Degeneration and repair 1061
Dementia with Lewy bodies: cerebrospinal fluid suppresses neuronal network activity Stephan Theissa,b, Walter Maetzlerc,d,f, Christian Deuschlec,d, Holger Lerchee, Henner Koche and Marcel Dihnée One of the core clinical criteria of Dementia with Lewy bodies (DLB) are fluctuations of cognition. Underlying processes may be rather functional than neurodegenerative, reflected by, for example, factors present in cerebrospinal fluid (CSF). The aim of this study was to identify in-vitro neuronal network activity (ivNNA) changes of CSF from DLB patients compared with patients with Parkinson’s disease (PD) and controls. Primary neuronal mouse cultures were grown on microelectrode arrays to record ivNNA when exposed to respective CSF samples. If exposed to CSF of DLB patients, ivNNA showed a reduced spike rate and burst rate compared with CSF of PD patients and controls. Our data are suggestive of the presence of functional factors in the CSF of DLB patients that differentiate network activity from PD patients and controls. Future studies should evaluate whether this pilot observation might be related to fluctuations of cognition in DLB. NeuroReport
Introduction Dementia with Lewy bodies (DLB) is a neurodegenerative disease with pathological aggregations of α-synuclein. The clinical course is characterized by early cognitive deficits, with Parkinsonism appearing never or no more than 12 months before the onset of cognitive symptoms [1]. This course is distinguishing DLB from Parkinson’s disease (PD), in which motor symptoms usually precede cognitive symptoms for more than 5–10 years. One of the core criteria of DLB is fluctuating cognition. The pathology underlying this feature is not well understood. The rapid change over days or even hours favors soluble factors instead of the neurodegenerative process itself to be crucially involved. This hypothesis is supported by own work, where we showed that ammonium chloride influences in-vitro neuronal network activity (ivNNA) [2]. Ammonium chloride is known to play a pivotal role in hepatic encephalopathy, a disease that is also characterized by fluctuating states of consciousness [2]. The abovementioned neuronal activity was measured at the network level with excitatory and inhibitory neurons contributing toward network bursting [3]. We generated functionally interconnected neuronal populations and recorded their ivNNA [4], and could already show that human cerebrospinal fluid (CSF) from different diseases can modulate ivNNA [5,6]. As there is direct contact 0959-4965 Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
28:1061–1065 Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved. NeuroReport 2017, 28:1061–1065 Keywords: cerebrospinal fluid, fluctuation of cognition, Parkinson’s disease a Institute of Clinical Neuroscience and Medical Psychology, Medical Faculty, Heinrich Heine University, bResult Medical GmbH, Düsseldorf, cDepartment of Neurodegeneration, Center for Neurology and Hertie Institute for Clinical Brain Research, University of Tübingen, dDZNE, German Center for Neurodegenerative Diseases, eDepartment of Epileptology, Center for Neurology and Hertie Institute for Clinical Brain Research, University of Tübingen, Tübingen and fDepartment of Neurology, Kiel University, Keil, Germany
Correspondence to Henner Koch, MD, Department of Epileptology, Hertie Institute for Clinical Brain Research, University of Tübingen, Otfried-Müller Strasse 27, 72076 Tübingen, Germany Tel: + 49 707 129 81984; fax: + 49 707 129 4488; e-mail: [email protected] Received 21 June 2017 accepted 7 July 2017
between brain parenchyma and CSF, CSF may have the potential to inform us about potentially existing soluble factors with an influence of cognitive fluctuations. Here, we investigated ivNNA in neuronal populations exposed to CSF of DLB patients, and compared these values with CSF obtained from patients with PD and older adults without neurodegenerative diseases (controls).
Materials and methods Microelectrode arrays
To generate ivNNA, hippocampal neurons prepared from E17 mice were plated on microelectrode arrays (MEAs) and cultured for up to 6 weeks according to Hedrich et al. [7]. About 3 weeks after plating, synchronous network burst firing was detected by MEAs. MEAs had a square grid of 60 planar Ti/TiN electrodes of 30 µm diameter and 200 µm spacing. The input impedance of electrodes was 30–50 kΩ according to the specifications of the manufacturer (Multi Channel Systems, Reutlingen, Germany). Signals from all 60 electrodes were sampled simultaneously at 25 kHz, visualized, and stored using the standard software MC_Rack provided by Multi Channel Systems. Spike detection was performed off-line using the SPANNER software suite (Result Medical, Düsseldorf, Germany) [7]. The number of spikes per minute was calculated as a global measure of action potential activity in the entire network. DOI: 10.1097/WNR.0000000000000890
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
1062 NeuroReport 2017, Vol 28 No 16
Coordinated firing across the entire network was quantified by the number of network bursts per minute, determined from peaks in the network firing rate [8]: spikes were aggregated from all electrodes in nonoverlapping 5 ms bins and smoothed by a Gaussian kernel with 100 ms SD (custom-built MATLAB software, see [7]). For each network burst, its peak firing rate (PFR, spikes/second) was determined as the maximum network firing rate within this burst so that shorter network bursts with more spikes had higher PFR. Cerebrospinal fluid collection
