Human iPSC-derived Neurons form Synchronously Bursting Cultures and Display a Seizurogenic Response to Excitatory

Pharmacology

Kile P Mangan, Tromondae K Feaster, Lisa Harms, Elisabeth Enghofer, Christian Kannemeier, Coby B Carlson and Blake D AnsonCellular Dynamics International, Inc., A FUJIFILM Company, Madison, WI USA

The mammalian brain works appropriately only when there is a proper balance between excitation and inhibition. An imbalancein the ratio of excitatory-to-inhibitory neurons (referred to as the E/I ratio) is associated with numerous neurologicalabnormalities and deficits. Increased E/I ratios result in higher excitability and leads to prolonged neocortical circuit activity,stimulus hypersensitivity, cognitive impairments, and even epilepsy (Hagerman, et al. 2002; Gibson, et al., 2008; Zhang, et al.2011). Of equal interest and importance, decreased E/I ratios result in a stronger inhibitory drive and have been linked toimpaired social interactions, autistic behaviors, and mental retardation (Tabuchi, et al. 2007; Dani, et al. 2005). It is well-definedthat during neuronal development that the E/I ratio changes and evolves over time, with excitation decreasing and inhibitionincreasing, and any deviations to this natural process may give rise to neurological disorders.Despite the urgency to advance our understanding of the mechanisms that govern neural network formation and synchronousbursting behaviors, a major challenge in neuroscience research and the identification of new drugs is access to clinically-relevantcell models. The advent of induced pluripotent stem cell (iPSC) technology now grants us access to previously unattainablehuman neuronal cell types. Using this technology, we have generated iPSC-derived frontal cortical neurons of an appropriate,brain-similar E/I ratio (~80% excitation), that develop and display network-level, synapse-coupled, coordinated neuronalactivity in vitro. Using multi-electrode arrays (MEA) to measure the electrophysiological activity of these glutamatergic neurons,we are able to capture periodical and synchronized bursting patterns and analyze these data via numerous parameters (includingPoisson and ISI statistics). These neuronal cultures contain excitatory synapses (as shown by HOMER1 and synapsin co-staining)that modulate the synaptically-driven and synchronous bursting behaviors observed on the MEA. Importantly, these signals canbe blocked by AP5 and DNQX.In the interest of understanding how excitatory pharmacology alters network-level neuronal interactions, we investigatedalterations in the regular bursting properties of glutamatergic neurons after addition of (0.003 to 300 micromolar dose response)various excitatory agents (e.g. bicuculline, chlorpromazine, pentylenetetrazol, amoxapine, 4-aminopyridine, and glutamic acid),common anti-epileptic drugs (e.g. ganaxalone and valproate), as well as a negative control (acetaminophen) and vehicle controls(e.g. ethanol and DMSO). Activity metrics displaying dose-dependent responses with pharmacology include: ‘single-channel’Poisson bursts (rate, intensity and duration), ‘network-level’ ISI bursts (rate, intensity and duration), and synchrony measures.The presented data illustrates how human iPSC-technology can be leveraged to create an unprecedented investigatory space forunderstanding the intricacies of how excitatory synapses control network-connected populations of human neurons.

Abstract

iPSC-Derived Neuronal Cell Types

≥ 70% Glutamatergic population determined by single cell gene expression

≥90% Pure Neurons- Easily implemented in excitotoxic compound screening

≥90% CD44 (+) / S100ß (+)

- Expression of key astrocytic markers

- Clearance of extracellular glutamate

- Stimulation with TNFα, IL-1β and IFNγ upregulates the expression of the pro-inflammatory cytokine IL-6

MIXED INTO CO-

CULTURE!120k GNCs + 20k Astros

iCe

llG

luta

Ne

uro

ns

iCe

llA

stro

cyte

s

Calcein AM EthD-2

Glutamate Uptake

0 50 100

200

400

0

1

2

3

4

5- inhibitor

+ inhibitor

Replace media ( Phenol red free DMEM with glucose and sodium pyruvate (no glutamine/GlutaMax)) with or without 80µl of 200µM DL-TBOA per well.

Add 20µl of media (untreated) or 5X of Glutamic acid (0.5µM or 1mM for 100 & 200µM ea) to each well and gently tap the plate to mix well.Incubate at 37°C for 1 hour.

