3. Cell Culture Techniques
Second Edition
Edited by
Michael Aschner
DepartmentofMolecularPharmacology,AlbertEinsteinCollegeofMedicine,
Bronx,NY,USA
Lucio Costa
DepartmentofEnvironment/STE100,UniversityofWashington,Seattle,WA,USA
5. v
Experimentallifescienceshavetwobasicfoundations:conceptsandtools.TheNeuromethods
series focuses on the tools and techniques unique to the investigation of the nervous system
and excitable cells. It will not, however, shortchange the concept side of things as care has
been taken to integrate these tools within the context of the concepts and questions under
investigation. In this way, the series is unique in that it not only collects protocols but also
includes theoretical background information and critiques which led to the methods and
their development. Thus, it gives the reader a better understanding of the origin of the
techniques and their potential future development. The Neuromethods publishing pro-
gram strikes a balance between recent and exciting developments like those concerning new
animal models of disease, imaging, in vivo methods, and more established techniques,
including immunocytochemistry and electrophysiological technologies. New trainees in
neurosciences still need a sound footing in these older methods in order to apply a critical
approach to their results.
Under the guidance of its founders, Alan Boulton and Glen Baker, the Neuromethods
series has been a success since its first volume published through Humana Press in 1985.
The series continues to flourish through many changes over the years. It is now published
under the umbrella of Springer Protocols. While methods involving brain research have
changed a lot since the series started, the publishing environment and technology have
changed even more radically. Neuromethods has the distinct layout and style of the Springer
Protocols program, designed specifically for readability and ease of reference in a laboratory
setting.
The careful application of methods is potentially the most important step in the process
of scientific inquiry. In the past, new methodologies led the way in developing new disci-
plines in the biological and medical sciences. For example, physiology emerged out of
anatomy in the nineteenth century by harnessing new methods based on the newly discov-
ered phenomenon of electricity. Nowadays, the relationships between disciplines and meth-
ods are more complex. Methods are now widely shared between disciplines and research
areas. New developments in electronic publishing make it possible for scientists who
encounter new methods to quickly find sources of information electronically. The design of
individual volumes and chapters in this series takes this new access technology into account.
Springer Protocols makes it possible to download single protocols separately. In addition,
Springer makes its print-on-demand technology available globally. A print copy can there-
fore be acquired quickly and for a competitive price anywhere in the world.
Saskatoon, SK, Canada Wolfgang Walz
Preface to the Series
6. vii
Research on the fundamental mechanisms underlying neurodevelopment and optimal brain
function requires comprehensive mechanisms. Some research is carried out in animal mod-
els, but recent efforts have been directed at optimizing tissue culture methods to allow for
the reduction of animal usage. Thus, neuromethodologies serve multiple neuroscience
disciplines.
We have called on a group of respected and internationally recognized researchers, each
with in-depth expertise in implementing specific neuromethod models and techniques, to
contribute to this volume. The chapters are geared to provide technical information as well
as discussions on the requirements, advantages, and limitations of these neuromethods.
The chapters should serve students, experienced researchers, as well as those endowed in
making risk decisions. We hope that the readers will find this volume an important comple-
ment to their current repertoire of techniques and methods in their research on various
aspects of neurosciences.
Bronx, NY, USA Michael Aschner
Seattle, WA, USA Lucio Costa
Preface
7. ix
Preface to the Series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
1 In Vitro Blood-Brain Barrier Functional Assays in a Human
iPSC-Based Model������������������������������������������������������������������������������������������� 1
Emma H. Neal, Yajuan Shi, and Ethan S. Lippmann
2 In Vitro Techniques for Assessing Neurotoxicity Using Human
iPSC-Derived Neuronal Models����������������������������������������������������������������������� 17
Anke M. Tukker, Fiona M. J. Wijnolts, Aart de Groot, Richard W. Wubbolts,
and Remco H. S. Westerink
3 Oxidative Stress Signatures in Human Stem Cell-Derived Neurons ����������������� 37
M. Diana Neely and Aaron B. Bowman
4 Glial Reactivity in Response to Neurotoxins: Relevance and Methods��������������� 51
Lindsay T. Michalovicz and James P. O’Callaghan
5 Neuron-Glia Interactions Studied with In Vitro Co-Cultures��������������������������� 69
S. Mancino, M. M. Serafini, and Barbara Viviani
6 Assessment of Mitochondrial Stress in Neurons: Proximity Ligation
Assays to Detect Recruitment of Stress-
Responsive Proteins
to Mitochondria ��������������������������������������������������������������������������������������������� 87
Monica Rodriguez-Silva, Kristen T. Ashourian, Anthony D. Smith,
and Jeremy W. Chambers
7 Epigenetic Changes in Cultures: Neurons and Astrocytes��������������������������������� 119
David P. Gavin, Xiaolu Zhang, and Marina Guizzetti
8 The Neurosphere Assay as an In Vitro Method for Developmental
Neurotoxicity (DNT) Evaluation��������������������������������������������������������������������� 141
Laura Nimtz, Jördis Klose, Stefan Masjosthusmann, Marta Barenys,
and Ellen Fritsche
9 Zebrafish as a Tool to Assess Developmental Neurotoxicity����������������������������� 169
Keturah G. Kiper and Jennifer L. Freeman
10 Rat Brain Slices: An Optimum Biological Preparation for Acute
Neurotoxicological Studies ����������������������������������������������������������������������������� 195
Gabriela Aguilera-Portillo, Aline Colonnello-Montero, Marisol Maya-
López,
Edgar Rangel-López, and Abel Santamaría
11 Electrophysiological Neuromethodologies������������������������������������������������������� 209
Yukun Yuan and William D. Atchison
12 A Method for Sampling Rat Cerebrospinal Fluid with Minimal Blood
Contamination: A Critical Tool for Biomarker Studies������������������������������������� 233
Zhen He, John Panos, James Raymick, Tetyana Konak, Li Cui,
Diane B. Miller, James P. O’Callaghan, Serguei Liachenko, Merle G. Paule,
and Syed Z. Imam
Contents
8. x
13 Transporter Studies: Brain Punching Technique����������������������������������������������� 245
Cherish A. Taylor and Somshuvra Mukhopadhyay
14 miRNA as a Marker for In Vitro Neurotoxicity Testing
and Related Neurological Disorders����������������������������������������������������������������� 255
Lena Smirnova and Alexandra Maertens
15 Application of Non-Animal Methods to More Effective Neurotoxicity
Testing for Regulatory Purposes ��������������������������������������������������������������������� 283
Anna Bal-Price and Francesca Pistollato
Index��������������������������������������������������������������������������������������������������������������������� 301
Contents
9. xi
Gabriela Aguilera-Portillo • Laboratorio de Aminoácidos Excitadores, Instituto
Nacional de Neurología y Neurocirugía, Mexico City, Mexico
Kristen T. Ashourian • Department of Environmental Health Sciences, Robert Stempel
College of Public Health Social Work, Florida International University, Miami, FL,
USA
William D. Atchison • Department of Pharmacology and Toxicology, Michigan State
University, East Lansing, MI, USA
Anna Bal-Price • European Commission Joint Research Centre, Ispra, Italy
Marta Barenys • INSA·UB and Department of Pharmacology, Toxicology and
Therapeutic Chemistry, Faculty of Pharmacy and Food Sciences, University of Barcelona,
Barcelona, Spain
Aaron B. Bowman • Vanderbilt University Medical Center, Department of Pediatrics,
Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt Kennedy Center,
Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt Brain Institute,
Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt Center for Stem
Cell Biology, Nashville, TN, USA; Vanderbilt University Medical Center, Department of
Neurology, Nashville, TN, USA; Vanderbilt University Medical Center, Department of
Biochemistry, Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt
Center for Molecular Toxicology, Nashville, TN, USA
Jeremy W. Chambers • Department of Environmental Health Sciences, Robert Stempel
College of Public Health Social Work, Florida International University, Miami, FL,
USA; Biomolecular Sciences Institute, Florida International University, Miami, FL,
USA
Aline Colonnello-Montero • Laboratorio de Aminoácidos Excitadores, Instituto
Nacional de Neurología y Neurocirugía, Mexico City, Mexico
Li Cui • Department of Microbiology and Immunology, University of Arkansas for Medical
Sciences, Little Rock, AR, USA
Aart de Groot • Neurotoxicology Research Group, Toxicology and Pharmacology Division,
Institute for Risk Assessment Sciences (IRAS), Faculty of Veterinary Medicine, Utrecht
University, Utrecht, The Netherlands
Jennifer L. Freeman • School of Health Sciences, Purdue University,
West Lafayette, IN, USA
Ellen Fritsche • IUF – Leibniz Research Institute for Environmental Medicine,
Duesseldorf, Germany; Heinrich-Heine-University, Duesseldorf, Germany
David P. Gavin • Jesse Brown Veterans Affairs Medical Center, Chicago, IL, USA;
Department of Psychiatry, Center for Alcohol Research in Epigenetics, University of
Illinois at Chicago, Chicago, IL, USA
Marina Guizzetti • Oregon Health Science University, Portland, OR, USA;
VA Portland Health Care System, Portland, OR, USA
Contributors
10. xii
Zhen He • Division of Neurotoxicology, National Center for Toxicological Research FDA,
Jefferson, AR, USA
Syed Z. Imam • Division of Neurotoxicology, National Center for Toxicological Research
FDA, Jefferson, AR, USA
Keturah G. Kiper • School of Health Sciences, Purdue University, West Lafayette, IN,
USA
Jördis Klose • IUF – Leibniz Research Institute for Environmental Medicine,
Duesseldorf, Germany
Tetyana Konak • Division of Neurotoxicology, National Center for Toxicological Research
FDA, Jefferson, AR, USA
Serguei Liachenko • Division of Neurotoxicology, National Center for Toxicological
Research FDA, Jefferson, AR, USA
Ethan S. Lippmann • Department of Chemical and Biomolecular Engineering, Vanderbilt
University, Nashville, TN, USA; Department of Biomedical Engineering, Vanderbilt
University, Nashville, TN, USA
Alexandra Maertens • Center for Alternatives to Animal Testing, Bloomberg School of
Public Health, Johns Hopkins University, Baltimore, MD, USA
S. Mancino • Cellular and Systems Neurobiology laboratory, CEDOC, Chronic Diseases