We collected CSF samples by diagnostic routine lumbar puncture. The use of CSF samples for research purposes was approved by the ethics board of the University of Tübingen (663/2012BO2). Controls were subjected to lumbar puncture to exclude intracranial hemorrhage or infection. PD patients without evidence of dementia and motor/nonmotor fluctuations fulfilled the UKBBS diagnostic criteria [9]. All DLB patients, diagnosed according to MacKeith et al. [1], had fluctuations in cognition and alertness. The median age of controls, PD, and DLB patients was 74, 77, and 75 years. The medians of Mini-Mental State Examination values (0–30) were 30 in controls, 29 in PD, and 16 in DLB patients (P < 0.001; Kruskal–Wallis test). Details of the collection process and storage have been published previously [10]. In brief, samples were processed on ice within 30 min after collection and centrifuged for Table 1
10 min at 4000g at + 4°C. Then, the supernatant was pipetted in a new tube and aliquots of 400 µl were stored at − 80°C within a 60-min time frame. Comparison of routine CSF parameters between the groups indicated no significant differences for all except one parameter: Amyloid-β1–42 (Table 1). As expected, the latter was lower in DLB than in PD and controls, respectively. Cerebrospinal fluid analysis and statistics
Each experiment lasted 25 min, comprising five 5-min MEA recordings. During the first 5 min, a screening recording was performed in culture medium to verify the quality of network bursting by inspecting the live signal. Then, the whole medium was exchanged against the artificial CSF (aCSF) sample (100 µl, 150 mM sodium, 1 mM calcium, 3 mM potassium, 1 mM magnesium, 10 mM HEPES, 10 mM glucose). Before application of aCSF, cultures were washed twice with 0.01 M sodium phosphate-buffered saline. For baseline measurements, aCSF was applied and recorded 5 + 5 min to ensure stable activity. Only the second 5-min interval was used for analysis. To provide a homogenous baseline activity in and between all groups, only MEA chips with a spike rate between 600/min and 9000/min and a network burst rate between 25/5 min and 165/5 min when exposed to aCSF were used. In earlier experiments, we could show that baseline activity was constant in an individual MEA chip also after several applications of aCSF. Baseline
Routine clinical and laboratory data of cohorts PD
Individuals (N) (female/male) Age [median (range)] (years) Age [mean (SD)] (years) Aao Parkinsonism [Median (range)] (years) Aao Parkinsonism [mean (SD)] (years) Duration of Parkinsonism [median (range)] (years) Duration of Parkinsonism [mean (SD)] (years) H&Y stage (1–5) [median (range)] H&Y stage (1–5) [mean (SD)] Aao dementia [median (range)] (years) Aao dementia [mean (SD)] (years) Duration of dementia [median (range)] (years) Duration of dementia [mean (SD)] (years) MMSE (0–30) [median (range)] MMSE (0–30) [mean (SD)] CSF cell number (n/3) [median (range)] CSF cell number (n/3) [mean (SD)] CSF albumin [median (range)] (mg/dl) CSF albumin [mean (SD)] (mg/dl) IgG index [median (range)] IgG index [mean (SD)] Aβ1–42 [median (range)] (pg/ml) Aβ1–42 [median (SD)] (pg/ml) h-Tau [median (range)] (pg/ml) h-Tau [median (SD)] (pg/ml) p-Tau [median (range)] (pg/ml) p-Tau [median (SD)] (pg/ml)
7 77 76 71 69 5 7 2 2
29 28 2 2 259 238 0.48 0.48 714 678 264 337 44 43
(2/5) (70–80) (4) (50–76) (3) (2–24) (8) (2–3) (1) – – – – (26–29) (1) (0–6) (2) (145–380) (62) (0.40–0.55) (0.06) (319–829) (193) (139–1074) (350) (27–63) (14)
DLB 7 75 76 72 72 3 3
72 72 3 3 16 17 2 1 246 262 0.42 0.42 348 358 357 377 68 60
(3/4) (70–81) (5) (65–79) (5) (0–7) (2) – – (65–81) (5) (0–7) (2) (11–23)# (5)# (0–6) (2) (159–522) (120) (0.39–0.47) (0.03) (205–425)# (88)# (153–556) (148) (22–86) (19)
Controls
P value
7 (3/4) 74 (73–79) 75 (3) – – – – – – – – – – 30 (30–30)§ § 30 (0) 3 (0–6) 3 (2) 237 (224–276) 242 (19) 0.46 (0.38–0.50) 0.45 (0.01) 752 (597–1147)§ 744 (111)§ 241 (121–329) 217 (72) 45 (28–57) 43 (10)
0.33a 0.87 0.89 0.53 0.39 0.36 0.18 – – – – – – < 0.001 < 0.001 0.36 0.43 0.83 0.88 0.24 0.13 0.001 0.005 0.40 0.26 0.08 0.09
Routine clinical and laboratory data of the included cohorts, presented with mean and SD/median and range except number of participants (frequency). ANOVA/ Kruskal–Wallis test (in case of number of participants) and, if significant, post-hoc Student’s t-test/Wilcoxon signed rank test were used to compare groups. Aao, age at onset; Aβ, amyloid-β; ANOVA, analysis of variance; CSF, cerebrospinal fluid; H&Y, Hoehn & Yahr; MMSE, Mini-Mental State Examination. a 2 χ -test. # Post-hoc P < 0.017 against Parkinson’s disease. § Post-hoc P < 0.017 against dementia with Lewy bodies.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Lewy-body dementia CSF reduces bursting Theiss et al. 1063
activity depends, however, on a constant pH, which was adjusted in each CSF sample to 7.4 before every MEA measurement. Afterwards, human CSF samples were applied and again recorded for 5 + 5 min. For each measurement and each parameter, the ratio ‘human CSF/ aCSFʼ was calculated. In each group [controls (cCSF), PD (pdCSF) or DLB (lbCSF)], seven CSF samples of seven different patients (overall 21 samples) were analyzed on different and independent MEAs. Statistical analyses were carried out using Graph Pad Prism 4 software (GraphPad Software Inc., La Jolla, California, USA). Demographics and clinical data are presented in Table 1 with mean and SD/median and range, and analysis of variance and post-hoc Student’s t-test/Kruskal–Wallis and post-hoc Wilcoxon signed rank tests were used to compare groups. Experimental results are presented using mean and SEM. We used the paired t-test for analyses of aCSF against the human CSF samples. For comparison of respective ratios, we used the Kruskal–Wallis test and (post-hoc) Dunn’s multiple comparison test.
