We present a two-part protocol that combines fluorescent calcium imaging with in situ hybridization, allowing the experimenter to correlate patterns of calcium activity with gene expression profiles on a single-cell level.
Spontaneous intracellular calcium activity can be observed in a variety of cell types and is proposed to play critical roles in a variety of physiological processes. In particular, appropriate regulation of calcium activity patterns during embryogenesis is necessary for many aspects of vertebrate neural development, including proper neural tube closure, synaptogenesis, and neurotransmitter phenotype specification. While the observation that calcium activity patterns can differ in both frequency and amplitude suggests a compelling mechanism by which these fluxes might transmit encoded signals to downstream effectors and regulate gene expression, existing population-level approaches have lacked the precision necessary to further explore this possibility. Furthermore, these approaches limit studies of the role of cell-cell interactions by precluding the ability to assay the state of neuronal determination in the absence of cell-cell contact. Therefore, we have established an experimental workflow that pairs time-lapse calcium imaging of dissociated neuronal explants with a fluorescence in situ hybridization assay, allowing the unambiguous correlation of calcium activity pattern with molecular phenotype on a single-cell level. We were successfully able to use this approach to distinguish and characterize specific calcium activity patterns associated with differentiating neural cells and neural progenitor cells, respectively; beyond this, however, the experimental framework described in this article could be readily adapted to investigate correlations between any time-series activity profile and expression of a gene or genes of interest.
Free cytosolic calcium is critical to a variety of biological processes, ranging from cell proliferation and migration to apoptosis and autophagy1,2,3. Within these pathways, calcium can exert downstream effects on gene expression by interacting with calcium-binding domains to induce conformational changes that modulate protein activity and interactions. For example, a neuronal calcium sensor known as the Downstream Regulatory Element Antagonist Modulator (DREAM) is held in an unfolded intermediate conformation when bound by calcium, preventing it from interacting with its protein and DNA targets4. Beyond serving as a simple signaling molecule, however, the dynamic nature of intracellular calcium transients allows these activity patterns to encode more complex amplitude- or frequency-based signals5,6. Nuclear translocation of the transcription factor nuclear factor of activated T-cells (NFAT) is enhanced by high-frequency calcium oscillations but inhibited by low-frequency oscillations7. Compellingly, recent work has suggested that NFAT may actually responsive to cumulative calcium exposure8. Both calcineurin and Ca2+/calmodulin-dependent protein kinase II (CaMKII) also exhibit distinct responses to calcium transients of a specific frequency, duration, or amplitude9. To add an additional level of regulatory complexity, computational models suggest that many downstream calcium-binding proteins become more or less frequency-dependent in response to the presence or absence of binding competitors10,11.
Within the developing nervous system, two main classes of calcium activity behaviors have been defined and associated with specific biological processes. Calcium influxes are classified as "spikes" if they occur within individual cells, reach a peak intensity of ~400% of baseline within five seconds, and exhibit double exponential decay12. This type of signal is associated primarily with neurotransmitter phenotype specification13. In contrast, "waves" are defined as slower, less extreme calcium transients in which a cell's intracellular calcium concentration rises to ~200% of baseline over a period of thirty seconds or more, then decays over several minutes12. These signals often propagate across multiple neighboring cells, and their presence has been associated with neurite outgrowth and cell proliferation14,15. However, although these two classes have been defined based on characteristic kinetic profiles, it remains unclear exactly which characteristics of these patterns are actually being detected by cells and translated by downstream effectors.
Understanding the relationship between intracellular calcium oscillations and gene expression would provide crucial insight into one of the regulatory mechanisms that ensures appropriate development and patterning of the nervous system. To this end, studies of the embryonic spinal cord have demonstrated that increased calcium spike activity during development is associated with higher levels of inhibitory neurons, while decreased calcium spike activity is associated with higher levels of excitatory neurons13. However, these population-level assays have not been used to associate calcium activity with gene expression on a single-cell level.