Research Centre, NOVA Medical School, Faculdade de Ciências Médicas Universidade
NOVA de Lisboa, Lisbon, Portugal
Stefan Masjosthusmann • IUF – Leibniz Research Institute for Environmental
Medicine, Duesseldorf, Germany
Marisol Maya-López • Laboratorio de Aminoácidos Excitadores, Instituto Nacional de
Neurología y Neurocirugía, Mexico City, Mexico
Lindsay T. Michalovicz • Health Effects Laboratory Division, Centers for Disease
Control and Prevention – National Institute for Occupational Safety and Health,
Morgantown, WV, USA
Diane B. Miller • CDC/NIOSH, Morgantown, WV, USA
Somshuvra Mukhopadhyay • Division of Pharmacology Toxicology, College of
Pharmacy, Austin, TX, USA; Institute for Cellular Molecular Biology, Austin, TX,
USA; Institute for Neuroscience, The University of Texas at Austin, Austin, TX, USA
Emma H. Neal • Department of Chemical and Biomolecular Engineering, Vanderbilt
University, Nashville, TN, USA
M. Diana Neely • Vanderbilt University Medical Center, Department of Pediatrics,
Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt Kennedy Center,
Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt Brain Institute,
Nashville, TN, USA; Vanderbilt University Medical Center, Vanderbilt Center for Stem
Cell Biology, Nashville, TN, USA
Laura Nimtz • IUF – Leibniz Research Institute for Environmental Medicine,
Duesseldorf, Germany
James P. O’Callaghan • Health Effects Laboratory Division, Centers for Disease Control
and Prevention – National Institute for Occupational Safety and Health, Morgantown,
WV, USA; CDC/NIOSH, Morgantown, WV, USA
John Panos • Division of Neurotoxicology, National Center for Toxicological Research
FDA, Jefferson, AR, USA
Merle G. Paule • Division of Neurotoxicology, National Center for Toxicological
Research FDA, Jefferson, AR, USA
Contributors
11. xiii
Francesca Pistollato • European Commission Joint Research Centre, Ispra, Italy
Edgar Rangel-López • Laboratorio de Aminoácidos Excitadores, Instituto Nacional de
Neurología y Neurocirugía, Mexico City, Mexico
James Raymick • Division of Neurotoxicology, National Center for Toxicological Research
FDA, Jefferson, AR, USA
Monica Rodriguez-Silva • Department of Environmental Health Sciences, Robert
Stempel College of Public Health Social Work, Florida International University,
Miami, FL, USA
Abel Santamaría • Laboratorio de Aminoácidos Excitadores, Instituto Nacional de
Neurología y Neurocirugía, Mexico City, Mexico
M. M. Serafini • Department of Pharmacological and Biomolecular Sciences, University
of Milan, Milan, Italy
Yajuan Shi • Department of Chemical and Biomolecular Engineering, Vanderbilt
University, Nashville, TN, USA
Lena Smirnova • Center for Alternatives to Animal Testing, Bloomberg School of Public
Health, Johns Hopkins University, Baltimore, MD, USA
Anthony D. Smith • Department of Environmental Health Sciences, Robert Stempel
College of Public Health Social Work, Florida International University, Miami, FL,
USA
Cherish A. Taylor • Division of Pharmacology Toxicology, College of Pharmacy,
Austin, TX, USA; Institute for Cellular Molecular Biology, Austin, TX, USA;
Institute for Neuroscience, The University of Texas at Austin, Austin, TX, USA
Anke M. Tukker • Neurotoxicology Research Group, Toxicology and Pharmacology
Division, Institute for Risk Assessment Sciences (IRAS), Faculty of Veterinary Medicine,
Utrecht University, Utrecht, The Netherlands
Barbara Viviani • Department of Pharmacological and Biomolecular Sciences,
University of Milan, Milan, Italy
Remco H. S. Westerink • Neurotoxicology Research Group, Toxicology and Pharmacology
Division, Institute for Risk Assessment Sciences (IRAS), Faculty of Veterinary Medicine,
Utrecht University, Utrecht, The Netherlands
Fiona M. J. Wijnolts • Neurotoxicology Research Group, Toxicology and Pharmacology
Division, Institute for Risk Assessment Sciences (IRAS), Faculty of Veterinary Medicine,
Utrecht University, Utrecht, The Netherlands
Richard W. Wubbolts • Centre for Cell Imaging (CCI), Department of Biochemistry
and Cell Biology, Faculty of Veterinary Medicine, Utrecht University, Utrecht,
The Netherlands
Yukun Yuan • Department of Pharmacology and Toxicology, Michigan State University,
East Lansing, MI, USA
Xiaolu Zhang • Oregon Health Science University, Portland, OR, USA; VA Portland
Health Care System, Portland, OR, USA
Contributors
13. 2
(NVU), which consists of pericytes, astrocytes, neurons, and
microglia [4].
Due to its restrictive nature, the BBB represents a major obsta-
cle for delivering drugs to the brain. In addition, dysfunction of
the BBB is implicated in a number of neurodegenerative diseases,
including but not limited to Alzheimer’s disease (AD), Parkinson’s
disease (PD), Huntington’s disease (HD), amyotrophic lateral
sclerosis (ALS), and multiple sclerosis (MS). For the sake of brev-
ity, only AD and PD are described herein. In the two-hit vascular
hypothesis of AD, initial blood vessel damage contributes to BBB
dysfunction and diminished brain perfusion, which can subse-
quently influence amyloid-β (Aβ) accumulation and neuronal
injury [5]. Several neuroimaging results support this notion of
BBB dysfunction in AD. Higher gadolinium leakage, observed by
dynamic contrast-enhanced magnetic resonance imaging (DCE-
MRI), has confirmed increased BBB permeability in several gray
and white matter regions of early AD patients [6]. Also, lobar cere-
bral microbleeds, which indicate BBB damage, are also found in
most AD patients [7, 8]. Whereas glucose transporter member 1
(GLUT-1) is selectively expressed on BBB endothelium,
18
F-fluorodeoxyglucose positron emission tomography (FDG-
PET) studies show diminished glucose uptake and reduced
GLUT-1 levels [9, 10] in humans [11] and transgenic mouse
models [12] of AD, which indicate BBB disruption preceding neu-
rodegeneration. Meanwhile, impaired p-glycoprotein, which clears
xenobiotic compounds across the BBB, may also be involved in the
pathogenesis of AD. On the basis of verapamil-PET results, dimin-
ished p-glycoprotein function is observed in the patients with mild
AD [13] and AD [14]. Meanwhile, PD, the second most common
neurodegenerative disorder in the elderly that is characterized by
aggregation of α-synuclein and degeneration of dopaminergic neu-
rons, also displays elements of BBB dysfunction. Decreased
p-
glycoprotein activity in the midbrain, which indicates a loss of
BBB integrity, has been shown in patients with PD [15].
These issues of drug delivery and possible involvement of the
microvasculature in neurodegeneration have led to the rise of
in vitro BBB models, which are a useful complement to protracted
in vivo studies. Cultured BMECs can be used to assess the relative
permeability of drug candidates and transport rates of biologics
with much higher throughput than similar measurements con-
ducted in animals [16]. Meanwhile, NVU cell-cell interactions and
their respective influence on BBB properties, as well as disease-
relevant perturbations of BMEC function, can also be studied
more easily in an in vitro setting. Unfortunately, in vitro models
have several drawbacks, such as (1) they often do not properly
mimic the phenotype of the BBB in vivo and (2) they can be diffi-
cult to construct owing to a lack of quality cells. Most in vitro BBB
Emma H. Neal et al.
14. 3
models have historically been constructed from primary
animal
BMECs, including bovine [17], porcine [18, 19], rat [20, 21], and
mouse [22]. For comparison, the BBB in vivo has a theoretical
maximum transendothelial electrical resistance (TEER, a measure
of ion permeability) of 8000 Ω × cm2
[23], and primary BMECs
typically exhibit ranges of TEER from ~200 to 1800 Ω × cm2
[24].
Small molecule permeability also indicates reasonable tightness in
these models, and most BMEC sources exhibit active transport
function. However, primary BMECs can only be isolated in low
yield (because the vasculature represents only a small fraction of
total brain volume [25]), and in many cases the isolation proce-
dures are tedious and difficult to carry out. Moreover, especially
with respect to drug screens and disease studies, there are concerns
that species differences may prohibit extrapolation of outcomes to
human biology [26]. These concerns have led to examination of
primary human BMECs for constructing more representative
in vitro models, but because they are only available from surgically
resected tissue or cadavers, such BMECs are not suitable for high-
throughput studies. Immortalized human BMECs have also been
explored to overcome these issues, but the immortalization pro-
cess has a negative effect on the passive barrier phenotype, yielding
TEER values of 30–40 Ω × cm2
that are more similar to peripheral
endothelium than BBB [27].
Recently, endothelial cells possessing a BBB-like phenotype
were derived from human-induced pluripotent stem cells (iPSCs)
[28]. These so-called iPSC-derived BMECs represent an attractive
alternative to primary human and immortalized BMECs. iPSCs
can self-renew indefinitely, thus providing an unlimited source of
BMECs similar to the immortalization procedure but with
improved passive barrier properties. First-generation iPSC-derived
BMECs yielded an average TEER of ~850 Ω × cm2
when co-
cultured with rat astrocytes and also demonstrated other hallmarks
of the BBB, including efflux transporter activity and relative per-
meability to a cohort of small molecules that correlated well with
in vivo uptake in rodents [28]. In subsequent studies, the addition
of retinoic acid (RA) was shown to drastically enhance the passive
barrier phenotype, yielding TEER of ~3000 Ω × cm2
in monocul-
ture and 5000 Ω × cm2
in co-culture with human pericytes, astro-
cytes, and neurons [29]. Further iterations of the differentiation
process have resulted in defined seeding densities and medium,
which accelerate the differentiation process without compromising
barrier function [30, 31]. In this book chapter, we detail the most
recent iteration of the iPSC-to-BMEC differentiation process and
outline the primary methods used to characterize barrier function,
including TEER measurements, permeability measurements, and
assessments of efflux transporter activity.
In vitro BBB Assays
15. 4
2 Materials
1. Any quality iPSC line that is properly karyotyped and exhibits
minimal spontaneous differentiation.
2. mTeSR1 (STEMCELL Technologies, Cat#85850) or E8
medium (Thermo Fisher Scientific, Cat#A1517001). We rou-
tinely use E8 medium but have also used mTeSR1 medium in
past publications.
3. Matrigel (Fisher Scientific, Cat#CB-40230), reconstituted to
1 mg aliquots and frozen at −80 °C. These aliquots are used to
coat plates for both iPSC maintenance and differentiation.
One aliquot of Matrigel is sufficient to coat two full 6-well
plates.
4. Versene (Thermo Fisher Scientific, Cat#15040-066).
5. 6-well tissue culture polystyrene dishes.
6. ROCK inhibitor Y27632 (10 μM; RD Systems, Cat#1254).
Stock solutions of 10 mM are prepared in sterile ddH2O and
stored at −80 °C.