Results Twenty-one days after plating, anatomical and functional neuronal networks had formed
E17-murine hippocampal cells and networks matured at individual rates after plating on microelectrode arrays, and they showed gradually increasing states of functionality: 1–2 weeks after plating formerly immature neurons, increasingly dense networks of neurons and astrocytes were generated (Fig. 1a). At this stage, spontaneous neuronal firing was observed on single electrodes. Twenty-one to 48 days after plating, neuronal activity comprised synchronous bursts: densely packed spike activity separated by mostly silent periods (Fig. 1b) on multiple spatially separated electrodes (Fig. 1c). This time course is in line with earlier observations [4]. Experiments were conducted exclusively with cultures that showed synchronous network bursting. Human cerebrospinal fluid samples enhance network bursting
The experimental design is shown in Fig. 1d. Baseline recordings were performed when exposed to aCSF. Then, aCSF was exchanged against pure human CSF samples. After replacing aCSF with cCSF (n = 7), global network activity (spikes/min) as well as the number of network bursts and the peak firing rate were significantly increased (Fig. 1e–g). This is in line with results from earlier studies [6]. Increase in global network activity, measured in the spike rate and burst rate, under the influence of cCSF is further illustrated by spike raster plots (Fig. 1h and i). When exposed to cCSF, activity was still mostly organized in network bursts as it was under the influence of aCSF. However, we also observed an increase in asynchronous spike activity between the network bursts, which was not quantified further.
Cerebrospinal fluid from patients with dementia with Lewy bodies leads to lower levels of in-vitro neuronal network activity compared with cerebrospinal fluid from patients with Parkinson’s disease
We then applied CSF from patients with PD (pdCSF) and CSF from patients with DLB (lbCSF) in separate sets of experimental groups, again after baseline activity assessment with aCSF (Fig. 1d). The means of baseline activities between groups were similar (Fig. 1j). pdCSF was associated with significantly increased spike and burst rates compared with aCSF (Fig. 1k). lbCSF did not induce significant changes in the network activity compared with aCSF (Fig. 1l). Respective ratios were then compared between the DLB, PD, and control groups. Neuronal activity, as measured with the spike rate, was significantly different between groups (Kruskal–Wallis, P = 0.0022). The spike rate was lower in MEAs exposed to lbCSF than in those exposed to cCSF (Dunn’s P = 0.0421) and pdCSF (P = 0.0071; Fig. 1m). Also, network bursts differed significantly between groups (Kruskal–Wallis, P = 0.0279). Burst rate was significantly different between lbCSF-exposed and pdCSF-exposed MEAs (Dunn’s P = 0.0285). The PFR ratios were not significantly different between groups.
