Introduction

The conservation of biological processes and mechanisms across various species, safety and ethical considerations, improvements in transgenic tools and increasing affordability have contributed to the growing use of animal models in translational biomedical research1,2. The principal disadvantage is that animal studies fail to capture faithfully the complexity of human conditions and disease mechanisms2,3. Over the recent decade, there has been a growing interest in utilizing human cells and tissue in biomedical research, with increasing applications of progenitor cells from embryos4 and stem cell-derived models5. Indeed, studies in human cell cultures provide a better approximation of molecular biology and genetics of human diseases and development, with excellent control of experimental conditions and accessibility6. The principal downside of these models is that they are overly simplified and come with ethical issues related to the use of human material, along with potential legal and regulatory challenges7,8. Furthermore, the research involving human pluripotent stem cells (hPSCs) and induced PSC-derived models is labour-intensive and expensive9.

Immortalized human cell lines have attracted interest as potential alternatives, with increasing numbers of reports using NT2, PC12, N1E-115, NG108-15, and SH-SY5Y cell lines to address a range of mechanistic and translational questions10,11,12,13,14. Robustness, affordability, and genetic continuity have made some of these lines attractive to the field of neuroscience research. Derived from the SK-N-SH progenitor of a bone marrow biopsy of the neuroblastoma of a child, the immortalized SH-SY5Y line is widely used to address a range of neurochemical15,16, molecular16,17,18 and developmental19 neurobiology and translational questions20,21. In cultures, these cells show excellent viability and can be readily induced into neurons, exhibiting characteristic morphological and neurochemical phenotypes17,18,22. SH-SY5Y cells are robust, sustain homogeneity, differentiate into several neuron types if conditions are favourable, and can be maintained at low costs17,23. Commercial accessibility of human SH-SY5Y cells eliminates ethical or legal hurdles, making the cell line one of the most widely used in vitro neuronal models17,18,24,25.

Developmental studies demonstrated reliable sequential differentiation of SH-SY5Y cells into morphologically, neurochemically and molecularly mature neurons17,18,24. Analysis of markers confirms their acquisition of dendrites and axons, which form specialized synaptic contacts17,26. There is, however, a significant gap in functional data and demonstration of activity dynamics of SH-SY5Y cells during differentiation27. Since neuronal development and wiring entail complex electrical and neurochemical activity and intercellular signaling, it is critical to elucidate the functional alterations of developing SH-SY5Y cells to relate structural and molecular changes with underlying processes and mechanisms28. Calcium imaging has proven highly instructive, with spontaneous and evoked transients measuring neuronal excitability and various activity states. The dynamics of Ca2+ play a pivotal role in multiple signaling processes, regulating differentiation and developmental mechanisms, cellular homeostasis and plasticity29,30.

In this study, we combined immunofluorescence confocal microscopy with live Ca2+ imaging of sequentially differentiating SH-SY5Y cell-derived human neurons17,18,23 to elucidate the relationship between evoked and spontaneous calcium dynamics with structural and molecular transformation of the progenitors into mature neurons. The rationale for analyzing various differentiation stages is to determine the dynamics and relationship of functional changes associated with transformation of SH-SY5Y cells into neurons exhibiting distinct molecular and morphological characteristics.

Materials and methods

Materials and reagents

SH-SY5Y cell line (CRL-2266, ATCC), Dulbecco’s modified eagle medium (DMEM, Thermo 41965-039), neurobasal-A (10888022 Thermo), N2 (11520536 Thermo), trypsin (25200056 Thermo), fetal bovine serum (16000044, Thermo), mounting media with DAPI (F6057 Sigma), penicillin streptomycin solution (P0781 Sigma), phosphate buffered saline (PBS, 18912014, Thermo), L-glutamine (25030032, Thermo) Nunc™ cell-culture treated multi-dishes (140675, Thermo), T-25 cell-culture treated flask (156367, Thermo), DMSO (32160405, Sigma), paraformaldehyde 16% (043368.9 L Thermo), carbachol (C4382-1G Sigma), ionomycin (124222 Thermo), brain derived neurotrophic factor (BDNF, 248-BDB-010/CF, Biotechne), retinoic acid (RA, 0695/50 Biotechne), Corning® Costar™ 24-well clear TC-treated multiple well plates (10380932, Fisher Scientific).