7. E6 medium (Thermo Fisher Scientific, Cat#A1516401).
8. Human endothelial serum-free medium (hESFM) (Thermo
Fisher Scientific, Cat#11111-044).
9. Platelet-poor plasma-derived serum (PDS) (Fisher Scientific,
Cat#AAJ64483AE). 100 mL of PDS is centrifuged at 30,000 g
for 30 min to remove particulates, then sterile-filtered using
0.22 μm pore vacuum bottle-top filters (Fisher Scientific,
Cat#S2GPT05RE) and stored in 5 mL aliquots at −80 °C.
10. Retinoic acid (RA) (10 μM; Sigma-Aldrich, Cat#R2625).
Stock solutions of 10 mM are prepared in DMSO (Sigma-
Aldrich, Cat#D8418) and stored at −80 °C.
11. Basic fibroblast growth factor (bFGF) (20 ng/mL; Peprotech,
Cat#100-18b). Stock aliquots are prepared by adding 1 mg of
bFGF to 10 mL of 5 mM Tris buffer (filter-sterilized, pH ~7.6)
and 20 μL of human albumin solution (Sigma-Aldrich,
Cat#A7223). Five hundred micro liter aliquots are then at
−80 °C.
12. Extracellular matrix (ECM) coating comprised of collagen IV
(Sigma-Aldrich, Cat#C5533) and fibronectin (Sigma-Aldrich,
Cat#F1141). One milligram per milliliter collagen IV stock
solution is prepared by dissolving 5 mg into 5 mL of 0.5 mg/
mL sterile acetic acid and is stored at 4 °C. The final coating is
prepared fresh by mixing 5 parts ddH2O, 4 parts collagen IV
stock solution, and 1 part fibronectin.
13. Endothelial cell culture medium: EM++ consists of hESFM,
1% PDS, 20 ng/mL bFGF, and 10 μM RA, while EM−−
2.1 iPSC
Maintenance
and Differentiation
Emma H. Neal et al.
16. 5
consists of hESFM and 1% PDS. RA is added to EM++ imme-
diately before use. All other components can be added to
EM++ and EM−−, and each medium is stable for at least
2 weeks.
14. Accutase (Thermo Fisher Scientific; Cat#A1110501). Accutase
can be aliquoted into 15 mL conicals and stored at −20 °C.
1. Transwell filters (12-well, 1.12 cm2
, 0.4 μm pore size, PET)
(Fisher Scientific, Cat#07-200-161).
2. EVOM voltohmmeter (World Precision Instruments,
Model#EVOM2).
3. Electrodes or Endohm cell culture cup chamber (World
Precision Instruments, #STX2 or ENDOHM-12).
1. Dilute sodium fluorescein (Sigma-Aldrich, Cat#F6377) to
10 mM in distilled water to create stock solutions. Sterile filter
and store in an opaque box at 4 °C.
2. Dilute fluorescently labeled dextrans to 1 mg/mL in DPBS
(Thermo Fisher Scientific, Cat#14190-144) to create stock
solutions. We frequently use a 3 kDa dextran labeled with
Alexa Fluor 680 (Thermo Fisher Scientific, Cat#D34681) but
other sized dextrans can be employed. Store in an opaque box
at −20 °C.
3. Prepare transport buffer: distilled water with 0.12 M NaCl,
25 mM NaHCO3, 3 mM KCl, 2 mM MgSO4, 2 mM CaCl2,
0.4 mM K2HPO4, 1 mM HEPES, and 0.1% bovine serum
albumin. All salts and albumin can be ordered in powdered
form from any reputable source (Fisher Scientific or
Sigma-Aldrich).
4. Dilute chosen radiolabeled compounds to 0.4 μCi in transport
buffer. Radiolabeled compounds can be ordered from American
Radiolabeled Compounds or PerkinElmer.
5. Scintillation counter.
1. Dilute rhodamine 123 (R123) (Thermo Fisher Scientific;
Cat#R302) to a 10 mM stock solution using sterile distilled
water and store at −20 °C.
2. Dilute cyclosporin A (Fisher Scientific; Cat#11-011-00) to a
10 mM stock solution using sterile DMSO and store at −20 °C.
3. Dilute H2DCFDA (Fisher Scientific, Cat#D399) to a 10 mM
stock solution using sterile DMSO and store at −20 °C.
4. Dilute MK-571 (Sigma-Aldrich; Cat#M7571) to a 10 mM
stock solution using sterile distilled water and store at −20 °C.
2.2 TEER
Measurements
2.3 Permeability
Measurements
2.4 Efflux Activity
Measurements
In vitro BBB Assays
17. 6
5. Prepare lysis buffer (DPBS + 5% TX-100): Triton X-100 (TX-
100) (Fisher Scientific, Cat#9002-93-1). Store at room
temperature.
6. DAPI (Thermo Fisher Scientific, Cat#D1306).
3 Methods
1. iPSCs are maintained in E8 medium on Matrigel-coated plates.
Medium is changed every day, and iPSCs are passaged at a 1:6
to 1:12 ratio into freshly coated plates every 3–4 days when
60–80% confluence is reached. Passage ratios are empirically
determined based on the growth rate of the iPSC line.
2. Upon reaching 60–80% confluence, cells are ready to be seeded
for differentiation.
3. Collect 1 mL of spent medium per well of iPSCs that will be
utilized for seeding. For example, if passaging 3 wells for seeding,
collect 3 mL of spent medium. Place this spent medium in a
plastic conical. Aspirate the remaining medium from each well.
4. Wash each well with 2 mL of DPBS. Aspirate DPBS.
5. Add 1 mL of Accutase to each well. Incubate cells at 37 °C for
3–5 min.
6. Using a P1000 pipet, collect the iPSCs from each well by gen-
tly spraying across the surface two to three times to detach the
cells. Transfer all cells to the collected spent medium to neu-
tralize the Accutase.
7. Pellet cells via centrifugation at 1000 RPM for 4 min. Aspirate
supernatant.
8. Resuspend cells in 1 mL of E8 medium.
9. Determine cell density using a hemocytometer or automated
cell counter, such as a Countess II.
10. Plate cells at a density of 150,000 live cells per well of a 6-well
plate in E8 medium supplemented with 10 μM Y27632.
11. Twenty-four hours after plating, initiate differentiation by
changing medium to E6 medium.
12. Change E6 medium every 24 h for 4 days.
13. After 4 days in E6 medium, change medium to EM++. Do not
change medium for 48 h.
14. Twenty-four hours after changing medium to EM++ (day 5 of
differentiation), prepare Transwell filters and/or plates for
BMEC purification by coating with 200 μL ECM/Transwell
filter or plates at the volumes listed in Table 1.
(a)
If coating plates the day before subculture (e.g., over-
night), we recommend adding an equal volume of water
3.1 Differentiation
of iPSCs to BBB
Endothelium
Emma H. Neal et al.
18. 7
to each ECM-coated plate to prevent the wells from dry-
ing out during the coating process.
(b)
Transwell filters must be coated for a minimum of 4 h at
37 °C. Plates must be coated for a minimum of 1 h at 37 °C.
15. Forty-eight hours after changing medium to EM++ (day 6 of
differentiation), cells are ready for subculture.
16. Aspirate ECM solution from coated plates and Transwell fil-
ters. Leave lids slightly ajar and move to the back of the lami-
nar flow hood. Transwell filters must dry for a minimum of
20 min. Plates must dry for a minimum of 5 min but no more
than 30 min to avoid overdrying the wells.
17. Retrieve cells from the incubator and collect 1 mL of spent
medium from each well to be subcultured. Aspirate remaining
medium.
18. Wash each well with 2 mL of DPBS. Aspirate.
19. Add 1 mL of Accutase to each well and return cells to
incubator.
20. Incubate cells in Accutase for 20–45 min (time varies by iPSC
line used and must be determined empirically) until cells have
lifted off the plate in a single cell suspension.
21. After filters have dried for a minimum of 20 min, rewet filters
with 0.5 mL hESFM/filter.
22. Gently collect cells by spraying across each well two to three
times using a P1000 pipet and transfer the cell suspension to
the collected spent medium.
23. Collect cells via centrifugation at 1000 RPM for 4 min.
24. While cells are centrifuging, aspirate the hESFM from each
filter. Add 1.5 mL of EM++ to each basolateral chamber. Omit
this step if only using plates.
25. Aspirate supernatant from the cell pellet and resuspend cells in
an appropriate volume of EM++ (0.5 mL/Transwell filter; see
Table 1.1 for recommended well plate volumes). For 6- and
12-wells and 12-well Transwell filter inserts, cells are seeded
based on a split ratio:
(a)
1 well of a 6-well plate is split to 1 well of a 6-well plate
(1:1).
(b)
1 well of a 6-well plate is split to 3 wells of a 12-well plate
(1:3) or 3 Transwell filters (1:3).
(c)
For smaller plates (24-, 48-, or 96-wells), seed 1 million
cells/cm2
.
(d)
Multiply split ratio by the working volume found in
Table 1 to arrive at total volume of EM++ in which to
resuspend cells.
In vitro BBB Assays
19. 8
26. Twenty-four hours after subculture, induce barrier by chang-
ing medium from EM++ to EM−−. Begin measuring TEER at
this time, which we typically refer to as day 0.
27. Once the medium is changed from EM++ to EM−−, which
removes the bFGF and RA, the user should wait 24 h for bar-
rier properties to spike. TEER will peak at 24 h and often dip
drastically over 2–3 days, then slowly recover over the course
of a week to reach or exceed the reading from 24 h. Thus,
TEER can be measured longitudinally, but we typically mea-
sure permeability and efflux activity 24 h after removal of bFGF
and RA during the first maximum peak.
1. Sterilize Chopstix using 70% ethanol. Shake Chopstix to dry
ethanol from the surface of the electrodes.
2. Let Chopstix equilibrate in hESFM for 2 min.
3. To measure TEER, position the electrodes such that the
shorter end is above the Transwell filter and the longer end is
below. The short electrode must be fully submerged in
medium. Take care not to poke the filter. Measure the resis-
tance until the reading has fully stabilized. Repeat the mea-
surement at the two other positions on the filter.
4. Subtract each resistance measurement from the resistance mea-
surement of an empty filter to yield the TEER value associated
with the BMEC monolayer. In our experience, there is no dif-
ference in TEER for empty filters that are coated or uncoated
with collagen IV and fibronectin. Then, multiply each TEER
value by the surface area of the filter to yield the final value in
Ω × cm2
.
5. TEER can be measured every 24 h to noninvasively track the
stability of the BMECs over time.