Approaching these questions on the level of the single cell offers several distinct advantages over previous work. For one, the ability to assess calcium activity and gene expression in many cells individually allows the full repertoire of distinct activity patterns to be observed without being obfuscated by a bulk-level measurement. Additionally, studying these relationships in single-cell primary culture means that cell-autonomous links between calcium activity and gene expression will be maintained, while interactions requiring cell-cell communication will be abrogated. Therefore, this approach allows these cell-autonomous mechanisms to be studied in isolation. However, it also allows the role of non-cell-autonomous calcium activity to be elucidated and interrogated. For example, cells can be dissected from an embryo at the neural plate stage, cultured until sibling controls reach the neural tube stage, and then compared to cells that have been freshly dissected from a neural-tube-stage embryo. This allows direct comparison of cells that retained cell-cell communication across a key developmental period to those in which cell-cell communication was abolished.
In seeking to address the limitations of previous experimental approaches, we developed a protocol that would enable the assessment of both calcium activity and gene expression in individual neural progenitor cells, facilitating the correlation of specific activity patterns with subsequent differentiation programs. Neural tissue was dissected from Xenopus laevis at various stages of neural development, dissociated into single cells, and imaged via confocal microscopy in the presence of a fluorescent calcium indicator. Following live-cell imaging, samples were fixed and assayed via fluorescence in situ hybridization (FISH) to detect expression of a gene or isoform of interest. Importantly, individual cells can be tracked across both imaging experiments, meaning that a cell's calcium activity profile and its gene expression level can be associated with one another (Figure 1). The protocol reported here is intended to probe relationships between calcium activity patterns and gene expression across embryonic neurodevelopment in Xenopus laevis. However, the broader experimental framework (single-cell time-course imaging followed by FISH and image coregistration) can be modified and applied to virtually any cell type, fluorescent reporter, and gene of interest.
All work involving animals was performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the College of William and Mary.
1. Animal Care and Embryo Handling
2. Embryo Dissection and Sample Preparation
3. Calcium Imaging
NOTE: Calcium imaging was performed using an inverted confocal microscope (Table of Materials).
4. Gene Expression Analysis: Probe Synthesis
5. Gene Expression Analysis: Fluorescence In Situ Hybridization
NOTE: All washes should be performed with approximately 1 mL of solution using a sterile, individually wrapped transfer pipette. The pipette should be positioned at the edge of the plate when removing or adding solution, and washes should be performed as gently as possible to ensure that cells are not dislodged from the plate surface and lost.
6. Imaging Cells
NOTE: Imaging was performed using an inverted confocal microscope.
7. Data Processing
NOTE: Data processing was performed using Nikon Elements software.
A successful example of dissociated cells prepared for calcium imaging can be seen in Figure 2A. Cells are densely plated, allowing the maximum amount of information to be collected from each image, but not so densely plated that individual cells cannot be confidently distinguished form one another. Fluorescence is detected for each defined cell over the 2 h imaging period. Visualization of a composite plot containing the traces for all cells recorded in an experiment reveals the degree to which bulk or population measurements can obscure more nuanced patterns of spiking behavior (Figure 2B). When the recorded profiles of individual cells are isolated, examples of the irregular spiking activity characteristic of neural progenitor cells can be clearly identified. Unlike mature neurons, embryonic neuronal cells exhibit irregular, highly variable and complex nature of calcium activity (Figure 2B). In order to quantify this complexity, application of different data analysis methods have been applied17,18, including diverse parameters to define a spike (Figure 2C).
Successful fluorescence in situ hybridization, including successful design and synthesis of an antisense mRNA probe, can be assessed by comparing the experimental plate against a background control incubated with a non-binding sense RNA control (Figure 3A,B). A positive probe control can also be performed by processing a cell type known to express the target mRNA at detectable levels.
Identification of the same cell across calcium and FISH imaging requires that cells retain roughly the same position during probe hybridization and processing. If plates are handled roughly or washes are performed too forcefully, cells can be dislodged from the plate surface and either lost when solution is discarded or deposited on a different location on the plate, making it impossible for them to be matched across images (Figure 4A). If this disruption affects only some of the cells in the field of view, it may still be possible to detect and assign some cells within the image (Figure 4B). However, the maximum amount of data is gained from an experiment in which FISH is performed carefully and few cells are lost or repositioned between images (Figure 4C).
Once data has been collected to describe both the calcium activity and gene expression of a reasonable number of cells at the developmental stage(s) of interest, further analyses can be performed to assess correlations between these two features (Figure 1). A number of metrics have been applied to quantify calcium activity patterns, including spike counting/frequency, average power, Hurst exponent estimation, and Markovian entropy measurement17,18. Gene expression can be defined quantitatively by absolute fluorescence level or graded on a binary (yes/no) scale, depending on the experimental questions being addressed.