Culturing and differentiation of neuroblastoma cells

The frozen SH-SY5Y cells from ATCC were thawed in T25 culture flasks in a standard growth medium containing DMEM. The flusk media of all cells utilized for structural and molecular differentiation studies were supplemented with 10% FBS (Table 1); the media of cells cultured for live imaging experiments was supplemented with 15% FBS (Table 2). Both groups also contained 1% ATB and were maintained at 37 °C in 5% CO2. Upon reaching 80–90% confluence, cells were treated with a trypsin-EDTA solution, plated at the desired seeding density and split into multiple groups (Tables 1 and 2), with cell counting made using a haemocytometer.

Table 1 Representation of the 12-day timeline of differentiation of SH-SY5Y cells used for immunostaining, with details of experimental conditions.
Full size table
Table 2 Representation of the 12-day timeline of differentiation of SH-SY5Y cells used for the calcium imaging experiment with details of experimental conditions.
Full size table

As detailed in Tables 1 and 2 and discussed in our earlier report17, at day 0, SH-SY5Y cells were seeded at a 100,000-cell density per well in 24-well polymer plates in the medium containing DMEM31. Cultures were maintained at 37 °C with 5% CO2. We set up three groups of cultures for studying morphological and molecular differentiation (Table 1) and four groups for live calcium imaging (Table 2). The first group includes (1) control cultures of undifferentiated SH-SY5Y cells in DMEM with gradual FBS withdrawal from 5 to 0.5% with 1% ATB over 12 days; (2) RA-treated SH-SY5Y cells exposed to RA (10µM) in DMEM medium with FBS starvation from 5 to 0.5% over 12 days in the presence of 1% ATB and (3) RA + BDNF-treated SH-SY5Y cells exposed to a RA (10µM) for 6 days in DMEM medium with gradual FBS withdrawal from 5 to 2% for 6 days, followed by addition of BDNF (50ng/ml) for 6 days in Neurobasal medium supplemented with 1% N2, L-glutamine (2mM) and 1% ATB (Table 1)17. The second group set for calcium imaging, we compared actively proliferating cells maintained in 15% FBS, with FBS-starved controls, RA and RA-BDNF groups (Table 2). The adopted experimental protocol enables the determination of the effects of FSB starvation and differentiation of SH-SY5Y cells into neurons on Ca2+ dynamics. Experiments were conducted in triplicate with conditions kept consistent to ensure reproducibility of the results.

Immunofluorescence staining of neuroblastoma cells

On day 12, cells were washed three times in 24-well plates with phosphate-buffered saline pH 7.4 (1X PBS) and fixed using 4% PFA in 1X PBS for 20 min, followed by 3 × washes with PBS and permeabilized with 0.1% Triton X-100 in 1X PBS for 10 min at room temperature. The step was followed by 3 × washes with PBS and binding by 2% FBS in PBS for 1 h at RT. Fixed cells were then incubated overnight with primary antibodies at 4 °C, which were subsequently removed, and after 3 × washes, treated with secondary antibodies overnight at 4 °C. Afterwards, secondary antibodies were removed, and cells were washed 3 × with PBS, followed by covering with a mounting medium containing nuclear stain DAPI. After drying and cover slipping, the samples were viewed with a Confocal (ZEISS LSM 880) microscope in epifluorescence mode. Confocal images were acquired on the same microscope in laser scanning mode with a 40 × NA = 1.3 oil immersion Zeiss objective. All experiments were conducted in triplicate, with conditions kept constant. For excitation of fluorescent-labelled secondary antibodies, we used 488 nm (Tau and PSD95), 568 nm (BT and Synapsin) and 647 nm (Map2) laser lines with images sampled in the same mode at a spatial resolution of 30 nm per pixel and dwell time of 1.5 µs.