3.2 TEER
Measurements
Table 1
Volumes of ECM coating and cell culture media required for the
subculture phase, based on plate size
Plate type for
subculture
Volume of ECM solution
for coating
Working volume of EC
media for cell culture
6-well 800 μL 2 mL
12-well 250 μL 1 mL
24-well 200 μL 500 μL
48-well 100 μL 400 μL
96-well 50 μL 200 μL
Emma H. Neal et al.
20. 9
1. Prepare a working solution of 10 μM sodium fluorescein by
diluting the stock solution. Then, further dilute working solu-
tion 1:80 in a separate conical by adding 62.5 μL of working
solution to 4.94 mL of EM−−.
2. Using the dilute solution prepared above, construct a standard
curve according to the specifications in Table 2.
3. Place each dilution into a well of a 96-well plate.
4. Upon collection of all permeability samples, measure the fluo-
rescence of each well using a plate reader. Construct a standard
curve by plotting fluorescence versus moles.
5. To prepare BMECs, aspirate medium from the apical and
basolateral sides of each Transwell filter and replace with fresh
EM−−. Let the cells equilibrate for 1 h to allow TEER to sta-
bilize after the media change.
(a)
Three Transwell filters seeded with BMECs are required to
calculate an average permeability.
(b)
One Transwell filter coated with ECM but not seeded with
cells should be used per experiment to account for mass
transfer resistance from the filter alone.
6. After allowing the filters to equilibrate, measure TEER to
ensure the monolayers are still intact.
7. Aspirate medium from the apical chamber of each filter and
add 0.5 mL of sodium fluorescein solution. Immediately
remove 200 μL from the basolateral chamber of each well
being tested and transfer it to the 96-well plate containing the
wells prepared to construct the standard curve. Add 200 μL of
fresh EM−− to each basolateral chamber and return all plates
to the incubator.
8. Every 30 min, remove 200 μL of media from each basolateral
chamber, transfer the medium to the 96-well plate, and add
200 μL of fresh media to each basolateral chamber to replace
the volume removed. Media replenishment in the basolateral
chamber is necessary to maintain hydrostatic pressure on the
top and bottom of the filter, which could otherwise influence
mass transfer rates.
9. Continue the assay for a total of 2 h (five total time points
measured, including t = 0).
10. At the conclusion of the assay, measure the fluorescence in
each well of the 96-well plate using a plate reader.
(a)
For sodium fluorescein, λexcitation = 460 nm and
λemission = 515 nm.
11. Calculate the average fluorescence at each time point.
3.3 Permeability
Measurements
3.3.1 Sodium
Fluorescein Permeability
In vitro BBB Assays
21. 10
12. Convert fluorescence to moles for each time point using the
calibration curve.
(a)
Remember that fluorescence from each sample is indica-
tive of the 1.5 mL volume in the basolateral chamber
where the sample was collected, while the calibration curve
was constructed using volumes of 200 μL. Thus, the fluo-
rescence of each sample needs to be scaled appropriately
based on these volumetric differences to properly deter-
mine the number of moles.
(b)
Remember to account for the fact that you removed fluo-
rescein from the basolateral chamber each time you took a
measurement, which would result in a decrease in fluores-
cence. To correct each time point, add the moles from the
previous time point multiplied by (0.2/1.5) to account for
the fraction removed.
13. Plot moles versus time for each condition as well as the empty
filter. Calculate the slope of the linear line of best fit for each
condition and the empty filter. Slope will be in units of moles
per time.
14. Using the assumption that the concentration of sodium fluo-
rescein remains relatively constant in the apical chamber due to
the restricted flux across the monolayer and the short time
frame of the assay, the permeability coefficient (P) can be cal-
culated as follows:
dQ
dT
P S
= ( )( )( )
Cd
where dQ/dT is the slope calculated in the previous step, S is
the area of the Transwell filter, and Cd is the concentration of
fluorescein in the apical chamber.
(a)
Ensure the use of consistent units when calculating P. In
our experience, inconsistent units are the most common
route for misinterpreting the results in this assay.
Table 2
Dilution ratios and expected fluorescein compositions used to construct
a calibration curve for permeability experiments
EM−− (μL) Fluorescein (μL) Fluorescein (μmol) Fluorescein (μM)
0 200 0.000025 0.125022616
50 150 0.00001875 0.093766962
100 100 0.0000125 0.062511308
150 50 0.00000625 0.031255654
200 0 0 0
Emma H. Neal et al.
22. 11
15. To obtain the permeability of the BMEC monolayer, calculate
Pe using the following equation:
1 1 1
P P P
e t f
= -
where Pe is the permeability of the monolayer, Pf is the perme-
ability of the empty filter, and Pt is the collective permeability
of cells cultured on top of filters. If the iPSC-derived BMEC
monolayer is high quality, expected permeability for fluores-
cein is in the range of 10−7
cm/s.
1. The permeability of fluorescently labeled dextrans (such as
Thermo Fisher Scientific #D34682) can be determined similar
to sodium fluorescein described in Sect. 3.3.1. We provide
details on how to calculate permeability for 3 kDa dextran
below. For other molecular weight dextrans conjugated with
different fluorophores, working concentrations that produce
appropriate fluorescent signal for permeability assays must be
empirically determined. This can be accomplished by assessing
the permeation of the fluorescent molecule through a mono-
layer of BMECs at various concentrations. If no fluorescent
signal is observed in the basolateral chamber after 1 h, the con-
centration of starting material in the apical chamber needs to
be increased such that transport of the dextran can be reliably
quantified.
2. If using the 3 kDa labeled dextran listed above, dilute the
1 mg/mL dextran stock solution to 4.11 μM by adding 125 μL
of stock solution to 10 mL EM−−. This working solution
should be prepared fresh for each experiment.
3. Further dilute the dextran working solution by adding 152 μL
of working solution to 4.85 mL of EM−−.
4. Use this diluted working solution to prepare serial dilutions as
described in Sect. 3.3.1. These serial dilutions will be used to
construct a calibration curve as described in Table 2.
5. Using the working solution prepared in Sect. 3.3.2., conduct
the permeability assay and subsequent permeability calcula-
tions in the same manner as the sodium fluorescein experi-
ments in Sect. 3.3.1. Remember that the initial concentration
of dextran will be different compared to the sodium fluores-
cein calculations.
1. Twenty-four hours after barrier induction, aspirate medium
from apical and basolateral chambers of Transwell filters to be
assayed.
2. Wash the apical chamber twice with 0.5 mL of transport buffer
pre-warmed to 37 °C. Add 0.5 mL to each apical chamber and
1.5 mL of transport buffer to each basolateral chamber.
3.3.2 Fluorescently
Labeled Dextran
Permeability
3.3.3 Radiolabeled
Compound Permeability
In vitro BBB Assays
23. 12
3. Incubate cells for 1 h at 37 °C to allow TEER to stabilize.
4. Immediately prior to beginning the assay, measure TEER as
previously described.
5. Dilute chosen radiolabeled compounds to 0.4 μCi in transport
buffer. Total volume will depend on the number of filters being
assayed. Take 200 μL of this stock solution and place in a scin-
tillation vial. It represents the initial concentration of radiola-
beled compound that will be used to calculate the permeability
coefficient.
6. Aspirate transport buffer from apical chamber and replace with
500 μL of the radiolabeled compound solution.
7. Immediately remove 200 μL of transport buffer from the baso-
lateral chamber, transfer to a scintillation vial, and replace with
fresh transport buffer. Return cells to incubator.
8. Remove 200 μL of transport buffer from the basolateral cham-
ber every 15 min for 1 h. Transfer each aliquot to a scintilla-
tion vial.
9. At the end of the experiment (typically 1 h), dilute each scintil-
lation vial with 800 μL of transport buffer and measure radio-
activity on a scintillation counter.
10. Calculate Pe as described in Sect. 3.3.1. A calibration curve
does not need to be constructed. Plot counts per minute
(CPM) as a function of time and calculate the linear slope and
then divide by the filter area and CPM of the initial solution
(adjusted to a per mL basis) to yield the combined permeabil-
ity coefficient for BMECs and the filter. The permeability coef-
ficient of the BMECs can then be calculated by removing the
mass transfer resistance of an empty filter, as described earlier.
1. As described earlier, iPSC-derived BMECs are seeded onto
24-well plates in EM++ for 24 h and changed to EM−− for
24 h to obtain maximum barrier induction. Four wells are
required for each experimental condition.
2. Stock solutions of efflux inhibitors (cyclosporin A and MK-571)
are diluted in EM−− to concentrations of 10 μM. BMECs are
then incubated with each inhibitor for 1 h at 37 °C. Control
BMECs are left untouched.
3. While the BMECs are incubating with inhibitor, prepare two
solutions, one containing EM−− and 10 μM of efflux trans-
porter substrate and one containing EM−−, 10 μM substrate
and 10 μM inhibitor. Make sure the substrate and inhibitor are
appropriately matched for each efflux transporter. Three milli-
liters of media is required per experimental condition.
4. Media is aspirated from each well, and BMECs are then incu-
bated with 10 μM R123 or 10 μM H2DCFDA with or without
3.4 Efflux
Transporter Activity
3.4.1 Substrate
Accumulation
Emma H. Neal et al.
24. 13
their respective inhibitors for 1 h at 37 °C. The volume in each
well is 0.5 mL. Media containing inhibitors should be added to
wells that were preincubated with inhibitors.
5. At the culmination of the experiment, three wells of BMECs
per condition are washed three times with DPBS, then lysed
with 200 μL of lysis buffer for 10 min, and transferred to a
96-well plate. Fluorescence is measured using a plate reader.
6. The remaining wells (one per condition) are fixed with 100%
ice-cold methanol for 10 min, washed three times with PBS,
stained with DAPI, washed three times again with DPBS, and
imaged. Six images should be acquired per well in varying loca-
tions for each condition. Calculate the number of cells per well
based on the surface area, and normalize fluorescent com-
pound uptake on a per cell basis. If efflux transporters are
expressed and functionally active, R123 and H2DCFDA
uptake should be increased in the presence of each respective
transporter inhibitor.
1. As described earlier, iPSC-derived BMECs are seeded onto
Transwell filters in EM++ for 24 h and changed to EM−− for
24 h to obtain maximum barrier induction.
2. Efflux inhibitors (cyclosporin A and MK-571) are diluted to
concentrations of 10 μM by removing 200 μL of EM−− from
the side of the filter where transport is being assessed, adding
an appropriate volume of efflux inhibitor to produce a 10 μM
concentration for the total volume on the desired side of the
filter, and returning the medium to its original location.
BMECs are then incubated with each inhibitor for 1 h at
37 °C. Inhibitors are only added on the side of the filter where
directional transport is being assessed. For example, if the user
is examining polarized transport in the apical-to-basolateral
direction, inhibitor should only be added on top of the filter.
Control BMECs are left untouched.
