Patients and animals
All patients underwent neurosurgical resections because of pharmaco-resistant epilepsy or tumor removal. Written informed consent to use spare neocortical tissue acquired during the surgical approach was obtained from all patients. The study was reviewed and approved by the local ethic committee (EK067/20). All ethical regulations relevant to human research participants were followed. For this study, we collected data from 21 patients (14 females, 7 males; age ranging from 8 to 75 years old) (Supplementary Tab. S1). The cases were meticulously selected to fulfill two main criteria: 1) availability of spare tissue based on the needed surgical approach; and 2) normal appearance of the tissue according to radiological and intraoperative criteria (absence of edema, absence of necrosis, and sufficient distance to any putative intracerebral lesion). In addition, samples from tumor cases were neuropathologically reviewed to rule out the presence of tumor cells in the examined neocortical specimen.
All experimental procedures involving animals were performed in accordance with the guidelines of the Federation of European Laboratory Animal Science Association, the EU Directive 2010/63/EU, and the German animal welfare law. In this study, Wistar rats (Charles River, either sex) aged 40–55 postnatal days were anesthetized with isoflurane and then decapitated. Rats were obtained from Charles River and kept under a 12-h light–dark cycle, with food and water available ad libitum.
Slice preparation
Human cortex was carefully micro-dissected and resected with minimal use of bipolar forceps to ensure tissue integrity. Resected neocortical tissue from the temporal or frontal cortex was directly placed in an ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM): 110 choline chloride, 26 NaHCO3, 10 D-glucose, 11.6 Na-ascorbate, 7 MgCl2, 3.1 Na-pyruvate, 2.5 KCl, 1.25 NaH2PO4, and 0.5 CaCl2) (325 mOsm/l, pH 7,45) and transported to the laboratory. Slice preparation commenced within 10 min after tissue resection. The pia was carefully removed from the human tissue block using forceps and the pia-white matter (WM) axis was identified. 300 µm thick slices were prepared using a Leica VT1200 vibratome in ice-cold ACSF solution containing 206 mM sucrose, 2.5 mM KCl, 1.25 mM NaH2PO4, 3mM MgCl2, 1 mM CaCl2, 25 mM NaHCO3, 12mM N-acetyl-L-cysteine, and 25 mM glucose (325 mOsm/l, pH 7,45). During slicing, the solution was constantly bubbled with carbogen gas (95% O2 and 5% CO2). After cutting, slices were incubated for 30 min at 31–33℃ and then at room temperature in ACSF containing (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 1 MgCl2, 2 CaCl2, 25 NaHCO3, 25 D-glucose, 3 myo-inositol, 2 sodium pyruvate, and 0.4 ascorbic acid (300 mOsm/l; 95% O2 and 5% CO2). To maintain adequate oxygenation and a physiological pH level, slices were kept in carbogenated ACSF (95% O2 and 5% CO2) during the transportation.
Rat brains were quickly removed and placed in an ice-cold sucrose containing ACSF. The experimental procedures used here have been described in detail previously 64. 300 µm thick coronal slices of the prelimbic medial prefrontal cortex (mPFC) and temporal association cortex were cut and incubated using the same procedures and solutions as described above for human slices.
Organotypic slice cultures of human neocortex
Preparation and cultivation of slice cultures of human neocortex followed previously published protocols 65. In brief, the neocortex was carefully micro-dissected and resected with only minimal use of bipolar forceps to ensure tissue integrity, directly transferred into ice-cold artificial cerebrospinal fluid (ACSF) (in mM: 110 choline chloride, 26 NaHCO3, 10 D-glucose, 11.6 Na-ascorbate, 7 MgCl2, 3.1 Na-pyruvate, 2.5 KCl, 1.25 NaH2PO4, and 0.5 CaCl2) equilibrated with carbogen (95% O2, 5% CO2) and immediately transported to the laboratory. Tissue was kept always submerged in cool and carbogenated ACSF. After removal of the pia, tissue blocks were trimmed perpendicular to the cortical surface and 250 µm thick slices were prepared using a live tissue vibratome. After the cortical tissue was sliced as described above, slices were cut into several evenly sized pieces. Subsequently, slices were transferred onto culture membranes (uncoated 30 mm Millicell-CM tissue culture inserts with 0.4 µm pores, Millipore) and kept in six-well culture dishes (BD Biosciences). For the first hour following the slicing procedure, slices were cultured on 1.5 ml intermediate step HEPES media (48% DMEM/F-12 (Life Technologies), 48% Neurobasal (Life Technologies), 1x N-2 (Capricorn Scientific), 1x B-27 (Capricorn Scientific), 1x Glutamax (Life Technologies), 1x NEAA (Life Technologies) + 20 mM HEPES before changing to 1.5 ml hCSF per well without any supplements. No antibiotics or antimycotics were used during cultivation. The plates were stored in an incubator (MCO-170AICUVH-PE, PHC Corporation) at 37°C with 5% CO2 and 100% humidity. For MEA recordings, slice cultures were transferred into the recording chamber of a MEA Setup (described below).