Live calcium imaging

On day 12 (Table 2), the media of cell cultures were replaced with Fluo-4 (2µM) containing media for 30 min before imaging. After washing cells with PBS, fresh Fluo-4-free media was added, and the plates were mounted on the microscope stage for live Ca2+ imaging inside the imaging chamber with a controlled environment. Micrographs were captured in time series mode at 1 Hz on ZEISS LSM 880 confocal laser scanning microscope with a 40 × oil immersion objective. A 488 nm laser was used for excitation of the fluorophore. In experiments that monitored spontaneous Ca2+ oscillations, recordings were made continuously for over 400 s without stimulants. In studies of evoked Ca2+ transients, all conditions were kept the same except that after collecting 100 frames, cholinergic agonist carbachol (50 µM)32 or Ca2+ ionophore ionomycin (5 µM)33,34 was added, followed by recording over 300 s. All collected data were analyzed offline using Image-J (NIH Image-J, Version Win64). Data processing and analysis were carried out using IBM SPSS Statistics software (Version v29.0.1.0). Analysis of the frequency of spontaneous Ca2+ oscillations and amplitude-kinetic parameters of the evoked Ca2+ transients was carried out using Easy Electrophysiology software (https://www.easyelectrophysiology.com/) in semiautomatic or automatic modes. Individual transients of Ca2+ were detected by setting an empirical threshold to capture events exceeding the baseline noise intensity deviations by a factor of 335. The reliability of the detection of individual events was verified manually and compared with that of those detected using a threshold-based algorithm.

Image processing, data analysis and presentation

Experimental data were collected from triplicate studies. Sholl analysis was used to count systematically the number of neurites and their branches as reported in our earlier paper36. The number of putative synaptic contacts was counted manually over normalised areas within 15–50 μm range of the soma. For fluorescence intensity measurements, we used multiple regions of interest (ROIs), with numerical values extracted for offline analysis with IBM SPSS. A similar approach was used for colocalization analysis, applying ImageJ JACoP colocalization plugin. All numerical values (i.e. Pearson’s correlation coefficient and Mander’s colocalization coefficient) were extracted and analyzed offline using IBM SPSS Statistics software. The results are presented as mean SE, with the statistical significance determined using Student’s t-tests or Tukey’s post hoc tests, with p < 0.05 defined as significant. Figures were prepared using Adobe Illustrator Artwork 16.0 in the Adobe Creative Suite Version 6 program. EndNote X8.2 was used for reference formatting per journal guidelines.

Results

General characterization of SH-SY5Y cell cultures

SH-SY5Y cells were maintained under different conditions to induce various developmental stages of neurons for morphological, molecular and functional studies (see Methods and Tables 1 and 2). Cultures retained in control media are taken as non-differentiated controls. RA treatment of SH-SY5Y cells induced the differentiation process, leading to morphological changes like neurite outgrowth and a decrease in cell proliferation15,17,23,37. These alterations result from activation of retinoic acid receptors and their downstream targets, promoting the development of neuron-like phenotypes. Finally, exposure of cultures to a cocktail of RA with BDNF has been reported to promote the development of specific neuronal characteristics such as differentiation of neurites into axons and dendrites and formation of synaptic contacts17,18,23,37,38.