Discussion Fluctuations in cognition represent one of the core diagnostic criteria that separate DLB from PD [11] and are a well-known and often severely disabling feature of DLB. Underlying mechanisms are unknown. As the neurodegenerative process itself is slowly progressive and does not explain hour-to-hour and day-to-day changes, it is tempting to speculate that additional mechanisms are responsible for this phenomenon. Those mechanisms could include soluble, disease-related factors within the CSF that are subject to fluctuations within the time periods mentioned. Soluble factors within the CSF can in principle reach the parenchymal compartment of the brain and, thus, influence behavior. This was shown in several animal studies [12–14]. To simulate and measure functional brain–CSF interaction in vitro, murine neuronal populations were cultured on MEAs. Thus, extracellular electrophysiological network activity under the influence of human CSF could be recorded. Here, we found that CSF samples of DLB patients showed reduced overall numbers of spikes and network bursts in in-vitro functional neuronal networks compared with CSF samples obtained from PD patients and controls. This finding suggests that soluble factors in lbCSF exist that might downregulate network bursting, or, vice versa, that factors are lacking in lbCSF that maintain such activity. Interestingly, ivNNA under the influence of lbCSF was still competent to reach high spike rates during the shorter time period of a network burst as the PFRs were similar in all groups. This means that the regulation of ivNNA under the influence of lbCSF is complex, leading to reduced overall network bursting and spike rates, but
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
1064 NeuroReport 2017, Vol 28 No 16
Fig. 1
(a) Phase-contrast image showing neuronal cultures on a microelectrode array (MEA) electrode field. (b) An exemplary recording trace shows burst duration (BD), spike amplitude (SA), interburst interval (IBI), and typical values of BD and SA in the enlarged trace fragment. (c) A typical 1-s snapshot of a recording during a network burst. Every rectangle illustrates the activity of a single electrode during the duration of 1 s. (d) Artificial cerebrospinal fluid (aCSF) was first pipetted on the MEA and after 10 min, exchanged against one of the human CSF (hCSF) samples. Asterisks mark the 5-min time periods of data collection entering statistics. (e) Absolute values of spikes/1 min, network bursts/5 min, and the peak firing rate (PFR) are shown for aCSF and cCSF as indicated. Spike raster plots are given for aCSF (h) and cCSF (i), showing an enhanced global activity under the influence of cCSF. Each ‘ + ʼ denotes a spike detected at a given time and electrode. (j) Absolute values are given for indicated experimental groups (k) pdCSF or (l) lbCSF with indicated parameters. (m) Ratios (human CSF/aCSF) are shown for the indicated experimental groups and parameters. Filled bars show mean values and error bars represent SEM. lbCSF, cerebrospinal fluid from patients with dementia with Lewy bodies; pdCSF, cerebrospinal fluid from patients with Parkinson’s disease.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Lewy-body dementia CSF reduces bursting Theiss et al. 1065
still allowing for high neuronal activity during a very short time interval, namely, a single burst. We focused on a descriptive level in this pilot report and did not conduct additional experiments to define the factors potentially responsible for the observed effect. This should be performed in future studies, which may also include CSF collected at clearly defined phases during reported fluctuations. Interesting candidates are small amino acids acting at the N-methyl-D-aspartate receptor, such as glycine [6].
Conclusion
References 1
2
3 4
5
6
Our study suggests that soluble factors exist that are functionally measurable in the CSF of DLB patients. These factors may be associated with fluctuations of cognition, which should be tested in prospective studies.
7
8
Acknowledgements The authors thank Gabriele Kuebart for technical support. Samples were obtained from the Neuro-Biobank of the University of Tuebingen, Germany (http://www.hih-tuebingen. de/ueber-uns/core-facilities/biobank/). This biobank is supported by the local University, the Hertie Institute and the DZNE. ST received support by the German Ministry of Education and Research (BMBF: FKZ 031B0010B) and the European Union (EuroTransBio9 project In-HEALTH).
9
10
11 12
13 14
Conflicts of interest
There are no conflicts of interest.
McKeith IG, Dickson DW, Lowe J, Emre M, O’Brien JT, Feldman H, et al. Diagnosis and management of dementia with Lewy bodies: third report of the DLB Consortium. Neurology 2005; 65:1863–1872. Schwarz CS, Ferrea S, Quasthoff K, Walter J, Gorg B, Haussinger D, et al. Ammonium chloride influences in vitro-neuronal network activity. Exp Neurol 2012; 235:368–373. Hartung HP, Dihne M. Volume transmission-mediated encephalopathies: a possible new concept? Arch Neurol 2012; 69:315–321. Illes S, Fleischer W, Siebler M, Hartung HP, Dihne M. Development and pharmacological modulation of embryonic stem cell-derived neuronal network activity. Exp Neurol 2007; 207:171–176. Jantzen SU, Ferrea S, Wach C, Quasthoff K, Illes S, Scherfeld D, et al. In vitro neuronal network activity in NMDA receptor encephalitis. BMC Neurosci 2013; 14:17. Otto F, Illes S, Opatz J, Laryea M, Theiss S, Hartung HP, et al. Cerebrospinal fluid of brain trauma patients inhibits in vitro neuronal network function via NMDA receptors. Ann Neurol 2009; 66:546–555. Hedrich UB, Liautard C, Kirschenbaum D, Pofahl M, Lavigne J, Liu Y, et al. Impaired action potential initiation in GABAergic interneurons causes hyperexcitable networks in an epileptic mouse model carrying a human Na(V) 1.1 mutation. J Neurosci 2014; 34:14874–14889. Schock SC, Jolin-Dahel KS, Schock PC, Theiss S, Arbuthnott GW, GarciaMunoz M, et al. Development of dissociated cryopreserved rat cortical neurons in vitro. J Neurosci Methods 2012; 205:324–333. