Foetal neural progenitors contribute to postnatal circuits formation ex vivo: an electrophysiological investigation

Neuronal progenitor cells (NPC) play an essential role in homeostasis of the central nervous system (CNS). Considering their ability to differentiate into specific lineages, their manipulation and control could have a major therapeutic impact for those CNS injuries or degenerative diseases characterized by neuronal cell loss. In this work, we established an in vitro co-culture and tested the ability of foetal NPC (fNPC) to integrate among post-mitotic hippocampal neurons and contribute to the electrical activity of the resulting networks. We performed extracellular electrophysiological recordings of the activity of neuronal networks and compared the properties of spontaneous spiking in hippocampal control cultures (HCC), fNPC, and mixed circuitries ex vivo. We further employed patch-clamp intracellular recordings to examine single-cell excitability. We report of the capability of fNPC to mature when combined to hippocampal neurons, shaping the profile of network activity, a result suggestive of newly formed connectivity ex vivo.


Introduction
Neuronal Replacement Therapy (NRT) directly addresses neurodegeneration by replacing dead cells with healthy new ones to restore compromised brain functions. Among NRT candidates, multipotent neural stem cells derived from embryonic tissue are the most promising. However, the mechanisms underlying the CNS functional improvements by NRT are not yet completely understood [1]. For this reason, we used a mixed population of foetal neural progenitor cells (fNPC) and cultured neurons dissociated from the rat hippocampus, as a new in vitro experimental model.

Cultures preparation
Primary HCC were obtained from Wistar rats at P2-3, as in [2]. fNPC were isolated from (E15) embryos, following [3]. pCCLsin.PPT.hPGK.EGFP.Wpre [4], a thirdgeneration self-inactivating (SIN) lentiviral vector, was employed to express the enhanced Green Fluorescence Protein (eGFP) in fNPC, under the control of phosphoglycerate kinase (PGK) promoter. Generation and titration of the vector were as described [5] and transduction took place the same day of dissection. Mixed cultures were obtained by blending together fNPC and hippocampal neurons in 1:1 ratio in order to keep the same density as controls (fNPC and HCC). Thus, in all three conditions, cells were plated at a density of 420 cells/ mm 2 on 0.01% poly-L-ornithine solution-coated glass coverslips (Orsatec GmbH, Bobingen, Germany) or at a density of 3000 cells/mm 2 on the inner area of MEAs, after coating with a 0.1% polyethylene-imine solution.
Pair recordings were performed by simultaneously patching two neurons, visualized in the same 40× microscopy field, alternatively holding one cell in currentclamp and the other in voltage-clamp mode. Brief current pulses were used to elicit an action potential in one cell, while monitoring the other (putative) postsynaptic cells for postsynaptic currents (PSCs).

Immunofluorescence
Labelling with β tubulin III and DAPI was performed as described [7].

Data analysis and statistics
MEA data were analysed offline as in [8]: briefly, raw voltage traces were bandpass filtered (0.2-3 kHz) and AP extracted as peaks exceeding 5 times the signal median [9]. Bursts of synchronous AP were later identified as episodes of threshold crossing for the instantaneous product between the AP count and the active-electrodes count, binned at 25 ms [10]. Samples subjected to the same conditions were pooled together and their values expressed as mean ± SEM with n = number of cells or of MEAs. Statistical analysis was performed using Graph-Pad Prism 6 (GraphPad Software, San Diego, USA). Depending on whether assumptions were met for parametric testing, we used either one-way ANOVA followed by a Bonferroni-corrected multiple comparisons procedure or Kruskal Wallis analysis followed by Dunn's multiple comparisons test. Significant differences were determined at P < 0.05. The convention to indicate the value of P is clearly indicated in the legend of Fig. 1. Numerical data were depicted (Fig. 1c-h) as box-andwhisker plots where the thick horizontal bar indicates the median value, the boxed area extends from the 25th to 75th percentiles while whiskers from the 10th to the 90th percentiles.