3. While the BMECs are incubating with inhibitor, prepare two
solutions, one containing EM−− with 10 μM of efflux trans-
porter substrate and one containing EM−− with 10 μM sub-
strate and 10 μM inhibitor. Make sure the substrate and
inhibitor are appropriately matched for each efflux transporter.
If performing apical-to-basolateral transport studies, 2 mL of
media will be required per condition. If performing basolateral-
to-
apical transport studies, 5 mL of media will be required per
condition.
4. Media is aspirated from the desired compartment and replaced
with media containing 10 μM R123 or 10 μM H2DCFDA
with or without their respective inhibitors and incubated for
1 h at 37 °C. Media containing inhibitors should be matched
to the compartments that were preincubated with inhibitors.
3.4.2 Directional
Transport
In vitro BBB Assays
25. 14
5. After 1 h, 200 μL of media is extracted from the opposite com-
partment of each filter and transferred to a 96-well plate.
TEER should be measured across each filter to ensure that the
BMEC monolayer remains intact. Fluorescence is then mea-
sured on a plate reader as described earlier. If efflux transport-
ers are expressed at the apical or basolateral membrane and
functionally active, R123 and H2DCFDA directional trans-
port should be increased in the presence of each respective
inhibitor.
4 Notes
1. We have qualitatively noticed that maximum TEER values
become lower as BMECs are differentiated from higher pas-
sage iPSCs, possibly due to the accumulation of mutations or
epigenetic modifications. We typically assign a hard cutoff at
passage 40 for most of our iPSC work.
2. When measuring TEER, resistance will become artificially
inflated as the plates cool down from 37 °C. To prevent bias,
we recommend measuring resistances on filters in a random
order or working as quickly as possible.
3. The quality of PDS has a substantial influence on the differen-
tiation process. Side-by-side comparisons of different lot num-
bers, tested on cultures differentiated from identical iPSC
populations, have yielded BMECs with 1000–2000 Ω × cm2
variations. A new lot of PDS should always be qualified with
pilot differentiations before moving forward with scaled-up
experiments.
Acknowledgments
Our research efforts in this area are supported by a NARSAD
Young Investigator Award from the Brain and Behavior Research
Foundation (ESL) and grant A20170945 from the Alzheimer’s
Disease Research Program through the BrightFocus Foundation
(ESL). EHN is supported by a National Science Foundation
Graduate Research Fellowship.
References
1. Abbott NJ, Rönnbäck L, Hansson E (2006)
Astrocyte–endothelial interactions at the
blood–brain barrier. Nat Rev Neurosci
7:41–53
2. Daneman R, Prat A (2015) The blood–brain
barrier. Cold Spring Harb Perspect Biol
7:a020412
3. Obermeier B, Daneman R, Ransohoff RM
(2013) Development, maintenance and disrup-
tion of the blood-brain barrier. Nat Med
19:1584–1596
4. Zlokovic BV (2008) The blood-brain barrier in
health and chronic neurodegenerative disor-
ders. Neuron 57:178–201
Emma H. Neal et al.
26. 15
5. Montagne A, Nation DA, Pa J et al (2016)
Brain imaging of neurovascular dysfunction in
Alzheimer’s disease. Acta Neuropathol (Berl)
131:687–707
6. van de Haar HJ, Burgmans S, Jansen JFA et al
(2016) Blood-brain barrier leakage in patients
with early Alzheimer disease. Radiology
281:527–535
7. Yates PA, Desmond PM, Phal PM et al (2014)
Incidence of cerebral microbleeds in preclinical
Alzheimer disease. Neurology 82:1266–1273
8. Uetani H, Hirai T, Hashimoto M et al (2013)
Prevalence and topography of small hypointense
foci suggesting microbleeds on 3T susceptibility-
weighted imaging in various types of dementia.
Am J Neuroradiol 34:984–989
9. Simpson IA, Chundu KR, Davies-Hill T et al
(1994) Decreased concentrations of GLUT1
and GLUT3 glucose transporters in the brains
of patients with Alzheimer’s disease. Ann
Neurol 35:546–551
10. Jagust WJ, Seab JP, Huesman RH et al (1991)
Diminished glucose transport in Alzheimer’s
disease: dynamic PET studies. J Cereb Blood
Flow Metab 11:323–330
11. Hunt A, Schönknecht P, Henze M et al (2007)
Reduced cerebral glucose metabolism in
patients at risk for Alzheimer’s disease.
Psychiatry Res Neuroimaging 155:147–154
12. Niwa K, Kazama K, Younkin SG et al (2002)
Alterations in cerebral blood flow and glucose
utilization in mice overexpressing the amyloid
precursor protein. Neurobiol Dis 9:61–68
13. Deo AK, Borson S, Link JM et al (2014)
Activity of P-glycoprotein, a β-amyloid trans-
porter at the blood-brain barrier, is compro-
mised in patients with mild Alzheimer disease.
J Nucl Med 55:1106–1111
14. van Assema DM, Lubberink M et al (2012)
Blood–brain barrier P-glycoprotein function in
Alzheimer’s disease. Brain 135:181–189
15. Kortekaas R, Leenders KL, van Oostrom JCH
et al (2005) Blood–brain barrier dysfunction in
parkinsonian midbrain in vivo. Ann Neurol
57:176–179
16. Naik P, Cucullo L (2012) In vitro blood–brain
barrier models: current and perspective tech-
nologies. J Pharm Sci 101:1337–1354
17. Dehouck M-P, Méresse S, Delorme P et al
(1990) An easier, reproducible, and mass-
production method to study the blood–brain
barrier in vitro. J Neurochem 54:1798–1801
18. Franke H, Galla H-J, Beuckmann CT (1999)
An improved low-permeability in vitro-model
of the blood–brain barrier: transport studies on
retinoids, sucrose, haloperidol, caffeine and
mannitol. Brain Res 818:65–71
19. Patabendige A, Skinner RA, Abbott NJ (2013)
Establishment of a simplified in vitro porcine
blood–brain barrier model with high transendo-
thelial electrical resistance. Brain Res 1521:1–15
20. Al Ahmad A, Taboada CB, Gassmann M et al
(2011) Astrocytes and pericytes differentially
modulate blood–brain barrier characteristics
during development and hypoxic insult.
J Cereb Blood Flow Metab 31:693–705
21. Al Ahmad A, Gassmann M, Ogunshola OO
(2009) Maintaining blood–brain barrier integ-
rity: pericytes perform better than astrocytes
during prolonged oxygen deprivation. J Cell
Physiol 218:612–622
22. Song L, Pachter JS (2003) Culture of murine
brain microvascular endothelial cells that main-
tain expression and cytoskeletal association of
tight junction-associated proteins. In Vitro
Cell Dev Biol Anim 39:313–320
23. SmithQR,RapoportSI(1986)Cerebrovascular
permeability coefficients to sodium, potassium,
and chloride. J Neurochem 46:1732–1742
24. Helms HC, Abbott NJ, Burek M et al (2016)
In vitro models of the blood–brain barrier: an
overview of commonly used brain endothelial
cell culture models and guidelines for their use.
J Cereb Blood Flow Metab 36:862–890
25. Pardridge WM (1998) CNS drug design based
on principles of blood-brain barrier transport.
J Neurochem 70:1781–1792
26. Syvänen S, Lindhe Ö, Palner M et al (2009)
Species differences in blood-brain barrier trans-
port of three positron emission tomography
radioligands with emphasis on P-glycoprotein
transport. Drug Metab Dispos 37:635–643
27. Weksler BB, Subileau EA, Perrière N et al
(2005) Blood-brain barrier-specific properties
of a human adult brain endothelial cell line.
FASEB J 19:1872–1874
28. Lippmann ES, Azarin SM, Kay JE et al (2012)
Derivation of blood-brain barrier endothelial
cells from human pluripotent stem cells. Nat
Biotechnol 30:783–791
29. Lippmann ES, Al-Ahmad A, Azarin SM et al
(2014) A retinoic acid-enhanced, multicellular
human blood-brain barrier model derived from
stem cell sources. Sci Rep 4:4160
30. Wilson HK, Canfield SG, Hjortness MK et al
(2015) Exploring the effects of cell seeding
density on the differentiation of human plu-
ripotent stem cells to brain microvascular endo-
thelial cells. Fluids Barriers CNS 12:13
31. Hollmann EK, Bailey AK, Potharazu AV et al
(2017) Accelerated differentiation of human
induced pluripotent stem cells to blood–brain
barrier endothelial cells. Fluids Barriers CNS
14:9
In vitro BBB Assays
28. 18
body. As soon as dendrites receive an excitatory chemical signal,
the neuron becomes activated and translates this chemical input
signal in an electrical signal, an action potential (AP). Via open-
ing of voltage-gated sodium and potassium channels, the AP
induces a change in membrane potential that travels via the cell
body along the axon to the synapse at the axon terminal. There,
the electrical signal (AP) is converted into a chemical signal that
will be transferred to (an)other neuron(s). The first step in con-
version from the electrical to the chemical signal involves open-
ing of voltage-gated calcium channels (VGCC), resulting in a
strong influx of calcium ions (Ca2+
). The resulting changes in the
intracellular Ca2+
concentration ([Ca2+
]i) are involved in a variety
of cellular processes such as excitability, plasticity, motility, and
viability [1, 2]. [Ca2+
]i is crucial for the regulation of neurotrans-
mission as Ca2+
influx through VGCCs triggers the release of
neurotransmitters from the presynaptic cell into the synaptic
cleft [2–4]. Neurotransmitters are chemical signaling molecules
that are stored in vesicles in the presynaptic neuron. There are
different types of excitatory and inhibitory neurotransmitters,
such as acetylcholine, dopamine, serotonin, glutamate, and
gamma-aminobutyric acid (GABA). After release into the synap-
tic cleft by fusion of the vesicles with the presynaptic plasma
membrane, neurotransmitters can bind to receptors on the post-
synaptic membrane. In the receiving cell, the signal can then be
converted in a new AP, or it can activate intracellular signaling
pathways. The chemical signal is terminated by degradation or
reuptake of the neurotransmitters from the synaptic cleft (for
review see [2, 4]).
Communication in neuronal networks thus critically depends
on the structure of neurons, intact neuronal membranes, and
regulation of cellular and molecular mechanisms underlying
neurotransmission. Additionally, proper neuronal communica-
tion also depends on supporting cells such as oligodendrocytes,
astrocytes, and microglia. The multicellular nature of the net-
works can be confirmed with techniques as immunocytochemis-
try, whereas proper intracellular signaling can be studied with
imaging techniques focusing on intracellular calcium levels and
the membrane potential. Finally, the resulting network activity
can be assessed with the use of multi-well microelectrode arrays
(mwMEA).
