells were washed once with warm PBS, stained for 30 minutes at 37 °C, and washed once with Hank's balanced salt solution containing Ca2+ (HBSS, Gibco, 14025050). Channel openings were capped with pieces of autoclaved Microseal sealing film until imaging.
The same HBSS buffer was used for perfusion of the chips. Therefore, 50 mL syringes were filled with pre-warmed HBSS under sterile conditions. At the Leica SP8 confocal laser scanning microscope (CLSM), the syringe, chip and an outlet container were connected to an AL-1000 syringe pump (World Precision Instruments), and the chip was left to rest for 10 minutes, allowing the cells to recover from any disturbing mechanical stimuli. For live-cell imaging, 12 μm xyzt-stacks were recorded every 3 seconds for 5 minutes. After 1 minute of static baseline condition, fluid flow was initiated at a flow rate Q of 12.93 mL min−1. With a buffer viscosity μ of 9.4 × 10−4 Pa s, this flow rate caused a shear stress τ of 2 Pa on the cells according to eqn (1), which can be used to calculate the wall shear stress in a parallel-plate flow chamber.32
(1) |
The time series maximum intensity projections (MIPs) were analyzed in ImageJ using the MultiMeasure plugin. Regions of interest (ROIs) were manually drawn around each cell and the intensity values exported. Since IDG-SW3 express Dmp1-GFP at this stage of differentiation, the baseline intensity was defined as the mean intensity during static conditions minus the baseline intensity measured for non-fluorescent human mesenchymal stem cells (Lonza) stained using the same protocol. This baseline was subtracted from the intensity measurements and Ca2+ spikes were defined as intensity values exceeding the 9-fold standard deviation of the measurements during the static first minute. This threshold was chosen 3-fold higher than by Degala et al.33 due to higher noise fluctuations. Finally, the total number of spikes per cell was computed.
After successful extraction of the gels from their molds, they were transferred into 24-well plates containing differentiation medium supplemented with 1 mM additional CaCl2, resulting in a final concentration of 3 mM. Culture at 37 °C and 5% CO2 was continued, and medium was exchanged twice per week.
At the Leica SP8 CLSM, the gels were imaged at two to three positions capturing 36 μm z-stacks. Automated cell area analysis was conducted in ImageJ. MIPs of both channels were segmented with a manually set, fixed threshold and the mean cell area was computed by dividing the area of actin particles by the number of nuclei. Dendrite quantification was performed using the IMARIS automated dendrite tracking software with a dendrite diameter set to 0.8 μm.
At the Leica SP8 CLSM, 60 μm z-stacks spanning the entire sample were acquired at two positions per section. Automated particle analysis in ImageJ was used to quantify the fluorescent signal. MIPs of both channels were segmented with a manually set, fixed threshold and the mean OsteoImage-stained area per cell was computed by dividing the total area of OsteoImage signal by the number of nuclei.
At the Leica SP8 CLSM, FRAP was performed using the FRAP module of the LAS X microscope software. A ROI was manually selected around the target cell for bleaching with 100% 488 nm laser power at the center height of the z-stack. Pre- and post-bleaching, 30 μm z-stacks with 5 slices were excited with the same 488 nm laser at 0.5% power. Two pre-bleach stacks were acquired with a 10 seconds interval, then the target cell was bleached using 2 iterations and finally post-bleach stacks were acquired for up to 7.5 minutes every 3 seconds. FRAP was performed in slow-relaxing gels for control cells and AGA-treated cells.
In ImageJ, MIPs were computed, and circular ROIs were drawn around the bleached cell, a background area and a reference cell that was not in contact with the bleached cell. Mean intensity values of the ROIs over time were extracted using the MultiMeasure plugin. The intensity values of the target cells were corrected for the background fluorescence and normalized to the reference cell in order to account for photobleaching during image acquisition. Then the fluorescence recovery after bleaching was computed normalized to the average intensity of the two pre-bleach stacks as 1 and the first stack after bleaching as 0.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3bm01781h |
This journal is © The Royal Society of Chemistry 2024 |