For structural analysis of cultures at different stages of differentiation, they were fixed with PFA and imaged in brightfield mode. Figure 1 summarises the results of these studies. Cultures maintained for 12 days in control media (Table 1) display large and irregular soma with sparse and short processes. Most cells in these cultures assemble in dense groups, indicating their active proliferation state (Fig. 1A1 and A2). Cultures treated with RA for 12 days exhibited smaller soma with distinctly elongated processes making occasional contacts (Fig. 1B1, Fig. 2). Finally, cells in cultures after 6 days of RA treatment exposed to RA + BDNF for six additional days showed the morphology of a typical neuron with more regular shape of soma, extensive branching of processes, which formed numerous contacts and varicosities (Fig. 1C1, Fig. 2). The results of counting the neurites and the number of contact points over normalized areas are summarized in (Fig. 1D–F). As can be seen, the control group developed fewer outgrowths per soma, whereas in RA and, especially, RA + BDNF group, the number of neurites per soma increased significantly (Fig. 1D). The higher incidence of neurite crossover in RA-BDNF differentiated cultures was subject to further analysis. Counting the number of contacts within randomly defined ROIs (15–50 m range of the soma) showed that the number of contacts is highest in cultures treated with RA + BDNF (× 20 controls) followed by RA (× 4 controls) treated cultures (Fig. 1E). Analysis of the relationship between the number of neurites and contacts showed a weak but significant correlation in control cultures. In RA and RA-BDNF-treated cultures, however, the correlation between these two variables was stronger and highly significant (Fig. 1F).

Fig. 1
figure 1

Morphological characterization of SH-SY5Y cell cultures. (A1–C2) Typical brightfield images of 12-day-old Control (A1-2), RA (B1-2) and RA + BDNF (C1-2) cultures. The top row has a wide field of view; the bottom row has an expanded view of boxed areas (dashed boxes). White arrows point to neurites; red arrowheads point to putative contacts. Scale bars top, 250 μm, bottom 40 μm). (D,E) Summary histograms of the number of neurites per soma and number of contacts over normalised areas, respectively, in Control, RA and RA + BDNF cultures. Histograms show means and ± SEM. N = 100 per group, three replicas. (F) The scatter plot shows the relationship between the number of contacts and the number of neurites per cell in Control, RA and RA + BDNF cultures.

Full size image
Fig. 2
figure 2

BT–Tau expression and distribution in SH-SY5Y cell cultures. (A1–C3) Typical confocal immunofluorescence micrographs of 12-day-old Control, RA and RA + BDNF cultures stained for β-tubulin (BT) and Tau protein (red and green, respectively). Blue - counterstaining of cells with DAPI. Scale bars: 80 μm. (D,E) Summary histograms of the intensity of staining for β-tubulin and Tau protein (red and green, respectively) in soma and neurites (SOMA and NEURITES) showing the distribution of signal intensity in Control, RA and RA + BDNF cultures. Histograms show ± SEM. N = 30 ROI per group, three replicas. (F) The bar graphs show the Pearson correlation values of β-tubulin and MAP2 protein in soma and neurites of control and RA-BDNF differentiated cells. Test for the significance of the relationship between two variables of each dataset with a paired Student t-test produced p < 0.05 for all datasets.

Full size image

Dynamics of structural proteins at various stages of developing SH-SY5Y cells

Next, we performed the immunostaining of SH-SY5Y cells at different developmental stages for generic and specific neuronal proteins. As above, after reaching the desired differentiation stages, cultures were fixed but stained for axonal tau, dendritic MAP-2 and skeletal protein β-tubulin39,40,41. The preparation of cells for staining was as described above (see Methods). Signal intensity and colocalization analysis were carried out using intensity or colocalization plugins of Image-J. Figure 2 shows representative micrographs and summarises the results of intensity and colocalization studies. Comparison of staining of three sets of 12-day-old undifferentiated cultures revealed a trace amount of β-tubulin and tau protein in the soma and rudimentary neurites of SH-SY5Y cells (Fig. 2A1–3,D, E). The expression levels of these proteins in cultures treated with RA or a cocktail of RA + BDNF showed a significant increase, with their levels in cultures exposed to RA and BDNF exceeding those of RA-treated and control cultures (Fig. 2B1-3-E). In both sets of cultures, there was a prominent rise in the expression of β-tubulin and tau in neurites, which was particularly notable in RA + BDNF differentiated cultures. Analysis of the colocalization of these structural proteins showed a lower index in outgrowths and soma of cells in control cultures as compared to RA + BDNF cultures (Fig. 2F). Similar studies of β-tubulin and MAP2 also revealed increase in their expression associated with differentiation. The MAP2 level was significantly higher in both soma and neurite of cells in RA-treated cultures as compared to RA-BDNF and controls (Fig. 3A1–E). Notably, the baseline colocalization of β-tubulin and MAP2 in controls was higher than the colocalization of β-tubulin with tau, with the colocalization index of MAP2-β-tubulin indistinguishable between controls and RA-BDNF-treated cultures (Figs. 2F and 3F). The results of differentiation-dependent changes in tau, MAP2 and β-tubulin are likely to reflect their developmental engagement in various structural elements and processes in different cellular compartments42.