Whole-cell recordings
Whole cell recordings were performed in acute slices 30 hours at most after slice preparation for human brain tissues and within 8 hours for rat brains. During whole-cell patch-clamp recordings, human or rat slices were continuously perfused (perfusion speed ∼ 5 ml/min) with ACSF bubbled with carbogen gas and maintained at 30–33℃. Patch pipettes were pulled from thick wall borosilicate glass capillaries and filled with an internal solution containing (in mM): 135 K-gluconate, 4 KCl, 10 HEPES, 10 phosphocreatine, 4 Mg-ATP, and 0.3 GTP (pH 7.4 with KOH, 290–300 mOsm). Neurons were visualized using either Dodt gradient contrast or infrared differential interference contrast microscopy. Human L2/3 neurons were identified and patched according to their somatic location (300–1200 µm from pia) 65. In rat acute prelimbic cortical slices, layer 2 is clearly distinguishable as a thin dark band that is densely packed with neuron somata. Layer 3 is about 2–3 times wider than layer 2 and has about the same width as layer 1. According to previous publications, layer 2/3 was located at a depth of 200 to 550 µm from the pia 66. Putative PCs and interneurons were differentiated on the basis of their intrinsic action potential (AP) firing pattern during recording and after post hoc histological staining also by their morphological appearance.
Whole-cell patch clamp recordings of human or rat L2/3 neurons were made using an EPC10 amplifier (HEKA). During recording, slices were perfused in ACSF at 31–33℃ containing (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 1 MgCl2, 2 CaCl2, 25 NaHCO3, 25 D-glucose, 3 myo-inositol, 2 sodium pyruvate, and 0.4 ascorbic acid. In a subset of experiments designed to induce ‘Up’ states, slices were perfused in a modified ACSF at 31–33℃ containing (in mM): 125 NaCl, 3.5 KCl, 1.25 NaH2PO4, 1 MgCl2, 1 CaCl2, 25 NaHCO3, 25 D-glucose, 3 myo-inositol, 2 sodium pyruvate, and 0.4 ascorbic acid. Signals were sampled at 10 kHz, filtered at 2.9 kHz using Patchmaster software (HEKA), and later analyzed offline using Igor Pro software (Wavemetrics). Recordings were performed using patch pipettes of resistance between 5 and 10 MΩ. Biocytin was added to the internal solution at a concentration of 3–5 mg/ml to stain patched neurons. A recording time > 15 min was necessary for an adequate diffusion of biocytin into dendrites and axons of patched cells 67.
Multi-electrode array (MEA) recordings
To perform the MEA recordings of the human cortical cultures, the brain slice was excised from the insert with the slice still attached to the culturing membrane. Subsequently, the slice was moved to the MEA chamber and placed onto the electrodes of the MEA Chip with the slice surface facing down. For fixation and improved contact with the electrodes, the slice was fixed in place by a weighted, close-meshed harp (ALA-HSG MEA-5BD, Multi Channel Systems MCS GmbH). Slices equilibrated at least 30 min on the chip with constant carbogenated ACSF (same as used for acute slices) perfusion at 30–33°C before MEA recordings were started. MEA recordings were performed using a 256–MEA (16 × 16 lattice) with electrode diameter of 30 µm and electrode spacing of 200 µm, thus covering a recording area of ∼3.2 × 3.2 mm2 (USB-MEA 256-System, Multi Channel Systems MCS GmbH). Recordings with the 256-MEA were performed at a sampling rate of 10–25 kHz using the Multi-Channel Experimenter (Multi Channel Systems MCS GmbH).