Its complexity and poor regenerative capacity make the ner-
vous system vulnerable to toxic insults caused by chemical, physi-
ological, and biological agents that are present in the surrounding
environment. Neurotoxicity is thus defined as an adverse effect
caused by any of these agents on the structure and/or function of
the nervous system. Nowadays, there is a broad range of approaches
and cell models to study neurotoxicity in vitro.
Anke M. Tukker et al.
29. 19
In vitro cell models should mimic the in vivo situation as closely as
possible. For neurotoxicity testing, this means that the in vitro
model must form functional neuronal networks with both inhibi-
tory and excitatory neurons as well as supporting cells. In order to
circumvent interspecies translation, cells from human origin are
the preferred option. Recently, human induced pluripotent stem
cell (hiPSC)-derived neurons became commercially available. A
benefit of these cells is that they do not require long differentiation
into neural progenitor cells and ultimately into functional neurons,
which can take several weeks [5, 6] till months [7, 8]. It has been
shown that these hiPSC-derived neurons exhibit the behavior and
function of mature neurons [7, 8]. Therefore, we chose to use
mixed hiPSC-derived neuronal models for the techniques described
in this chapter.
First of all, it is important to determine whether the cultured
cells form neuronal networks. This can be studied using fluores-
cent antibodies and confocal microscopy to detect specific target
proteins in the cell or on the cell membrane. Besides studying net-
work formation and complexity, immunofluorescent stainings can
be used to differentiate between (neuronal) cell types in
co-cultures.
Once the cells formed mixed neuronal networks, spontaneous
network activity and the effect of toxic insults can be studied by
looking at intracellular calcium homeostasis. Changes in [Ca2+
]i
can be analyzed by loading the cultured cells with a high-affinity
Ca2+
-responsive fluorescent dye. Similarly, fluorescent voltage-sen-
sitive dyes can be used to study changes in membrane potential as
an indication for the occurrence of electrical activity.
Another way to look at spontaneous neuronal network activity
and (network) bursting and toxic effects hereon is by the use of
electrophysiological methods. The introduction of mwMEAs pro-
vided a way to grow cells on a culture surface with an integrated
array of microelectrodes. This allows for simultaneous and nonin-
vasive recording of extracellular local field potentials at a millisec-
ond time scale at different locations in the network grown in vitro
(for review see [9]). Mammalian neuronal networks grown in vitro
on mwMEA display many characteristics of in vivo neurons, includ-
ing the development of spontaneous neuronal activity [10] and
synchronized bursting [11]. It has also been shown that these net-
works are responsive to neurotransmitters [12], indicating the
presence of a wide range of common neurotransmitter receptors.
This technique offers consistent reproducibility across different
laboratories [13, 14] and a high sensitivity and specificity [15, 16].
For these reasons mwMEAs are seen as a suitable and consistent
in vitro neurotoxicity screening method. Because measurements
take place in a sterile environment, this technique allows for both
acute [17–19] and chronic toxicity screening [20]. Most mwMEA
research was done with rat primary cortical cultures [21–24], but
1.1 Methods
to Study Neuronal
Network
Communication
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
30. 20
recently it has been shown that also hiPSC co-cultures grow on
mwMEA plates, develop spontaneous activity and bursting behav-
ior, and are suitable for neurotoxicity screening [25–27].
2 Materials
For all techniques described in this chapter, we used commercially
available hiPSC-derived neurons (iCell® Neurons and iCell®
Glutaneurons) and astrocytes (iCell® Astrocytes). All cells were
obtained from Cellular Dynamics International (Madison, WI,
USA). Cells were cultured at 37 °C in a humidified 5% CO2 incu-
bator. From previous experiments, we know that these cells grow
better on polyethylenimine (PEI)-coated surfaces compared to
poly-l-lysine (PLL)- or poly-l-ornithine (PLO)-coated materials.
We therefore pre-coated all our cell culture surfaces with 0.1% PEI
solution diluted in borate buffer (24 mM sodium borate/50 mM
boric acid in Milli-Q, pH adjusted to 8.4) unless stated otherwise.
Co-cultures were grown in BrainPhys™ medium supplemented
with 2% iCell Neuron supplement, 1% nervous system supplement,
1% penicillin-streptomycin, 1% N2 supplement, and 0.1% laminin
(L2020 Sigma-Aldrich, Zwijndrecht, The Netherlands). Astrocytes
(iCell®) were cultured in astrocyte medium (DMEM with high
glucose and 10% FBS, 1% N2 supplement, and 1%
penicillin-streptomycin).
Young astrocytes can proliferate rapidly and potentially over-
grow neuronal cultures. Astrocytes were therefore passaged two to
three times and stored in liquid nitrogen until use in co-culture
with the hiPSC-derived neurons. Young astrocytes were thawed by
gently swirling them for 2–3 min in a 37 °C water bath. Then, the
content of the vial was transferred to a sterile 50 mL tube. The vial
was rinsed three times with astrocyte medium. Total volume in the
50 mL tube was brought to 10 mL, and cells were centrifuged for
5 min at 1300 rpm. The cell pellet was dissolved in 6 mL astrocyte
medium, and cells were transferred to a 25 cm3
Geltrex™-coated
culture flask (Geltrex™ was added to cover the bottom of the flask
and incubated for 45–60 min at 37 °C in a humidified incubator,
after which the Geltrex™ was removed). These flasks were used for
a total of 1–1.4 × 106
cells. Astrocyte medium was replaced every
3–4 days. Cells were passaged by removing the medium, rinsing
with PBS and adding 1 mL 0.0125% trypsin for ~5 min to the
flask. During trypsin incubation the flask was placed at 37 °C in a
humidified 5% CO2 incubator. It is important to carefully check
that cells are detached since astrocytes adhere strongly to the flask
and may require more than 5 min incubation. Once the cells were
detached, 9 mL of medium was added to the cells. Cells were
counted and centrifuged for 5 min at 1300 rpm. For passaging,
3–4.2 × 106
cells were transferred to a 75 cm3
Geltrex™-coated
culture flask, and the volume was brought to 20 mL. In case cells
2.1 Cell Culture
Anke M. Tukker et al.
31. 21
were used for co-culturing and not for passaging, the cell pellet was
dissolved in Complete iCell Neurons Maintenance Medium sup-
plemented with 1% penicillin-
streptomycin, 2% iCell Neuron
medium supplement, and 1% laminin (L2020, Sigma-Aldrich,
Zwijndrecht, The Netherlands) into a 14,000 cells/μL solution.
Procedures for thawing of iCell® Neurons and iCell®
Glutaneurons are comparable. The vial with neurons was thawed
by gently swirling it for 2–3 min in a 37 °C water bath. The cell
suspension was transferred to a 50 mL tube, and the vial was rinsed
three times with Complete iCell Neurons Maintenance Medium.
The total volume in the 50 mL tube was brought to 10 mL, and
the tube was turned upside down twice. A sample for cell counting
was taken, and cells were centrifuged for 5 min at 1300 rpm. The
pellet of iCell® Neurons was dissolved in dotting medium (i.e.,
supplemented BrainPhys™ medium with 10% laminin) to a solu-
tion of 14,000 cells/μL. In parallel, iCell® Glutaneurons were dis-
solved in dotting medium to a solution of 13,000 cells/μL.
When all cells were thawed or detached, we created two types
of co-culture models that differ in the ratio of inhibitory neurons
by addition or absence of iCell® Neurons to create different pro-
files of neuronal activity (see [25] for details). First, a mixture was
made of ~13% iCell® Astrocytes, ~17% iCell® Neurons, and ~70%
iCell® Glutaneurons (culture model A) according to Table 1. Cells
were plated in 10 μL droplets (Table 2). In order to get the total
Table 1
Composition of cell models and plating density
Culture Cell type and % N/well Solution μL/well
A ~13% astrocytes 10,000 14,000 cells/μL 0.71 μL
~17% iCell® Neurons 13,000 14,000 cells/μL 0.93 μL
~70% iCell® Glutaneurons 52,000 13,000 cells/μL 4 μL
B ~15% astrocytes 11,250 14,000 cells/μL 0.80 μL
~85% iCell® Glutaneurons 63,750 13,000 cells/μL 4.55 μL
Table 2
Cell culture surfaces and medium volume
Culture surface Experiment Total volume
48-well MEA plate (Axion
Biosystems Inc., Atlanta,
GA, USA)
MEA recording 300 μL/well
μ-slide 8-well chambered coverslip
(ibidi GmbH, Planegg, Germany)
Immunocytochemistry 200 μL/well
Glass-bottom dishes (MatTek,
Ashland MA, USA)
Live fluorescence
imaging
2 mL/dish
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
32. 22
volume to 10 μL, 4.36 μL dotting medium was added. For culture
model B, a mixture was made of ~15% iCell® Astrocytes and ~85%
iCell® Glutaneurons (according to Table 1). In this case, 4.65 μL
dotting medium was added to bring the total plating volume to
10 μL. Following plating, droplets were allowed to adhere for 1 h
after which medium was added (Table 2). On DIV1, 50% of the
medium was refreshed with room temperature (RT) supplemented
BrainPhys™ medium. Hereafter, 50% medium changes took place
three times a week up till DIV23.
Since both co-culture models require several days to develop
functional neuronal networks, MEA and imaging experiments
should not be performed before DIV11. In our experience, the
optimum window for performing MEA and live imaging experi-
ments ranges from DIV14 to DIV23. It should be noted that the
optimum window for measurements differs between the various
available commercial models as well as culture conditions (e.g., cell
density and % astrocytes).
It should be noted that using a high ratio of astrocytes may
cause the cells to cluster, complicating imaging and MEA experi-
ments. Developmental curves can be measured from DIV4 onward
and can be used to determine the optimum window for assessing
acute neurotoxicity. We strongly recommend to always measure a
developmental curve before performing neurotoxicity assessment
when starting to work with new cells and/or new culture
protocols.
Immunofluorescent images of chemically fixed samples were cap-
tured with a Leica DMI4000 TCS SPEII confocal microscope. To
capture images, a 20× oil immersion objective (ACS APO IMM
NA 0.6) was used. The 20× objective allows for visualization of
multiple neurons in one frame with the connecting dendrites and
axons being clearly visible. When a 10× objective is used, too many
details and sensitivity are lost, whereas a 40× objective does not
capture the complexity of the network structure. It is recom-
mended to visually scan the complete chamber to make sure the
selected area is representative for the culture. We noticed that the
neuronal co-cultures tend to grow differently at the periphery of
the culture area compared to the center. In order to capture an
area that best matches the region where MEA measurements take
place, a region in the middle of the dish is chosen. Once a repre-
sentative area is found, it’s recommended to define an upper and
lower limit in order to make a picture based on z-stacks. We recom-
mend imaging a z-stack series, since the elaborate extensions of the
cells are poorly captured in a single plane. An axially extended view
provided by a maximal intensity projection visualizes the structures
better. Images were captured as .lif files using Leica Application
Suite Advanced Fluorescence software (LAS AF version 2.6.0;
Leica Microsystems GmbH, Wetzlar, Germany).