Fig. 3
figure 3

Quantification of BT – MAP2 expression and distribution in SH-SY5Y cell cultures. (A1–C3) Typical confocal immunofluorescence micrographs of 12-day-old Control, RA and RA + BDNF cultures stained for β-tubulin (BT) and MAP2 protein (red and green, respectively). Blue - counterstaining of cells with DAPI. Scale bars: 80 μm. (D,E) Summary histograms of the mean intensity of staining for β-tubulin and MAP2 protein (red and green, respectively) in soma and neurites (SOMA and NEURITES) showing their distribution in Control, RA and RA + BDNF cultures. Histograms show ± SEM. N = 30 ROI per group, three replicas. (F) The bar graphs show the Pearson correlation values of β-tubulin and MAP2 protein in soma and neurites of control and RA-BDNF differentiated cells. Analysis of the significance of the relationship between the two variables of each dataset with a paired Student t-test produced p < 0.05 for all datasets.

Full size image

Visualizing synaptic markers in SH-SY5Y-derived neurons

The development of an elaborate network of neurites with abundant contacts in RA + BDNF differentiated SH-SY5Y cultures suggests synaptic wiring of cells. To confirm formation of synaptic connections in RA + BDNF differentiated cultures, we have carried out immunostaining on 12-day-old SH-SY5Y cells for pre and postsynaptic markers and dendrite-specific protein MAP217,36. Figure 4 demonstrates representative confocal micrographs and summarises the results of the colocalization studies (A-D and E, F, respectively). We used a combination of immunostaining for MAP2 and ubiquitous presynaptic protein synapsin, as well as a combination of MAP2 and a key element of the postsynaptic scaffold of glutamatergic synapses, PSD9536,43. In all double-labelled specimens, MAP2-PSD95 and MAP2-synapsin expression was visible in soma and neurites, with abundant juxta-cellular PSD95 or synapsin positive puncta and varicosities also visible (Fig. A1–A3 and B1–B3). While PSD95 puncta were mainly positioned at (or colocalized with) MAP2, there were also numerous PSD-95 or synapsin-positive puncta unrelated to MAP2. Closer scrutiny of double-labelled puncta showed organized arrangements of synaptic markers with MAP2, with the latter known to be associated with dendrites36,44. To more rigorously characterize the labelling of putative synapses, after visualizing morphological contacts on brightfield mode, they were scanned in confocal mode, with colocalization analysis conducted using ROIs as specified in Methods (Fig. 4E, F). Measurements of the colocalization index of PSD95, synapsin, and MAP2 revealed a high degree of superposition between MAP2 and PSD95, and a lesser degree of superposition between MAP2 and synapsin in the ROIs corresponding to neurite contacts. This agrees with PSD95 and MAP2 being enriched in dendritic compartments of SH-SY5Y-derived human neurons45. The results of our confocal microscopy of SY-SY5Y-derived neurons confirm effective differentiation of RA + BDNF-treated SH-SY5Y progenitors into neurons expressing specific synaptic markers. The nature of neurotransmission and the mediators operating at contacts is a subject of our ongoing experiments.