Results from experiments collating calcium activity with the expression of neural progenitor marker genes revealed numerous associations between specific patterns of calcium activity and neurotransmitter phenotypes. At the neural plate stage (Stage 14), GABAergic cells expressing the inhibitory neuron marker gad1.1 exhibit calcium activity that is more regular and higher-amplitude than that of cells that lack gad1.1 expression (Figure 5A). Furthermore, while these gad1.1-expressing cells are associated with higher levels of high-amplitude spiking, low-amplitude spiking is more frequent in glutamatergic cells expressing the excitatory neuron marker slc17a7.
Figure 1: Schematic of experimental workflow. Scale bar = 100 μm. Images in panels 3-5 were taken from Paudel et al. (2019)17. Please click here to view a larger version of this figure.
Figure 2: Calcium imaging and example activity profiles. (A) Intracellular calcium activity as reported by Fluo4-AM. Each 2 h image is composed of 901 frames, with one representative frame show here. (B) Composite plot of fluorescence intensity over time in all cells within the imaged field of view. The traces clearly indicate photobleaching of indicator dye (Fluo4) overtime. Raster plot on top left shows representative traces of calcium activity after application of a de-trending algorithm developed by Eilers and Boelen19, where cells shown here exhibit diverse pattern of spiking behavior. (C) Application of different thresholds (150% and 200% of baseline, where the baseline is the average of de-trended fluorescent intensity) to define a spike (green and blue arrows). Scale bar = 100 μm. Please click here to view a larger version of this figure.
Figure 3: FISH imaging. (A) FISH performed with a non-binding sense RNA probe as a negative control. Imaging settings have been adjusted so that no cells appear fluorescent. Some fields of view may include non-cell debris with some fluorescence, such as seen in the top right and bottom right corners of (A); these can be ignored for the purpose of background setting. (B) The same imaging settings are then used to image an experimental plate (antisense RNA probe). Fluorescence under these conditions corresponds to gene expression above background. Scale bar = 100 μm. Please click here to view a larger version of this figure.
Figure 4: Image overlay and coregistration. Schematic representations of a sample imaged for calcium activity (cells represented by filled green circles) and after FISH (cells represented by shaded red circles). (A) Cells that have moved significantly during sample handling and processing cannot be reliably identified across the two images. (B) Cell disruption may affect only some cells in the field of view. Some cells can be clearly identified in both images, while others cannot be confidently matched. (C) If samples are handled carefully, most cells will remain undisturbed and can be identified in both images. Please click here to view a larger version of this figure.
Figure 5: An example of application of this method, boxplots showing associations between calcium activity and gene expression (GABA and Glut for genes gad1.1 and slc17a7 respectively) in neural plate stage Xenopus laevis. At stage 14, gad1.1-postive cells (GABA) exhibit higher-amplitude and more regular calcium activity as defined by (A) Markovian entropy18 and (B) spike counts using thresholds 125%, 150%, 200% and 300% of the average of the de-trended fluorescent intensity (baseline)17 than slc17a7 positive cells (Glut). Stars indicate statistically significant differences according to both Bonferroni-corrected two-sample Kolmogorov–Smirnov Test (p < 0.05) and Cohen's d statistics for effect size (n = 5 cultures and >100 cells; * 0.2 ≤ |d| < 0.5). The figure was redrawn and adapted from data set obtained from Paudel et al.17. Please click here to view a larger version of this figure.
Characteristic patterns of calcium activity have been observed in the cells that make up the developing nervous system, with specific types of activity associated with distinct neurodevelopmental processes. However, further understanding of the mechanisms by which these information-dense activity patterns are translated into transcriptional responses requires information about calcium activity and gene expression to be collected with single-cell resolution. While systems that exhibit more stereotypical calcium activity, such as mature neurons, can be reasonably assayed on a bulk level, the irregular patterns that characterize the embryonic nervous system are easily masked by less precise recordings.
The experimental framework established in this protocol is easily adaptable to a wide variety of cell types and fluorescent reporters. Tissue containing virtually any cell type or combination of cell types can be dissected from a model organism of interest and plated for single-cell imaging. In addition to allowing cell identification and isolating the effect of cell-autonomous processes, a primary cell culture approach allows the experimenter to define media components as desired. For example, experiments comparing the activity of neuronal precursors in 2 mM Ca2+ solution have been performed to investigate whether the relationships between spike frequency and neurotransmitter phenotype in the embryonic spinal cord can be recapitulated without the influence of cell-cell interactions13,20.