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992; 55:181–184. Maetzler W, Schmid SP, Wurster I, Liepelt I, Gaenslen A, Gasser T, et al. Reduced but not oxidized cerebrospinal fluid glutathione levels are lowered in Lewy body diseases. Mov Disord 2011; 26:176–181. Walker Z, Possin KL, Boeve BF, Aarsland D. Lewy body dementias. Lancet 2015; 386:1683–1697. Borges LF, Elliott PJ, Gill R, Iversen SD, Iversen LL. Selective extraction of small and large molecules from the cerebrospinal fluid by Purkinje neurons. Science 1985; 228:346–348. Pappenheimer JR. Bayliss-Starling Memorial Lecture (1982). Induction of sleep by muramyl peptides. J Physiol 1983; 336:1–11. Silver R, LeSauter J, Tresco PA, Lehman MN. A diffusible coupling signal from the transplanted suprachiasmatic nucleus controlling circadian locomotor rhythms. Nature 1996; 382:810–813.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Dementia with Lewy bodies: cerebrospinal fluid suppresses neuronal network activity Stephan Theissa,b, Walter Maetzlerc,d,f, Christian Deuschlec,d, Holger Lerchee, Henner Koche and Marcel Dihnée One of the core clinical criteria of Dementia with Lewy bodies (DLB) are fluctuations of cognition. Underlying processes may be rather functional than neurodegenerative, reflected by, for example, factors present in cerebrospinal fluid (CSF). The aim of this study was to identify in-vitro neuronal network activity (ivNNA) changes of CSF from DLB patients compared with patients with Parkinson’s disease (PD) and controls. Primary neuronal mouse cultures were grown on microelectrode arrays to record ivNNA when exposed to respective CSF samples. If exposed to CSF of DLB patients, ivNNA showed a reduced spike rate and burst rate compared with CSF of PD patients and controls. Our data are suggestive of the presence of functional factors in the CSF of DLB patients that differentiate network activity from PD patients and controls. Future studies should evaluate whether this pilot observation might be related to fluctuations of cognition in DLB. NeuroReport
Introduction Dementia with Lewy bodies (DLB) is a neurodegenerative disease with pathological aggregations of α-synuclein. The clinical course is characterized by early cognitive deficits, with Parkinsonism appearing never or no more than 12 months before the onset of cognitive symptoms [1]. This course is distinguishing DLB from Parkinson’s disease (PD), in which motor symptoms usually precede cognitive symptoms for more than 5–10 years. One of the core criteria of DLB is fluctuating cognition. The pathology underlying this feature is not well understood. The rapid change over days or even hours favors soluble factors instead of the neurodegenerative process itself to be crucially involved. This hypothesis is supported by own work, where we showed that ammonium chloride influences in-vitro neuronal network activity (ivNNA) [2]. Ammonium chloride is known to play a pivotal role in hepatic encephalopathy, a disease that is also characterized by fluctuating states of consciousness [2]. The abovementioned neuronal activity was measured at the network level with excitatory and inhibitory neurons contributing toward network bursting [3]. We generated functionally interconnected neuronal populations and recorded their ivNNA [4], and could already show that human cerebrospinal fluid (CSF) from different diseases can modulate ivNNA [5,6]. As there is direct contact 0959-4965 Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
28:1061–1065 Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved. NeuroReport 2017, 28:1061–1065 Keywords: cerebrospinal fluid, fluctuation of cognition, Parkinson’s disease a Institute of Clinical Neuroscience and Medical Psychology, Medical Faculty, Heinrich Heine University, bResult Medical GmbH, Düsseldorf, cDepartment of Neurodegeneration, Center for Neurology and Hertie Institute for Clinical Brain Research, University of Tübingen, dDZNE, German Center for Neurodegenerative Diseases, eDepartment of Epileptology, Center for Neurology and Hertie Institute for Clinical Brain Research, University of Tübingen, Tübingen and fDepartment of Neurology, Kiel University, Keil, Germany
Correspondence to Henner Koch, MD, Department of Epileptology, Hertie Institute for Clinical Brain Research, University of Tübingen, Otfried-Müller Strasse 27, 72076 Tübingen, Germany Tel: + 49 707 129 81984; fax: + 49 707 129 4488; e-mail: [email protected] Received 21 June 2017 accepted 7 July 2017
between brain parenchyma and CSF, CSF may have the potential to inform us about potentially existing soluble factors with an influence of cognitive fluctuations. Here, we investigated ivNNA in neuronal populations exposed to CSF of DLB patients, and compared these values with CSF obtained from patients with PD and older adults without neurodegenerative diseases (controls).
Materials and methods Microelectrode arrays
To generate ivNNA, hippocampal neurons prepared from E17 mice were plated on microelectrode arrays (MEAs) and cultured for up to 6 weeks according to Hedrich et al. [7]. About 3 weeks after plating, synchronous network burst firing was detected by MEAs. MEAs had a square grid of 60 planar Ti/TiN electrodes of 30 µm diameter and 200 µm spacing. The input impedance of electrodes was 30–50 kΩ according to the specifications of the manufacturer (Multi Channel Systems, Reutlingen, Germany). Signals from all 60 electrodes were sampled simultaneously at 25 kHz, visualized, and stored using the standard software MC_Rack provided by Multi Channel Systems. Spike detection was performed off-line using the SPANNER software suite (Result Medical, Düsseldorf, Germany) [7]. The number of spikes per minute was calculated as a global measure of action potential activity in the entire network. DOI: 10.1097/WNR.0000000000000890
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
1062 NeuroReport 2017, Vol 28 No 16