Results
We compared three conditions: HCC (n = 8), fNPC (n = 8), and mixed cultures (n = 8). HCC and eGFP-tagged fNPC reorganized ex vivo into interconnected networks ( Fig. 1). Bursts of AP, occurring synchronously in cells across the network were detected by distinct MEA microelectrodes (Fig. 1b). These bursts emerge due to excitatory and inhibitory synaptic transmission interplay and intrinsic cell properties [11]. Pharmacological blockade of GABA A receptors by BIC, altered network activity patterns (not shown). When cultured alone, fNPC displayed rare or no spiking activity, even under BIC.
The number of microelectrodes detecting repeated extracellular AP was then taken, for each MEA (n = 8 per condition), as a measure of synaptic and neuronal viability across the network (not shown). We observed a much higher number of active electrodes in mixed (45.0 ± 4.69) and HCC (38.5 ± 2.83) than in fNPC cultures (5.00 ± 1.95). Under BIC, the number of active electrodes was still higher in mixed (46.6 ± 3.26) and HCC (40.9 ± 5.27) than in fNPC cultures (5.63 ± 1.82). Next, we quantified the frequency of occurrence of spontaneous AP for every culture (Fig. 1c-d). We observed that the AP frequency differed, although not significantly, comparing HCC (76.7 ± 22.6 Hz) and mixed cultures (151 ± 32.2 Hz). Instead, the difference was significant both with HCC and mixed cultures, compared to fNPC cultures (2.66 ± 1.44 Hz).
The emergence of AP episodic synchronization, a well-known correlate of synaptogenesis [12,13], was studied by analysing the rate of occurrence of AP "bursts", across each MEA (Fig. 1e). Interestingly, fNPC cultures never displayed bursting. Burst rate was significantly higher in mixed cultures (0.547 ± 0.131 Hz) than in HCC (0.163 ± 0.063 Hz). Burst lengths in mixed cultures (296.7 ± 67.3 ms) and HCC (186.3 ± 58.4 ms) did not differ significantly (p = 0.27). Under BIC, no bursting occurred in fNPC cultures (Fig. 1f), while a very similar degree of episodic synchronization was found in mixed (0.615 ± 0.094 Hz) and HCC (0.410 ± 0.109 Hz). Moreover, BIC had a significant effect on burst rate in HCC but not in mixed cultures.
We then probed the active membrane properties upon injecting current pulses, while monitoring the membrane potential. All cell types responded with AP. However, despite similar peak amplitudes (average 100.8 ± 1.8 mV), AP rise time was significantly faster in fNPC (2.1 ± 0.1 ms; n = 63) than among all the other experimental conditions, while a significative different degree of maturation could be observed, comparing these values to fNPC in mixed cultures (2.6 ± 0.1 ms; n = 44). HCC responded with AP rise time of 2.8 ± 0.1 in all conditions.
Finally, we performed simultaneous pair recordings (n = 20) in mixed cultures to characterise the monosynaptic connectivity between fNPC and HCC ( Fig. 1i-j). In 50% of the cases HCC neurons synapsed to fNPC and in 25% of the cases a reciprocal connection was found (Fig. 1i).

Discussion
We evaluated whether fNPC integrate among and communicate with post-mitotic hippocampal neurons, by establishing a novel in vitro co-culture system and electrophysiological assay. Our recordings suggest the formation of functional hybrid networks differing from control cultures, resulting in qualitatively and quantitatively different emerging spontaneous AP synchronization.
Voltage-clamp intracellular recordings indicated a significant alteration of cell membrane passive properties, as fNPC grew in mixed cultures compared to control cultures, implying overall a maturation through a significant increase in cell size and number of ion channels of the cell membranes. Current-clamp experiments further suggest cellular maturation, as fNPC's excitability was examined in response to a depolarizing stimulus. Pair recordings further supported our conclusions, by documenting the synaptic integration of fNPC in HCC microcircuits. Finally, it is tempting to speculate that the effect of BIC on the burst rate links fNPC differentiation to the excitatory/inhibitory balance of network activity, a possibility that is also supported by the lack of significant BIC effects on mixed cultures. This might be reinforced by the role of the reciprocal synaptic connectivity where fNPC are involved. Further studies, involving immunohistochemistry and an expansion of our paired-cell recordings performed here would still be needed to confirm these hypotheses.