While this protocol leverages the fluorescent marker Fluo4-AM to detect intracellular calcium activity, depending upon the selection criteria, users can select other commercially available markers21, including genetically encoded calcium indicators. Similarly, alternative markers could be used to monitor dynamic changes to concentration of an ion of interest (including K+, Na+, and Zn2+), membrane potential, or cellular pH. Imaging settings and image duration can be modified as necessary.
Although we correlated calcium activity and neuronal phenotype as a specific application, this method is also applicable for a variety of other cellular properties. For example, fluorescence in situ hybridization can be performed with probes against any gene of interest, including the neuronal marker ChAT or the transcription factor Engrailed, allowing sensitive detection of a customizable panel of mRNA species. These probes can be designed to be isoform-specific, supporting additional target specificity if desired. Double FISH can be performed using probes conjugated to several two different fluorophores, allowing the simultaneous assessment of the expression of multiple genes. However, the additional washes required by this type of experiment are associated with an increased chance of cell loss or movement and require experience and delicacy to be performed successfully.
Regardless of any experiment-specific modifications made to this protocol, there are several key steps that require careful attention. Dissections should be performed with care to remove all contaminating tissues or cell populations; because spatial patterning is lost when the explants are dissociated, any remaining cells from neighboring tissues will become interspersed with and indistinguishable from the cells of interest. After cells are plated, samples should be handled as gently as possible to prevent cells from being dislodged. Most importantly, this means that all solution changes should be performed slowly and carefully, with the pipette placed at the edge of plate when solution is being removed and added. This will ensure that cells can be confidently identified in both calcium and FISH images. If cells are disrupted during processing, it may be impossible to identify some or all of the corresponding cells between the two images. We advise erring on the side of caution with these assignments, such that only unambiguously corresponding cells are used for further analysis.
Depending on the biological question being addressed, a variety of analysis approaches may be appropriate. Time-series calcium activity can be processed and quantified in a variety of ways, with experimenter flexibility in choosing de-trending parameters, analysis metrics, and analysis parameters (for example, the % of baseline threshold used to define a calcium spike). Correlations between calcium activity and level of gene expression can be drawn by analyzing gene expression as an absolute or relative fluorescence value extracted from the FISH image. Alternatively, correlations between calcium activity and gene expression (presence/absence) can be drawn by defining a fluorescence threshold for positive gene expression signal and assigning 'yes' or 'no' identifiers to individual cells. As a whole, this experimental schema provides an incredibly flexible pipeline for the collection and preliminary analysis of time-series data in conjunction with cell-matched gene expression data. Such experiments will be critical for better understanding the complex relationships between cellular dynamics and transcriptional changes, as exemplified by the identification of calcium activity patterns characteristic of inhibitory-fated and excitatory-fated neuronal precursors in embryonic Xenopus laevis.
The authors have nothing to disclose.
We thank Wendy Herbst and Lindsay Schleifer for their contributions to the development of these protocols. This work was supported by grants from the National Institutes of Health (1R15NS067566-01, 1R15HD077624-01 and 1R15HD096415-01) to MSS.