Rinse with 1X DPBS once and add 40µl of 1X DPBS per wellAdd 5µl of Inactivation solution per well. Shake the plate for 1 minute and incubate for 5 min.Add 5µl of Neutralization solution and shake.

Prepare Glutamate Assay master mix and add 50µl of the mix per well.Incubate in dark for 1 hour and read the luminescence.

Glutamic Acid (M)

Fo

ld C

han

ge

Glutamate Uptake

0 50 100

200

400

0

1

2

3

4

5- inhibitor

+ inhibitor

Replace media ( Phenol red free DMEM with glucose and sodium pyruvate (no glutamine/GlutaMax)) with or without 80µl of 200µM DL-TBOA per well.

Add 20µl of media (untreated) or 5X of Glutamic acid (0.5µM or 1mM for 100 & 200µM ea) to each well and gently tap the plate to mix well.Incubate at 37°C for 1 hour.

Rinse with 1X DPBS once and add 40µl of 1X DPBS per wellAdd 5µl of Inactivation solution per well. Shake the plate for 1 minute and incubate for 5 min.Add 5µl of Neutralization solution and shake.

Prepare Glutamate Assay master mix and add 50µl of the mix per well.Incubate in dark for 1 hour and read the luminescence.

Glutamic Acid (M)

Fo

ld C

han

ge

Network-Level Synchronization Detected via MEAs

Electrode diameter = 50 mInterpolar distance = 350 m48 wells per plate16 electrodes per well 768 electrodes per MEA12.5 kHz sampling rate9.6 million data points per sec

GNC + Astrocytes

Elec

tro

de

#P

eak

Inte

nsi

ty (

Hz)

Minutes

GNC Mono-culture

Elec

tro

de

#P

eak

Inte

nsi

ty (

Hz)

Minutes

Multi-Elecrode Arrays (MEAs) were utilized to assess theactivity levels and bursting behaviors of iCell GlutaNeuron& iCell Astrocyte co-cultures. iPSC-derived neuronal co-cultures develop synchronized bursting behaviorsbeginning near DIV11 and reach robust, re-producible andreliable bursting levels near DIV16. Co-cultures display aslightly 1) more organized and 2) cleaner burstingphenotype compared to mono-cultures. Appropriately, co-culture conditions decrease mean Firing Rate but increasebursting behaviors both at a channel-level (Poisson) as wellas at the network-level (ISI) resolution. We used Axion’sMaestro MEA system and Neural Metric analysis softwareto assess bursting behaviors, along with an in-house all-point histogram (500 msec bins) algorithm (*) that detectsbursting peak time-points (o).

Seizurogenic Assay DevelopmentSeizurogenic Assay workflow, analysis and results are presented for all pharmacology tested, including vehicle, control,excitatory (i.e. seizurogenic) and anti-epileptic (AEDs) drugs. iCell GlutaNeurons and iCell Astrocytes were mixed upon thaw anddotted together onto 48-well MEA plates. Evaluation of seizurogenic pharmacology was performed on DIV 19 or 20 by adding10 µL (30X solutions) to wells containing 300 µL of Brainphys Medium, assessing 6 different concentrations of each drug [0.003,0.03, 0.3, 3, 30, 300 µM]. Each plate contained positive control (bicuculline [200 µM]) treated wells, as well as untreated wells.‘Before’ and ‘Treatment’ 8-minute recordings were collected, with pre-incubation periods of 10 minutes preceding eachrecording. Spike Files (6 SD) were collected and processed for spike and bursting metrics via Axion Neural Metric analysis and byan all-points histogram burst-peak detection suite (CDI NeuroAnalyzer). Differences from baseline were normalized to vehiclecontrol and are presented for all drug concentrations, for each metric. Example analysis flow and raster plots (middle) areshown for bicuculline. Axion’s ‘Area Under Normalized Cross-Correlation’ statistic offers possible *golden variable* needed forseizurogenic qualification. Far Right: Filled radar graphs (increasing concentration going clock-wise) for all pharmacology arepresented depicting absolute value changes from baseline of all metrics. *Note control and vehicle display no changes frombaseline, while seizurogenic pharmacology alters spike and burst metrics >40 fold .