Drug application
NE (10 µM or 30 µM) and ACh (10 µM, 15 µM or 30 µM) were bath applied for 150–300 s through the perfusion system during whole-cell patch clamp or MEA recordings. In a subset of human neurons, propranolol (20 µM), tropicamide (TRO, 1 µM), tetrodotoxin (TTX, 0.5 µM), cyanquixaline (CNQX, 10µM), gabazine (1 µM), mecamylamine (1 µM) or prazosin (2 µM) were bath applied for 200–600 s to study the underlying pharmacological mechanisms. Drugs were purchased from Sigma-Aldrich or Tocris.
Histological staining
After recordings, brain slices containing biocytin-filled neurons were fixed for at least 24 h at 4 ℃ in 100 mM phosphate buffer solution (PBS, pH 7.4) containing 4% paraformaldehyde (PFA). After rinsing several times in 100 mM PBS, slices were treated with 1% H2O2 in PBS for about 20 min to reduce any endogenous peroxidase activity. Slices were rinsed repeatedly with PBS and then incubated in 1% avidin-biotinylated horseradish peroxidase (Vector ABC staining kit, Vector Lab. Inc.) containing 0.1% Triton X-100 for 1 h at room temperature. The reaction was catalyzed using 0.5 mg/ml 3,3-diaminobenzidine (DAB; Sigma-Aldrich) as a chromogen. Subsequently, slices were rinsed with 100 mM PBS, followed by slow dehydration with ethanol in increasing concentrations, and finally in xylene for 2–4 h. After that, slices were embedded using Eukitt medium (Otto Kindler GmbH).
In a subset of experiments, we tried to identify the expression of the molecular marker - a calcium-binding protein parvalbumin (PV) in human layer 2/3 interneurons. To this end, during electrophysiological recordings, Alexa Fluor 594 dye (1:500, Invitrogen) was added to the internal solution for post hoc identification of patched neurons. After recording, slices (300 µm) were fixed with 4% PFA in 100 mM PBS for at least 24 h at 4 ℃ and then permeabilized in 1% milk powder solution containing 0.5% Triton X-100 and 100 mM PBS. Primary and secondary antibodies were diluted in the permeabilization solution (0.5% Triton X-100 and 100 mM PBS) shortly before the antibody incubation. For single-cell PV staining, slices were incubated overnight with Rabbit-anti-PV primary antibody (1:120, ab11427, Abcam) at 4 ℃ and then rinsed thoroughly with 100 mM PBS. Subsequently, slices were treated with Donkey-anti-Rabbit Alexa Fluor secondary antibodies (1:400, A21207, Invitrogen) for 2–3 h at room temperature in the dark. After rinsing with 100 mM PBS, the slices were embedded in Fluoromount. Fluorescence images were taken using the Olympus CellSens platform. The position of the patched neurons was identified by the biocytin conjugated Alexa dye so that the expression of PV could be examined in biocytin-stained neurons. After acquiring fluorescent images, slices were incubated in 100 mM PBS overnight and then used for subsequent histological processing as described above.
Morphological 3D reconstructions
Using NEUROLUCIDA® software (MBF Bioscience, Williston, VT, USA), morphological reconstructions of biocytin filled human layer 2/3 interneurons were made at a magnification of 1000-fold (100-fold oil-immersion objective and 10-fold eyepiece) on an upright microscope. Neurons were selected for reconstruction based on the quality of biocytin labelling when background staining was minimal. Neurons with major truncations due to slicing were excluded. Embedding using Eukitt medium reduced fading of cytoarchitectonic features and enhanced contrast between layers 67. This allowed the reconstruction of different layer borders along with the neuronal reconstructions. Furthermore, the position of soma and layers were confirmed by superimposing the Dodt gradient contrast or differential interference contrast images taken during the recording. The tissue shrinkage was corrected using correction factors of 1.1 in the x–y direction and 2.1 in the z direction 67. Analysis of 3D-reconstructed neurons was done with NEUROEXPLORER® software (MBF Bioscience, Williston, VT, USA).
Data analysis
Single-cell recording data analysis
Custom written macros for Igor Pro 6 (WaveMetrics) were used to analyze the recorded electrophysiological signals. The resting membrane potential (Vm) of the neuron was measured directly after breakthrough to establish the whole-cell configuration with no current injection. The input resistance was calculated as the slope of the linear fit to the current–voltage relationship. For the analysis of single spike characteristics such as threshold, amplitude and half-width, a step size increment of 10 pA for current injection was applied to ensure that the AP was elicited very close to its rheobase current. The spike threshold was defined as the point of maximal acceleration of the membrane potential using the second derivative (d2V/dt2), which is, the time point with the fastest voltage change. The spike amplitude was calculated as the difference in voltage from AP threshold to the peak during depolarization. The spike half-width was determined as the time difference between rising phase and decaying phase of the spike at half-maximum amplitude.