For Animal Husbandry & Cell Culture | |||
CHORULON (chorionic gonodotropin) | Merck Animal Health | ||
Gentamycin sulfate salt | Millipore Sigma | G1264 | |
Penicillin-Streptomycin (10,000 U/mL) | Thermo Fisher Scientific | 15140122 | |
Pyrex petri dishes, 100 mm x 20 mm | Millipore Sigma | CLS3160102 | |
Corning Falcon Easy-Grip Tissue Culture Dishes, 35mm | Fisher Scientific | 08-772A | |
Corning Falcon Easy-Grip Tissue Culture Dishes, 60mm | Fisher Scientific | 08-772F | |
Falcon Standard Tissue Culture Dishes | Fisher Scientific | 08-772E | |
Thermo Scientifc Nunc Cell Culture / Petri Dishes, 35x10mm Dish, Nunclon Delta | Fisher Scientific | 12-565-90 | |
Fisherbrand Standard Disposable Transfer Pipettes, Nongraduated; Length: 5.875 in.; Capacity: 7.7 mL | Fisher Scientific | 13-711-7M | |
Ethyl 3-aminobenzoate methanesulfonate | Millipore Sigma | E10521 | |
Collagenase B | Millipore Sigma | 11088807001 | |
Dumont #55 Forceps, Dumostar | Fine Science Tools | 11295-51 | |
Dumont #5 Forceps, Dumostar | Fine Science Tools | 11295-00 | |
Cellattice Micro-Ruled Cell Culture Surface | Nexcelom Bioscience | CLS5-25D-050 | |
For Calcium Imaging | |||
Fluo-4, AM, cell permeant | Thermo Fisher Scientific | F14201 | |
Pluronic F-127, 0.2 µm filtered (10% Solution in Water) | Thermo Fisher Scientific | P6866 | |
For RNA Probe Generation | |||
PureYield Plasmid Miniprep System | Promega | A1222 | |
rATP | Promega | P1132 | |
rCTP | Promega | P1142 | |
rGTP | Promega | P1152 | |
rUTP | Promega | P1162 | |
Digoxigenin-11-UTP | Millipore Sigma | 3359247910 | |
Rnase Inhibitor | Thermo Fisher Scientific | N8080119 | |
T3 RNA Polymerase | Promega | P2083 | |
T7 RNA Polymerase | Promega | P2075 | |
SP6 RNA Polymerase | Promega | P1085 | |
RQ1 Rnase-Free Dnase | Promega | M6101 | |
LiCl Precipitation Solution (7.5 M) | Thermo Fisher Scientific | AM9480 | |
For Fluorescence In Situ Hybridization | |||
Acetic Anhydride | Thermo Fisher Scientific | 320102 | |
Blocking Reagent | Millipore Sigma | 11096176001 | |
Anti-Digoxigenin-POD, Fab fragments | Millipore Sigma | 11207733910 | |
Cy3 Mono-Reactive NHS Ester | Millipore Sigma | GEPA13105 | |
Solution Components | |||
Calcium chloride, 96% extra pure, powder, anhydrous, ACROS Organixs | Fisher Scientific | AC349610 | |
Calcium chloride dihydrate | Millipore Sigma | C3306 | |
CHAPS hydrate | Millipore Sigma | C3023 | |
Denhardt's Solution (50X) | Thermo Fisher Scientific | 750018 | |
DTT, Molecular Grade (DL-Dithiothreitol) | Promega | P1171 | |
Ethylenediaminetetraacetic Acid, Disodium Salt Dihydrate | Fisher Scientific | S311 | |
Ethylene glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid | Millipore Sigma | E3889 | |
Formamide (Deionized) | Thermo Fisher Scientific | AM9342 | |
Herparin sodium salt from porcine intestinal mucosa | Millipore Sigma | H3393 | |
HEPES (Ultra Pure) | Thermo Fisher Scientific | 11344041 | |
Hydrogen peroxide solution | Millipore Sigma | H1109 | |
L-Cysteine | Millipore Sigma | 168149 | |
Magnesium chloride, pure, ACROS Organics | Fisher Scientific | AC223211000 | |
Magnesium sulfate, 97% pure, ACROS Organixs, anhydrous | Fisher Scientific | AC413480050 | |
Maleic Acid, 99%, ACROS Organics | Fisher Scientific | ACS125231000 | |
MOPS (Fine White Crystals/Molecular Biology), Fisher BioReagents | Fisher Scientific | BP308 | |
Potassium chloride | Millipore Sigma | P9541 | |
Ribonucleic acid from torula yeast, Type IX | Millipore Sigma | R3629 | |
Sodium chloride | Millipore Sigma | S7653 | |
Triethanolamine | Millipore Sigma | 90279 | |
Tris | Millipore Sigma | GE17-1321-01 | |
TWEEN 20 | Millipore Sigma | P9416 | |
Equipment | |||
Laminar Flow Hood | model of choice | ||
Dissecting Microscope | model of choice | ||
Inverted Fluorescence Microscope | Nikon | TE200 | |
NIS-Elements Imaging Software | Nikon | ||
Shaking Incubator | model of choice | ||
Refrigerated Centrifuge | model of choice | ||
Miscellaneous | |||
Corning bottle-top vaccum filter system, 0.22 μm pore, 500 mL bottle capacity | Millipore Sigma | CLS430769 | |
Falcon 50mL Conical Centrifuge Tubes | Fisher Scientific | 14-432-22 |