Coordinated firing across the entire network was quantified by the number of network bursts per minute, determined from peaks in the network firing rate [8]: spikes were aggregated from all electrodes in nonoverlapping 5 ms bins and smoothed by a Gaussian kernel with 100 ms SD (custom-built MATLAB software, see [7]). For each network burst, its peak firing rate (PFR, spikes/second) was determined as the maximum network firing rate within this burst so that shorter network bursts with more spikes had higher PFR. Cerebrospinal fluid collection
We collected CSF samples by diagnostic routine lumbar puncture. The use of CSF samples for research purposes was approved by the ethics board of the University of Tübingen (663/2012BO2). Controls were subjected to lumbar puncture to exclude intracranial hemorrhage or infection. PD patients without evidence of dementia and motor/nonmotor fluctuations fulfilled the UKBBS diagnostic criteria [9]. All DLB patients, diagnosed according to MacKeith et al. [1], had fluctuations in cognition and alertness. The median age of controls, PD, and DLB patients was 74, 77, and 75 years. The medians of Mini-Mental State Examination values (0–30) were 30 in controls, 29 in PD, and 16 in DLB patients (P < 0.001; Kruskal–Wallis test). Details of the collection process and storage have been published previously [10]. In brief, samples were processed on ice within 30 min after collection and centrifuged for Table 1
10 min at 4000g at + 4°C. Then, the supernatant was pipetted in a new tube and aliquots of 400 µl were stored at − 80°C within a 60-min time frame. Comparison of routine CSF parameters between the groups indicated no significant differences for all except one parameter: Amyloid-β1–42 (Table 1). As expected, the latter was lower in DLB than in PD and controls, respectively. Cerebrospinal fluid analysis and statistics
Each experiment lasted 25 min, comprising five 5-min MEA recordings. During the first 5 min, a screening recording was performed in culture medium to verify the quality of network bursting by inspecting the live signal. Then, the whole medium was exchanged against the artificial CSF (aCSF) sample (100 µl, 150 mM sodium, 1 mM calcium, 3 mM potassium, 1 mM magnesium, 10 mM HEPES, 10 mM glucose). Before application of aCSF, cultures were washed twice with 0.01 M sodium phosphate-buffered saline. For baseline measurements, aCSF was applied and recorded 5 + 5 min to ensure stable activity. Only the second 5-min interval was used for analysis. To provide a homogenous baseline activity in and between all groups, only MEA chips with a spike rate between 600/min and 9000/min and a network burst rate between 25/5 min and 165/5 min when exposed to aCSF were used. In earlier experiments, we could show that baseline activity was constant in an individual MEA chip also after several applications of aCSF. Baseline
Routine clinical and laboratory data of cohorts PD
Individuals (N) (female/male) Age [median (range)] (years) Age [mean (SD)] (years) Aao Parkinsonism [Median (range)] (years) Aao Parkinsonism [mean (SD)] (years) Duration of Parkinsonism [median (range)] (years) Duration of Parkinsonism [mean (SD)] (years) H&Y stage (1–5) [median (range)] H&Y stage (1–5) [mean (SD)] Aao dementia [median (range)] (years) Aao dementia [mean (SD)] (years) Duration of dementia [median (range)] (years) Duration of dementia [mean (SD)] (years) MMSE (0–30) [median (range)] MMSE (0–30) [mean (SD)] CSF cell number (n/3) [median (range)] CSF cell number (n/3) [mean (SD)] CSF albumin [median (range)] (mg/dl) CSF albumin [mean (SD)] (mg/dl) IgG index [median (range)] IgG index [mean (SD)] Aβ1–42 [median (range)] (pg/ml) Aβ1–42 [median (SD)] (pg/ml) h-Tau [median (range)] (pg/ml) h-Tau [median (SD)] (pg/ml) p-Tau [median (range)] (pg/ml) p-Tau [median (SD)] (pg/ml)
7 77 76 71 69 5 7 2 2
29 28 2 2 259 238 0.48 0.48 714 678 264 337 44 43
(2/5) (70–80) (4) (50–76) (3) (2–24) (8) (2–3) (1) – – – – (26–29) (1) (0–6) (2) (145–380) (62) (0.40–0.55) (0.06) (319–829) (193) (139–1074) (350) (27–63) (14)
DLB 7 75 76 72 72 3 3
72 72 3 3 16 17 2 1 246 262 0.42 0.42 348 358 357 377 68 60
(3/4) (70–81) (5) (65–79) (5) (0–7) (2) – – (65–81) (5) (0–7) (2) (11–23)# (5)# (0–6) (2) (159–522) (120) (0.39–0.47) (0.03) (205–425)# (88)# (153–556) (148) (22–86) (19)
Controls
P value
7 (3/4) 74 (73–79) 75 (3) – – – – – – – – – – 30 (30–30)§ § 30 (0) 3 (0–6) 3 (2) 237 (224–276) 242 (19) 0.46 (0.38–0.50) 0.45 (0.01) 752 (597–1147)§ 744 (111)§ 241 (121–329) 217 (72) 45 (28–57) 43 (10)
0.33a 0.87 0.89 0.53 0.39 0.36 0.18 – – – – – – < 0.001 < 0.001 0.36 0.43 0.83 0.88 0.24 0.13 0.001 0.005 0.40 0.26 0.08 0.09
Routine clinical and laboratory data of the included cohorts, presented with mean and SD/median and range except number of participants (frequency). ANOVA/ Kruskal–Wallis test (in case of number of participants) and, if significant, post-hoc Student’s t-test/Wilcoxon signed rank test were used to compare groups. Aao, age at onset; Aβ, amyloid-β; ANOVA, analysis of variance; CSF, cerebrospinal fluid; H&Y, Hoehn & Yahr; MMSE, Mini-Mental State Examination. a 2 χ -test. # Post-hoc P < 0.017 against Parkinson’s disease. § Post-hoc P < 0.017 against dementia with Lewy bodies.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Lewy-body dementia CSF reduces bursting Theiss et al. 1063
activity depends, however, on a constant pH, which was adjusted in each CSF sample to 7.4 before every MEA measurement. Afterwards, human CSF samples were applied and again recorded for 5 + 5 min. For each measurement and each parameter, the ratio ‘human CSF/ aCSFʼ was calculated. In each group [controls (cCSF), PD (pdCSF) or DLB (lbCSF)], seven CSF samples of seven different patients (overall 21 samples) were analyzed on different and independent MEAs. Statistical analyses were carried out using Graph Pad Prism 4 software (GraphPad Software Inc., La Jolla, California, USA). Demographics and clinical data are presented in Table 1 with mean and SD/median and range, and analysis of variance and post-hoc Student’s t-test/Kruskal–Wallis and post-hoc Wilcoxon signed rank tests were used to compare groups. Experimental results are presented using mean and SEM. We used the paired t-test for analyses of aCSF against the human CSF samples. For comparison of respective ratios, we used the Kruskal–Wallis test and (post-hoc) Dunn’s multiple comparison test.