The spontaneous activity was analyzed using the program SpAcAn (https://www.wavemetrics.com/project/SpAcAn). EPSPs and LRDs were distinguished by dramatic differences in event amplitude and decay time. A threshold of 0.2 mV was set manually for detecting EPSP events while a threshold of 3 mV was set for detecting LRDs. Recordings were not filtered to reduce noise before data analysis. When marking EPSPs, small EPSPs distributing in decay phrase but not rising phrase of LRDs were included into analysis. To study oscillatory network activities, we computed time-frequency representations of the signals by performing wavelet analysis using the Time-Frequency Toolkit (https://www.wavemetrics.com/project/TFPlot). Morlet wavelets were used for decomposition of recording signals as they provide an ideal compromise between time and frequency resolution 68.
MEA data analysis
MEA recordings were analyzed using custom-written programs in Python, detecting and quantifying the mean firing rate, number of active channels, bursting channels, and network bursts. First, the raw signal was filtered using a band-pass filter (Butterworth 2nd order). The spike identification was performed according to a threshold-based method using median absolute deviation (MAD) / 0.6745 x -5. Signal deviations were detected and aligned to the next minimum of the signal with a 1 ms dead time. The firing rate for each recording was defined as the number of recorded spikes divided by the duration of the recording (in s). Bursting channels were calculated using a modified adaptive network-wide cumulative moving average (CMA) approach in which all the electrodes of one MEA in multiple measurement time points were analyzed together 69.
In short, all inter-spike intervals (ISIs) were calculated and grouped into 5 ms bins for the whole recording. Next, a CMA over the histogram of the bins was calculated and a burst threshold was determined, therefore adapting the detection of bursts to the basic activity of the entire slice 69. The detection sensitivity was further increased by applying a minimum and maximum threshold of 60 ms and 140 ms, respectively. Whenever the threshold was undercut by at least three consecutive spikes, it was defined as a single-channel burst.
For a quantitative analysis of spiking synchronization, we used a graph theoretical approach to identify the degree of centrality of active channels 70–72. Specifically, we designated MEA contacts as nodes and the shared spiking time as edges to construct a graph representation of the recorded network activity. To construct edges, we grouped spike trains for each channel into 200 ms long bins and defined two channels as connected by an edge only if they both recorded spikes are in the same bin. By using this approach, we were able to examine the functional connectivity and activity patterns in the neuronal network.
We used mean degree centrality as a measure of the connectivity of nodes in the network (Python library NetworkX). Mean degree centrality (MDC) quantifies the average number of edges connecting a node to other nodes in the network. The degree of each node was defined by the number of edges connected to that node, i.e. the number of other MEA contacts that share a spiking time with that contact within a 200-ms bin. Finally, MDC was calculated by summing the degree of all nodes and dividing it by the total number of nodes in the network, which is
$$MDC=\frac{2\times numberoftotaledges}{numberoftotalnodes}$$
where the factor 2 is introduced because each edge is counted twice (once for each node it connects). By calculating the MDC for the whole MEA recording, we were able to assess the overall connectivity of the network and identify how nodes connectivity (used as a surrogate for synchronicity) changed upon application of NE and ACh compared to baseline.
Detection of LFPs was performed after low-pass filtering the signal of each channel (Butterworth 2nd order with a Nyquist frequency of 0.5 x sampling rate and a cut-off of 100 Hz); as a threshold, standard deviation of the low pass filtered signal multiplicated by three was used. Any deviation above or below this threshold with a minimum duration of 30 ms was defined as a LFP. The respective maximum and minimum deviation was defined as the amplitude of respective LFP.
All Python scripts are available on GitHub page (https://github.com/jonasort/MEA_analysis/tree/main/modified_common_script).
Statistical analysis
Data was either presented as box plots (n ≥ 10) or as bar histograms (n < 10). For box plots, the interquartile range (IQR) is shown as box, the range of values within 1.5∗IQR is shown as whiskers and the median is represented by a horizontal line in the box; for bar histograms, the mean ± SD is given. Wilcoxon Mann-Whitney U test was performed to access the difference between individual clusters. Statistical significance was set at P < 0.05, and n indicates the number of neurons/slices analyzed.