Plate Layout for ALL Plates (µM)

4-A

P [

30

0 µ

M]

Vehicle (H2O) BIC (30 µM)

-40

-30

-20

-10

0

10

20

30

40

mn

FRs

Act

Es

NA

BP

M

NA

Co

nn

chan

B

chan

Bsp

k

chan

BD

ur

nB

nB

urs

tPer

c

nB

urs

tDu

r

nEl

ect

Syn

chro

ny

t0003 t003 t03 t3 t30 t300. . .

Ra

tio

No

rma

lize

d t

o V

ehic

le

Bicuculline [µM]

Graphs depict ‘Ratio Normalized to Vehicle’ for eachconcentration [µM]. This data processing mirrors Axion’sNeural Metric analysis (right) ‘percentage change frombaseline’. Analysis includes variables depicting ‘Spikes’,‘Channel’, ‘Network’ and ‘Synchrony’ measures. Below:Example all-points histogram (i.e. ‘Velocity’) graphs displaychanges in bursting behaviors following addition ofexcitatory pharmacology. *Note variable changes followingpharmacological addition,

Glu

tam

ate

[0.3

µM

]

Glu

tam

ate

[3 µ

M]

60 secs

200 Hz

Ch

lorp

rom

azin

e [3

µM

]

including increased burst duration, increased bursting frequency and / or complete destruction of bursting behaviors.

Bic

ucu

llin

eA

ceta

min

op

hen

Glu

tam

ate

Area Under Normalized Cross-Correlation

4-A

min

op

yrid

ine

Pharmacology Tested

✓ Bicuculline (BIC)✓ Chlorpromazine (CHL)✓ Picrotoxin (PTX)✓ Pentylenetetrazol (PTZ)✓ 4-Aminopyridine (4AP)✓ Glutamatic Acid (GA)

✓ Acetaminophen (ACE)✓ Ethanol (EtOH)✓ Methanol (data not shown)✓ DMSO

Qualifying Compounds: Filled Radar Graphs

Ratio Normalized to Vehicle Treatment

mnFRs ActEsNA BPM NA ConnchanB chanBspkchanBDur nBnBurstPerc nBurstDurnElect Synchrony

Ganaxolone

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Valproate

0

10

20

30

40t00003

t0003

t003

t03

tp3

t3

4-Aminopyridine

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Chlorpromazine

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Phenothiazine

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Glutamate

0

10

20

30

40t0003

t003

t03

t3

t30

t300

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Acetaminophen 1 Acetaminophen 2

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Bicuculline Bicuculline

1 Day

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Picrotoxin

0

10

20

30

40t0003

t003

t03

t3

t30

t300

Picrotoxin

1 Day

0

10

20

30

40t0003

t003

t03

t3

t30

t300

EtOH

0

10

20

30

40t00625

t0125

t025

t05

t1

t2

0

10

20

30

40t00312

t00625

t0125

t025

t05

t1

DMSO

Co

ntro

lsExcitato

ry Co

mp

ou

nd

s (1

0 m

in)

AED

s

0

10

20

30

40t0003

t003

t03

t3

t30

t300

-200 -150 -100 -50 50 100 150 200

-200

-100

100

200

R2 = 0.76

Post WashoutPTX

PTX

D22

D200

D1800

-100

-50

0

50

100

150

200

250

300

mnFRs ActEsNA BPM

NA ConnchanB

chanBspkchanBDur

nBnBurstPerc

nBurstDurnElect

Synchrony

-100

100

300% Changefrom Baseline

SpikesNetwork

‘Velocity’

Network‘ISI’

Channel‘Poisson’

SynchronyIndex

% C

ha

ng

e fr

om

Ba

selin

e

Network BurstFrequency (BPM)

Poisson BurstIntensity (Hz)

Post WashoutPTX (5 Minutes)

Vehicle66.7 33.3 16.7 8.3 4.1 2.2 1.1

1 H

ou

r P

ost

GB

Z (µ

M)

Qualifying Compounds

Summary and Conclusions

• iCell GlutaNeurons and iCell Astroyctes can be mixed together to generate a purely human neuronal co-culture• iCell GlutaNeurons and iCell Astrocytes co-cultures develop synchronized bursting cultures in vitro• Co-cultures develop a robust, reproducible network-level bursting phenotype within 3 weeks• Control and vehicle conditions do not alter co-culture synchronized bursting behaviors, while pharmacology that either

stimulates or ameliorates excitatory pathways does alter bursting behaviors• Co-cultures stimulated with excitatory pharmacology produce ‘seizurogenic phenotypes’