Results Twenty-one days after plating, anatomical and functional neuronal networks had formed
E17-murine hippocampal cells and networks matured at individual rates after plating on microelectrode arrays, and they showed gradually increasing states of functionality: 1–2 weeks after plating formerly immature neurons, increasingly dense networks of neurons and astrocytes were generated (Fig. 1a). At this stage, spontaneous neuronal firing was observed on single electrodes. Twenty-one to 48 days after plating, neuronal activity comprised synchronous bursts: densely packed spike activity separated by mostly silent periods (Fig. 1b) on multiple spatially separated electrodes (Fig. 1c). This time course is in line with earlier observations [4]. Experiments were conducted exclusively with cultures that showed synchronous network bursting. Human cerebrospinal fluid samples enhance network bursting
The experimental design is shown in Fig. 1d. Baseline recordings were performed when exposed to aCSF. Then, aCSF was exchanged against pure human CSF samples. After replacing aCSF with cCSF (n = 7), global network activity (spikes/min) as well as the number of network bursts and the peak firing rate were significantly increased (Fig. 1e–g). This is in line with results from earlier studies [6]. Increase in global network activity, measured in the spike rate and burst rate, under the influence of cCSF is further illustrated by spike raster plots (Fig. 1h and i). When exposed to cCSF, activity was still mostly organized in network bursts as it was under the influence of aCSF. However, we also observed an increase in asynchronous spike activity between the network bursts, which was not quantified further.
Cerebrospinal fluid from patients with dementia with Lewy bodies leads to lower levels of in-vitro neuronal network activity compared with cerebrospinal fluid from patients with Parkinson’s disease
We then applied CSF from patients with PD (pdCSF) and CSF from patients with DLB (lbCSF) in separate sets of experimental groups, again after baseline activity assessment with aCSF (Fig. 1d). The means of baseline activities between groups were similar (Fig. 1j). pdCSF was associated with significantly increased spike and burst rates compared with aCSF (Fig. 1k). lbCSF did not induce significant changes in the network activity compared with aCSF (Fig. 1l). Respective ratios were then compared between the DLB, PD, and control groups. Neuronal activity, as measured with the spike rate, was significantly different between groups (Kruskal–Wallis, P = 0.0022). The spike rate was lower in MEAs exposed to lbCSF than in those exposed to cCSF (Dunn’s P = 0.0421) and pdCSF (P = 0.0071; Fig. 1m). Also, network bursts differed significantly between groups (Kruskal–Wallis, P = 0.0279). Burst rate was significantly different between lbCSF-exposed and pdCSF-exposed MEAs (Dunn’s P = 0.0285). The PFR ratios were not significantly different between groups.
Discussion Fluctuations in cognition represent one of the core diagnostic criteria that separate DLB from PD [11] and are a well-known and often severely disabling feature of DLB. Underlying mechanisms are unknown. As the neurodegenerative process itself is slowly progressive and does not explain hour-to-hour and day-to-day changes, it is tempting to speculate that additional mechanisms are responsible for this phenomenon. Those mechanisms could include soluble, disease-related factors within the CSF that are subject to fluctuations within the time periods mentioned. Soluble factors within the CSF can in principle reach the parenchymal compartment of the brain and, thus, influence behavior. This was shown in several animal studies [12–14]. To simulate and measure functional brain–CSF interaction in vitro, murine neuronal populations were cultured on MEAs. Thus, extracellular electrophysiological network activity under the influence of human CSF could be recorded. Here, we found that CSF samples of DLB patients showed reduced overall numbers of spikes and network bursts in in-vitro functional neuronal networks compared with CSF samples obtained from PD patients and controls. This finding suggests that soluble factors in lbCSF exist that might downregulate network bursting, or, vice versa, that factors are lacking in lbCSF that maintain such activity. Interestingly, ivNNA under the influence of lbCSF was still competent to reach high spike rates during the shorter time period of a network burst as the PFRs were similar in all groups. This means that the regulation of ivNNA under the influence of lbCSF is complex, leading to reduced overall network bursting and spike rates, but
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
1064 NeuroReport 2017, Vol 28 No 16
Fig. 1