2.2 Confocal
Microscopy
for Immunocyto-
chemistry
Anke M. Tukker et al.
33. 23
Live changes in [Ca2+
]i were monitored with the fluorescent dye
Fura-2AM (Ex 340 and 380/Em 510, Life Technologies, Bleiswijk,
The Netherlands) using an Axiovert 35 M inverted microscope
(Zeiss, Göttingen, Germany) as described previously [28]. We
used a 40× oil immersion objective (Plan-NeoFluar NA 1.30) to
capture images. Light with an excitation wavelength of 340 and
380 nm evoked by a monochromator (TILL Photonics Polychrome
IV; TILL Photonics GmBH, Gräfelfing, Germany) was directed to
the sample via a 290 nm long-pass filter and beam splitter. From
there, the emitted light with a wavelength of 510 nm was directed
to a 440 nm long-pass filter and was collected every 0.5 s, i.e., at a
sample frequency of 2 Hz for each excitation wavelength, with an
Image SensiCam digital camera (TILL photonics GmBH). The
degree of fluorescence in Fura-2-loaded neuronal co-cultures
allowed us to sample with binning 2 × 2. TILLvisION (version
4.01) software was used to trigger the light source, camera, and
data acquisition.
For the cells to exhibit spontaneous network activity, experi-
ments must take place at 37 °C. Below this temperature no spon-
taneous calcium oscillations are visible (data not shown). In order
to keep our samples at a stable temperature, we equipped the
microscope with a custom-built heating system and bipolar tem-
perature control unit (TC-202, Medical Systems Corp, Greenvale,
New York, USA).
Changes in the ratio F340/F380 of selected regions of interest
reflecting the changes in [Ca2+
]i were further analyzed using
custom-
made MS Excel macros. Since the cells form networks,
spontaneous synchronous calcium oscillations are seen in all
regions.
To monitor calcium levels, X-Rhod-1 (Ex 580/Em 602, AM-ester
derivative, Life Technologies, Bleiswijk, The Netherlands) was
used and for membrane depolarization the dye FluoVolt (Ex 488/
Em 515, Life Technologies, Bleiswijk, The Netherlands).
Live cell microscopic fluorescence recordings were performed
on a commercial NIKON Ti system equipped with an EMCCD
camera (iXon Ultra897, ANDOR) using wide-field illumination. A
10× CFI Plan Fluor air objective (NA 0.3, WD 16 mm) was applied
to record the data. A LED light source (Lumencor Spectra X) illu-
minated the samples to deliver 470/24 nm and 575/25 nm light
bandwidths sequentially. Emitted fluorescence light was collected
through a quadband emission filter (Chroma, DAPI/FITC/
TRITC/Cy5 Quad). Triggered synchronization of the LEDs and
the camera was directed with the NIS-
Elements software ND
acquisition module (NIKON, version 4.6) via a NIDAQ commu-
nication board (National Instruments). Halogen illumination light
path was regulated by a fast shutter (Lambda SC, SUTTER
Instruments) to obtain differential interference contrast (DIC)
2.3 Live Cell [Ca2+
]i
Changes Using
a Temperature
Controlled Microscopy
Unit
2.4 Simultaneous
Live Cell Imaging
of Calcium Transients
and Membrane
Depolarization
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
34. 24
images without the delay of slow on and off glowing rates of these
lamps. Fast focus control is performed by a piezo stepping stage
device (Mad City Labs, Madison, WI, USA).
In the Spectra X module of the NIS-Elements software, the
470 nm and 575 nm light sources were controlled. Via the trig-
gered acquisition module, light levels and exposure times were set
(3%, ND4 and eight filters were available for careful illumination
modulation). The EMCCD camera capture area was cropped to
256 × 256, and 2 × 2 binning was applied at 20 ms exposures
(sampling in frame transfer mode at 17 Mhz horizontal pixel shift
readout frequency with 300 V amplification gain).
The system is equipped with temperature and CO2 control in
a stage top setup with a small water container to provide a humidi-
fied culture environment (TOKAI HIT, INUBG2EH-TiZB). The
lens heating option was omitted for the air objectives employed.
Both the automated focus control system of NIKON (perfect
focus system or pfs) and the compressor that provides air for the
vibration isolation table can create vibrations and were thus
switched off for the recordings.
Following data acquisition, recorded data was exported to MS
Excel. For further frequency data analysis, one representative trace
from a dish was chosen as all cells display synchronized oscillations
in calcium levels and membrane potential. To calculate changes in
peak amplitude, data was loaded in a custom-made MATLAB pro-
gram, and averages of multiple responsive cells in the dish were
calculated. Data are reported as mean ± SEM form n cells.
Cells were cultured on 48-well MEA plates (Fig. 1) as described
previously [12, 17]. Each well contains an electrode array of 16
nanotextured gold microelectrodes that are ~40–50 μm in diame-
ter with a 350 μm center-to-center spacing. Each well contains
four integrated ground electrodes. In total this yields 768 channels
in one plate that can be recorded simultaneously. A Maestro
768-channel amplifier with an integrated heating system, tempera-
ture controller, and data acquisition interface (Axion Biosystems
Inc., Atlanta, GA, USA) was used for recordings of neuronal activ-
ity. Data acquisition was managed with Axion’s Integrated Studio
(AxIS version 2.4.2.13) and recorded as .RAW files. Files were
obtained by sampling all channels simultaneously with a gain of
1200× and a sampling frequency of 12.5 kHz/channel using a
band-pass filter (200–5000 Hz). Notably, the large number of
electrodes, sampled at high frequency, will yield a large amount of
data (~1 GB/min recording). Consequently, sufficient storage
space should be available.
For data analysis .RAW files were re-recorded to obtain alpha
map files. During this re-recording, spikes were detected using the
AxIS spike detector (adaptive threshold crossing, Ada BandFIt v2).
A variable threshold spike detector was used with a threshold set at
2.5 Multi-well
Microelectrode Array
Recordings
Anke M. Tukker et al.
35. 25
7× standard deviation of the internal noise level (rms) on each elec-
trode. The obtained spike files were loaded in Neural Metric (ver-
sion 2.04, Axion Biosystems). In this program, bursts were detected
with the Poisson surprise method (minimal surprise S = 10 [29]).
Network bursts were extracted with the adaptive threshold method
(min # of spikes 10; min % of electrodes 25).
The Neural Metric output files (.csv files) were loaded in a
custom-
made MS Excel macro. We only used active electrodes
(MSR ≥ 6 spikes/min) from active wells (≥1 active electrode) for
further analysis. Electrodes were seen as bursting electrodes when
the minimum burst rate was ≥0.001 bursts/s. Only wells with ≥2
bursting electrodes were included for network bursting and syn-
chronicity analysis. Effects of test compounds were determined by
comparing the activity of the baseline of a well to the activity in that
well following exposure. During data analysis it is important to
Fig. 1 Experimental setup for measurements of spontaneous neuronal network activity. Axion’s Maestro plat-
form (a) was used to record neuronal activity of co-cultures grown in 48-well MEA plates (b). Each well con-
tains an electrode grid with 16 electrodes/well (c) for noninvasive extracellular field recordings. Live heat maps
are shown in AxIS software during recordings (d). Files are loaded in neural metric to determine (e) spikes
(black), bursts (blue), and network bursts (pink squares)
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
36. 26
correct for exposure artifacts [12]. To do so, the time it took to
expose all wells was excluded from data analysis. For example, if
exposing the plate took 2 min, the first 2 min of the recording was
not used for data analysis but only the subsequent 30 min. Then,
treatment ratios per well for different metric parameters (mean
spike rate [MSR], mean burst rate [MBR], and mean network burst
rate [MNBR]) were calculated by expressing the parameterexposure as
a percentage change of parameterbaseline. Next, treatment ratios were
normalized to vehicle control. Outliers were defined as not within
average ± 2× standard deviation and excluded for further analysis.
MEA data are expressed as mean ± SEM. A one-way ANOVA was
performed to determine statistical significant changes (p 0.05) in
MSR, MBR, and MNBR compared to vehicle control.
3 Methods and Results
Neuronal networks of culture model A were chemically fixed, and
specific antibodies were used to demonstrate the presence of neu-
rons and astrocytes. Neurons were identified using antibodies
against class III β-tubulin, which is found almost exclusively in
neurons. Anti-s100β is a protein specific for glial cells and was used
to identify astrocytes. Notably, many other antibodies can be used
to gain additional insight in the composition of the neuronal net-
work. For example, using antibodies against vGLUT (vesicular
glutamate transporter), vGAT (vesicular GABA transporter), or
tyrosine hydroxylase (marker for dopaminergic neurons) will pro-
vide information on the types of neurons present.
In order to stain the cultures, a staining protocol described
previously [25] was used. The cultures were chemically fixed at
DIV16 and 21 with 4% PFA in 0.1 M PBS (pH 7.4) for 15 min at
RT. We found that longer fixation times reduced epitope recogni-
tion by the antibodies. Following fixation, chambered coverslips
were quenched for PFA, permeabilized, and incubated for 20 min
at RT with 20 mM NH4Cl in blocking buffer (2% bovine serum
albumin and 0.1% saponin in PBS). Hereafter, chambers were
incubated overnight at 4 °C with the primary antibody. The fol-
lowing primary antibodies were used: mouse anti-S100β (final
dilution 1:500; Ab11178, Abcam, Cambridge, United Kingdom)
to stain astrocytes and rabbit anti-βIII tubulin (final dilution 1:250;
Ab18207, Abcam, Cambridge United Kingdom) to visualize the
iCell® Neurons and iCell® Glutaneurons. The following day,
chambers were washed thrice with blocking buffer and incubated
with donkey anti-rabbit Alexa Fluor® 488 (final dilution 1:100;
715-545-152, Life Technologies, Bleiswijk, The Netherlands) and
donkey anti-mouse Alexa Fluor® 594 (final dilution 1:100; 715-
585-151, Life Technologies, Bleiswijk, The Netherlands) for
45 min at RT in the dark. During the last 2–3 min of the incuba-
tion time, 200 nM DAPI (staining the nuclei) was added. Chambers
3.1 Immunocyto-
chemistry: Visualizing
Neuronal Networks
Anke M. Tukker et al.
37. 27
were washed again 3× with blocking buffer and sealed with two to
three droplets of FluorSave (Calbiochem, San Diego, CA, USA).