Fig. 4
figure 4

Expression and localisation of synaptic proteins in RA + BDNF-differentiated cultures. (A1–B3) Typical confocal immunofluorescence micrographs of 12-day-old RA + BDNF cultures triple-stained for synapsin, MAP2 and DAPI (red, green and blue, respectively). Arrowheads point to putative synaptic contacts. Scale bars: 80 μm. (C,D) Zoomed-in image of putative synaptic contacts double-stained for synapsin-MAP2 and PSD95-MAP2, respectively. Scale bars: 3 μm. (E,F) Summary histograms of the colocalization index of synapsin-MAP2 and PSD95-MAP2 show the degree of colocalization of these specific neuronal markers. Histograms show ± SEM. N = 30 ROI per group, three replicas. Analysis of the significance of the relationship between the two variables of each dataset with a paired Student t-test produced p < 0.05 for all datasets.

Full size image

Spontaneous and evoked calcium dynamics in SH-SY5Y cells

To investigate if structural and molecular differentiation of SH-SY5Y cells into neurons is accompanied by activity changes, we studied spontaneous and evoked Ca2+ dynamics using live imaging with Fluo-4 Ca2+ sensor. We compared Ca2+ dynamics of undifferentiated SH-SY5Y cultures with starved SH-SY5Y cultures or cultures treated with RA or RA + BDNF (Table 2, see also Methods). We observed significant differences in spontaneous Ca2+ dynamics with the switch of activity depending on the differentiation state. Figure 5 summarises the results of these experiments with corresponding frequency, amplitude and dye uptake distribution graphs. In control SH-SY5Y cultures, over ~ 70% of cells showed high-amplitude slow oscillation of Ca2+, with the remaining cells exhibiting random transients or keeping in a resting state (Fig. 5A, E–G).

Fig. 5
figure 5

Live imaging of spontaneous Ca2+ dynamics in undifferentiated controls, FBS-starved, RA and RA-BDNF differentiated cells. (A–D) Raw recording traces of Ca2+ dynamics from the soma of three representative cells from each experimental group were collected over 400 s in time series recording mode. Slow high-amplitude oscillations in undifferentiated SH-SY5Y cells (Controls) were rendered less regular and of variable amplitudes by FBS starvation (Control-Starvation) (A,B). (C) Treatment of cells in RA abolished the oscillatory activity, leaving sparse low-amplitude transients, while (D) cultures maintained in RA + BDNF showed random and low-amplitude Ca2+ transients. (E–G) Summary histogram of the amplitude and recorded number of events (frequency) in four experimental groups: Control undifferentiated with 15% FBS (white), Control FBS Starved (ST, red), RA (blue) and RA + BDNF (black). (G) Histograms of the initial dye uptake in four experimental groups. Histograms show ± SEM. N = 3 per group, with at least 30 cells in each group from three repeats.

Full size image

Given that FBS deprivation of SH-SY5Y cells promotes their differentiation and increases the expression of neuronal proteins46,47, we investigated whether a switch in Ca2+ dynamics accompanies these changes. In cultures under starvation, high-amplitude regular Ca2+ oscillations switched into less regular, low-amplitude transients, with occasional sharp spikes superimposed onto different phases of slow Ca2+ waves (Fig. 5A, B,E, F). In the RA group, we observed a dramatic reduction of Ca2+ activity, with elimination of high-amplitude spontaneous Ca2+ waves (Fig. 5C, E and F). The level of baseline noise was, however, elevated, with occasional low-amplitude transients visible. Importantly, in 12-day-old cultures differentiated with RA + BDNF, we observed a resurgence of low amplitude and random Ca2+ transients (Fig. 5D–F). Measurements of the Fluo-4 uptake by SH-SY5Y cells under different experimental conditions revealed considerable differences between various experimental groups, with the uptake of Fluo-4 highest in RA + BDNF-treated cultures and starved groups, as opposed to the lowest uptake in RA-treated SH-SY5Y cultures (Fig. 5G).