(a) Phase-contrast image showing neuronal cultures on a microelectrode array (MEA) electrode field. (b) An exemplary recording trace shows burst duration (BD), spike amplitude (SA), interburst interval (IBI), and typical values of BD and SA in the enlarged trace fragment. (c) A typical 1-s snapshot of a recording during a network burst. Every rectangle illustrates the activity of a single electrode during the duration of 1 s. (d) Artificial cerebrospinal fluid (aCSF) was first pipetted on the MEA and after 10 min, exchanged against one of the human CSF (hCSF) samples. Asterisks mark the 5-min time periods of data collection entering statistics. (e) Absolute values of spikes/1 min, network bursts/5 min, and the peak firing rate (PFR) are shown for aCSF and cCSF as indicated. Spike raster plots are given for aCSF (h) and cCSF (i), showing an enhanced global activity under the influence of cCSF. Each ‘ + ʼ denotes a spike detected at a given time and electrode. (j) Absolute values are given for indicated experimental groups (k) pdCSF or (l) lbCSF with indicated parameters. (m) Ratios (human CSF/aCSF) are shown for the indicated experimental groups and parameters. Filled bars show mean values and error bars represent SEM. lbCSF, cerebrospinal fluid from patients with dementia with Lewy bodies; pdCSF, cerebrospinal fluid from patients with Parkinson’s disease.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
Lewy-body dementia CSF reduces bursting Theiss et al. 1065
still allowing for high neuronal activity during a very short time interval, namely, a single burst. We focused on a descriptive level in this pilot report and did not conduct additional experiments to define the factors potentially responsible for the observed effect. This should be performed in future studies, which may also include CSF collected at clearly defined phases during reported fluctuations. Interesting candidates are small amino acids acting at the N-methyl-D-aspartate receptor, such as glycine [6].
Conclusion
References 1
2
3 4
5
6
Our study suggests that soluble factors exist that are functionally measurable in the CSF of DLB patients. These factors may be associated with fluctuations of cognition, which should be tested in prospective studies.
7
8
Acknowledgements The authors thank Gabriele Kuebart for technical support. Samples were obtained from the Neuro-Biobank of the University of Tuebingen, Germany (http://www.hih-tuebingen. de/ueber-uns/core-facilities/biobank/). This biobank is supported by the local University, the Hertie Institute and the DZNE. ST received support by the German Ministry of Education and Research (BMBF: FKZ 031B0010B) and the European Union (EuroTransBio9 project In-HEALTH).
9
10
11 12
13 14
Conflicts of interest
There are no conflicts of interest.
McKeith IG, Dickson DW, Lowe J, Emre M, O’Brien JT, Feldman H, et al. Diagnosis and management of dementia with Lewy bodies: third report of the DLB Consortium. Neurology 2005; 65:1863–1872. Schwarz CS, Ferrea S, Quasthoff K, Walter J, Gorg B, Haussinger D, et al. Ammonium chloride influences in vitro-neuronal network activity. Exp Neurol 2012; 235:368–373. Hartung HP, Dihne M. Volume transmission-mediated encephalopathies: a possible new concept? Arch Neurol 2012; 69:315–321. Illes S, Fleischer W, Siebler M, Hartung HP, Dihne M. Development and pharmacological modulation of embryonic stem cell-derived neuronal network activity. Exp Neurol 2007; 207:171–176. Jantzen SU, Ferrea S, Wach C, Quasthoff K, Illes S, Scherfeld D, et al. In vitro neuronal network activity in NMDA receptor encephalitis. BMC Neurosci 2013; 14:17. Otto F, Illes S, Opatz J, Laryea M, Theiss S, Hartung HP, et al. Cerebrospinal fluid of brain trauma patients inhibits in vitro neuronal network function via NMDA receptors. Ann Neurol 2009; 66:546–555. Hedrich UB, Liautard C, Kirschenbaum D, Pofahl M, Lavigne J, Liu Y, et al. Impaired action potential initiation in GABAergic interneurons causes hyperexcitable networks in an epileptic mouse model carrying a human Na(V) 1.1 mutation. J Neurosci 2014; 34:14874–14889. Schock SC, Jolin-Dahel KS, Schock PC, Theiss S, Arbuthnott GW, GarciaMunoz M, et al. Development of dissociated cryopreserved rat cortical neurons in vitro. J Neurosci Methods 2012; 205:324–333. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992; 55:181–184. Maetzler W, Schmid SP, Wurster I, Liepelt I, Gaenslen A, Gasser T, et al. Reduced but not oxidized cerebrospinal fluid glutathione levels are lowered in Lewy body diseases. Mov Disord 2011; 26:176–181. Walker Z, Possin KL, Boeve BF, Aarsland D. Lewy body dementias. Lancet 2015; 386:1683–1697. Borges LF, Elliott PJ, Gill R, Iversen SD, Iversen LL. Selective extraction of small and large molecules from the cerebrospinal fluid by Purkinje neurons. Science 1985; 228:346–348. Pappenheimer JR. Bayliss-Starling Memorial Lecture (1982). Induction of sleep by muramyl peptides. J Physiol 1983; 336:1–11. Silver R, LeSauter J, Tresco PA, Lehman MN. A diffusible coupling signal from the transplanted suprachiasmatic nucleus controlling circadian locomotor rhythms. Nature 1996; 382:810–813.
Copyright r 2017 Wolters Kluwer Health, Inc. All rights reserved.