To determine if different conditions of SH-SY5Y cultures also affect the stimulation-induced Ca2+ transients, we performed live calcium imaging before and under stimulation of cells with carbachol and ionomycin (Fig. 6). Both compounds are known to induce rise of intracellular Ca2+ but via different mechanisms34,48,49. The choice of concentrations of carbachol and ionomycin was based on published literature34,50. Figure 6A–D presents typical recordings of evoked Ca2+ transients with corresponding summary graphs of the amplitude distribution. As can be seen, the application of carbochol in all experimental sets of cultures induced a rapid rise of Ca2+ followed by its decline. Analysis of the evoked response amplitude showed that the peak response was highest in RA + BDNF-treated cultures, with amplitudes of control, starved and RA groups remaining within a similar range (Fig. 6A, C). There were also subtle differences in the kinetics of the Ca2+ transient decay, with RA and RA + BDNF differentiated cultures showing faster decline (Fig. 6A). The response of cells to ionomycin was consistent across all groups, with evoked Ca2+ transients displaying a slower rise with a much broader peak and slower decline (Fig. 6B, D).

Fig. 6
figure 6

Live imaging of Ca2+ transients in undifferentiated controls, FBS starved, RA and RA-BDNF differentiated cells evoked by carbachol and ionomycin. (A,B) Typical recordings of evoked by carbochol (A) and ionomycin (B) Ca2+ transients in Undifferentiated SH-SY5Y cells with 15% FBS, Starved (ST) RA and RA + BDNF. The grey bar above recordings indicates the time of the application of stimulants. Note the amplitude and kinetics differences of evoked Ca2+ transients in different experimental groups. (C,D) Box plots summarise the peak amplitude distribution of evoked response in undifferentiated or 15% FBS, FBS starved, RA and RA + BDNF group cultures stimulated by carbachol or ionomycin, respectively. NS non-significant.

Full size image

Discussion

Neuronal differentiation and maturation play a crucial role in the development and function of the nervous system. Overwhelming data suggests specific histochemical, molecular and biochemical alterations in neurons at various stages of their development26,36,51,52,53,54,55,56,57. However, a significant gap remains in understanding how these changes relate to the developmental adjustments of intrinsic and synaptic activity. In this study, we investigated whether spontaneous and evoked Ca2+ dynamics change in developing SH-SY5Y-derived neurons in vitro. In general, the molecular and morphological characteristics of differentiating SH-SY5Y cells observed in our experiments agree with those reported previously23,46,58,59. The results of live Ca2+ imaging are novel and present a significant milestone in elucidating functional changes accompanying structural and molecular remodeling of neurons derived from SH-SY5Y progenitors.

Like in earlier studies, serum deprivation of SH-SY5Y cells in our experiments set cultures on a differentiation path, initiating polarization of cells with outgrowth of rudimentary neurites 38,47. Supplementing RA in culture media promoted the elongation of neurites and increased the expression of structural proteins such as β-tubulin, MAP2, and tau, arrested G1 and G2 phases, leading to block of cell proliferation15,17,60,61. This process was promoted further by adding BDNF to the cultures, which is known to facilitate the polarisation and differentiation of SH-SY5Y cells17,18,62. The polarization of cells was accompanied by the development of dense branches of neurites with increase in the level of tau protein and synaptic markers. Interestingly, RA alone enhanced the MAP2 signal in SH-SY5Y cells more than in combination with BDNF, whereas the combination of RA and BDNF enhanced both MAP2 and tau expression. A possible explanation for this differential response could be that under RA, the rate of neurite growth is slower than that in the presence of BDNF, leading to stronger enrichment of the soma and rudimentary neurites with MAP2.

Because neuronal differentiation and integration into functional networks depend on their electrical activity, which drives release of signaling molecules57,63,64,65, we investigated Ca2+ dynamics in SH-SY5Y cells at various differentiation stages. Under our settings, the undifferentiated cells show high-amplitude spontaneous Ca2+ oscillations and smaller evoked Ca2+ transients, also reported in other dividing cells66,67. Ample data suggest that the rise in cytosolic free Ca2+ caused by its release from endoplasmic reticulum and mitochondria could play a pivotal role in cell proliferation68,69,70,71. Our imaging experiments show high-amplitude low-frequency Ca2+ oscillations in SH-SY5Y cells in the control environment, where they are maintained in active proliferation state, a process involving the activity of Ca2+-calmodulin complex66,72. The high-amplitude waves of intracellular Ca2+ are essential for the control of cell cycle progression, contributing to several aspects of cell division, including nuclear envelope breakdown and reformation, microtubule activity and cytoskeletal remodelling in the course of cell division73,74,75. They also play a crucial role in controlling molecular motors, driving cell division and migration76,77,78. Importantly, in cultures maintained in serum-deprived conditions, we observed a significant increase in the neurite outgrowth and expression of markers of mature neurons.

The attenuation of large-amplitude Ca2+ oscillations in SH-SY5Y cells under starvation and their nearly complete block under RA treatment agrees with the switch of cells from the undifferentiated state onto the differentiation path30,79. The latter is in line with increased expression of specific structural proteins such as tau, MAP2 and β-tubulin, stabilising neurites and developing neural networks80,81. The abolition of spontaneous high-amplitude Ca2+ oscillations by RA agrees with reports showing inhibition of voltage-gated calcium currents by retinoic acid82,83. It remains to be shown how supression of calcium dynamics contributes to setting SH-SY5Y cells on the path of differentiation into neurons. The remaining random and rapid Ca2+ transients in starved cultures possibly reflect the emergence of rudimentary neuronal activity57,84. With further differentiation of SH-SY5Y cells into neurons by a cocktail of RA + BDNF, the residual Ca2+ oscillations might reflect specific neuronal processes which are likely to contribute to electro-chemical integration of developing neurons into neural networks57,85,86,87.

The results of the analysis of evoked Ca2+ transients by carbachol are consistent with the developmental switch of the signalling mechanisms and functions in SH-SY5Y cells. Low amplitude carbachol-induced Ca2+ transients in undifferentiated and starved cells suggest limited expression of cholinergic receptors and their weaker coupling with intracellular Ca2+ stores88,89,90,91. In contrast, RA + BDNF-treated SH-SY5Y cells showed a response with the highest amplitude, indicating higher expression of cholinergic receptors and their more effective coupling with intracellular Ca2+ stores. The rapid onset of the evoked Ca2+ response is likely to involve also depolarisation and influx of this cation via voltage-gated mechanisms. At the same time, the protracted phase can be attributed to the release of Ca2+ from intracellular stores88,90,91,92. Accordingly, SH-SY5Y cells are known to express nicotinic (α3,5 and 7, and β2 and 4) and muscarinic (M1/2) receptors93,94. Alterations in their levels and signaling mechanisms during differentiation can significantly affect cellular response to acetylcholine and its analogues such as carbochol95. Unlike carbochol acting on cholinergic receptors, ionomycin is a calcium ionophore, which enables passage of Ca2+ across the plasma membrane without activation of specific receptors or transport mechanisms48. The similar amplitude-kinetic characteristics of Ca2+ transients evoked by ionomycin in SH-SY5Y cells across all experimental groups agree with non-specific mechanisms of the induction of calcium transients by this ionophore34,48. Taken together, these results show that the developmental switch of signalling in SH-SY5Y cells involves both spontaneous and evoked Ca2+ dinamics, with potential functional implications.

In conclusion, the results of our experiments demonstrate for the first time that during sequential differentiation of SH-SY5Y cells into human neurons, structural and molecular changes are accompanied by switch of Ca2+ activity. The impact of the changes of intrinsic and evoked Ca2+ dynamics on molecular and structural alterations and their functional significance for SH-SY5Y-derived neurons remains unclear and is a subject of future studies.