Now the chambers are ready for use. However, they can be stored
for months at 4 °C in the dark until use.
The images of the chemically fixed and stained co-cultures of
culture model A show that mixed neuronal networks with a high
degree of complexity are formed (Fig. 2a, b). It is clear that the
network is already strongly developed at DIV16 (Fig. 2a) and this
is maintained till DIV21 (Fig. 2b). The images also show that astro-
cytes do not overgrow the neuronal population. The ratio astro-
cytes to neurons remains comparable over time and is in line with
the plated ratio. Altogether, these results indicate that this co-cul-
ture consists of a mixed cell population with neurons as well as sup-
porting astrocytes that can be used for neurotoxicity testing.
Both dyes used for the [Ca2+
]i experiments described in this chap-
ter contain an acetoxymethyl (AM) ester. The AM group allows
the dye to cross the cell membrane, which allows the cells to be
loaded in a noninvasive manner. Following membrane crossing,
the AM group is cleaved from the dye by non-specific esterases in
the cytosol. The cleaved dye is no longer able to cross the cell
membrane and remains in the cytosol. The calcium-sensitive dye
Fura-2 is fluorescent green (510 nm), and fluorescence increases
upon Ca2+
binding following excitation at 340 nm, but fluores-
cence decreases by excitation at 380 nm. The resulting F340/380
ratio thus correlates directly with the [Ca2+
]i.
Cells cultured in glass-bottom dishes of culture model B were
loaded with 5 μM Fura-2AM for 1 h at 37 °C in a humidified 5%
CO2 incubator. After loading, cells were washed four times with
37 °C saline solution to remove excess dye. Then dishes with
loaded cells were placed on the stage of the inverted microscope in
the heating ring. The temperature sensor was placed in the dish,
3.2 Live Cell
Imaging: [Ca2+
]i
Changes in hiPSC-
Derived Co-cultures
Fig. 2 Immunofluorescent stainings of co-culture model A. At DIV16 (a) and DIV21 (b), cultures were stained
with β(III) tubulin (green) and S100β (red) to identify, respectively, neurons and astrocytes in the co-culture.
Nuclei were stained with DAPI (blue). Scale bar depicts 25 μm
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
38. 28
and cells were allowed to warm till 37 °C. As soon as this tempera-
ture was reached, a 5 min recording was started to measure spon-
taneous calcium oscillations.
With this method it is possible to detect spontaneous calcium
oscillations in hiPSC-derived neuronal co-cultures (Fig. 3). Sample
traces indicate the presence of spontaneous calcium oscillations.
Cells oscillate synchronously, indicating that cells are all part of a
single network.
X-rhodamine-1 (X-RhodAM) was used to study spontaneous cal-
cium oscillations. This dye works in a comparable manner as Fura-
2AM. However, the main difference is that Fura-2AM is a dual
wavelength dye and X-RhodAM a single wavelength dye in the red
color range (emission 595 nm). The latter has as advantage that it
can be imaged at a higher sample frequency and allowing it to be
used in combination with other sensor dyes. In order to study
simultaneous changes in membrane depolarization, we used the
dye FluoVolt. In general there are two types of probes that can be
used to do this: fast- or slow-response probes. The first type reacts
fast, but the magnitude of potential-dependent fluorescence
change is small. The latter type reacts slower, but the magnitude of
fluorescence fluctuation is high. FluoVolt is a dye that combines
characteristics of fast and slow probes as it reacts fast (millisecond
time scale) and yields a high magnitude of fluorescence with a small
change in membrane
potential (~25% change in fluorescence per
100 mV change in membrane potential).
3.3 Live Cell
Imaging:
Simultaneous Calcium
Oscillations
and Membrane
Depolarization
in hiPSC-Derived
Co-cultures
Fig. 3 Sample traces of representative neuronal cells (DIV23) showing spontane-
ous calcium oscillations. Each trace represents the oscillations of a single cell
over a 3 min time period. Cells oscillate synchronously, indicating they are part
of a single network
Anke M. Tukker et al.
39. 29
Cells of culture model B were cultured up till DIV21 in glass-
bottom dishes and used to record changes in calcium transients
and membrane voltage fluctuations. One day prior to measure-
ments, cells were transferred to an incubator close to the imaging
station to minimize transport stress and temperature variation at
the time of the experiment. At the day of measurements, start the
microscope peripheral modules (Spectra X, Mad City Labs stage
control, XY control joystick unit, halogen lamp, and TOKAI HIT
control unit) before powering the microscope stand and finally
turn on the PC. Distilled water was placed in the basin of the
TOKAI HIT sample holder. It is important to place a dummy sam-
ple in the holder to prevent condensation on optics below the
table. At this point, the climate control unit (37 °C for the basin,
40 °C for the TOP deck heating, 5% CO2) is switched on, and the
whole setup is left for acclimatization for 30 min. All experiments
took place at 37 °C.
During the acclimatization time, cultures were loaded with a
mixture of 5 μM X-Rhod-1, 5 μM FluoVolt, and PowerLoad solu-
tion (a solubilizing agent provided in FluoVolt kit) in life cell imag-
ing solution (LCIS; 140 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl2,
1.0 mM MgCl2, 20 mM HEPES, 20 mM glucose) for 15–20 min.
Following incubation, cells were washed four times with 1 mL
LCIS. After loading, the sample was placed in the TOKAI HIT
holder (UNIVD35) to search for a representative area and select
regions of interest (ROIs) to obtain live intensity preview plots
(Fig. 4). Next, a 10 min baseline recording was made prior to addi-
tion of test compound and a subsequent 10 min exposure record-
ing. For exposure recordings, a fresh stock solution of picrotoxin
(PTX) in EtOH was prepared on the day of the experiment. Stock
solution was further diluted in LCIS. Solvent concentration did
not exceed 0.1% v/v.
Data again confirm that neuronal co-cultures form networks,
since oscillations in membrane potential as well as calcium occur
for all cells at the same time (Fig. 5a, b left). Exposure (dilution
Fig. 4 Captures of neuronal cells loaded with FluoVolt (a), X-Rhod (b), and an overlay (c). Encircled areas are
selected regions of interest for measurements as further illustrated in Fig. 5
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
40. 30
1:10) to 10 μM (PTX) does not affect the frequency of oscillations.
However, it does result in a decrease of the amplitude of mem-
brane potential peaks to 57.2% (±3.4, n = 6) of baseline (Fig. 5a).
A stronger effect is seen on calcium oscillations, where amplitude
decreases to 37.5% (±3.2, n = 6) compared to baseline (Fig. 5b).
Basal fluorescence increases over time, likely as a result of ongoing
de-esterification of dye that still contained the AM module.
However, filter and/or subtraction methods can be used to elimi-
nate this trend in fluorescence from data analysis.
All MEA measurements took place at 37 °C with culture model
B. Experiments took place at DIV21 since this is the optimum
window in terms of activity (Table 3). It should be noted that dif-
ferent culture models develop differently and may therefore have a
different optimum window. For this reason, it is strongly recom-
mended to always make a developmental curve before starting
toxicity experiments. In order to determine effects of test com-
pounds on spontaneous network activity and (network) bursting of
the hiPSC-derived co-culture, a 30 min baseline recording was
made. Prior to the 30 min recording, plates were allowed to equili-
brate in the Maestro for ~5 min. Immediately following this 30 min
baseline recording, cells were exposed (dilution 1:10) to the test
compounds or the solvent control, and another 30 min recording
was made. Each well was exposed only once, since cumulative dosing
3.4 MEA: Assessing
Spontaneous Neuronal
Network Activity
in hiPSC-Derived
Co-cultures
Fig. 5 Selection of traces of cells selected in Fig. 4 for assessing changes in fluorescence of FluoVolt (a) and
X-Rhod-1 (b) during baseline (left) and exposure (right) to 10 μM PTX. Different cells oscillate simultaneously,
indicating that changes in membrane potential (a) and [Ca2+
]i (b) occur at the same time and the cells are part
of the same neuronal network
Anke M. Tukker et al.
41. 31
may confound results due to, e.g., receptor (de)sensitization. We
prepared fresh PTX stock solutions in EtOH and of strychnine in
supplemented BrainPhys™ medium prior to every experiment.
Stock solutions of PTX were further diluted in supplemented
BrainPhys™ medium such that solvent concentration never
exceeded 0.1% v/v.
Neuronal activity in culture model B can be modulated with
strychnine (Fig. 6a). Exposure to the highest tested concentration
strongly decreases MSR, MBR, and MNBR (Fig. 6a, right) as
compared to baseline activity (Fig. 6a, left). The inhibitory effect
of strychnine on MSR is concentration-dependent (Fig. 6b). This
in contrast to the MBR, which increases following exposure to low
concentrations of strychnine and decreases following exposure to
higher concentrations. MNBR decreases by all tested concentra-
tions of strychnine. Exposure to PTX (Fig. 6b) has little effect on
the MSR, except for exposure to 1 μM, which decreases the
MSR. The lowest test concentration of PTX causes a decrease in
MBR; however with increasing concentrations, the MBR increases
as well. On the other hand, all tested concentrations of PTX cause
a decrease in MNBR.
The different effects that strychnine and PTX have on culture
model B become clear from the inclusion of additional metric
parameters illustrated by a heat map (Fig. 7). This heat map also
indicates that for proper MEA data analysis, it is important to
include more parameters than just MSR, MBR, and MNBR.
4 Conclusions
Human iPSC-derived co-cultures develop functional and sponta-
neously active neuronal networks consisting of mature neurons [8,
30]. Our immunocytochemistry data demonstrate the mixed
nature of our co-culture models consisting of neurons and astro-
cytes that form complex, multicellular networks (Fig. 2).
Table 3
Development of spontaneous neuronal activity and (network) bursting at different DIVs. Data are
expressed as mean ± SEM
MSR MBR MNBR
Frequency (Hz)
% active
wells Frequency (Hz)
% active
wells Frequency (Hz)
% active
wells
DIV7 1.53 ± 0.14 93.8 0.04 ± 0.01 26.7 0.01 ± 0.01 25
DIV14 1.43 ± 0.14 93.8 0.02 ± 0 13.3 0.01 ± 0.01 100
DIV21 1.29 ± 0.18 93.8 0.05 ± 0.01 76.7 0.03 ± 0.03 65.2
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
42. Fig. 6 Toxicological modulation of spontaneous network activity, bursting, and network bursting. Cultures were
exposed at DIV21 to strychnine (a, b) or PTX (c). Spike raster plot depicting activity and (network) bursting
before (left) and following exposure (right) to 10 μM strychnine (a). Each row depicts one electrode, each tick
representing one spike (field potential) in a 200 s interval, bursts are depicted in blue, and network bursts are
encircled in pink squares. Data are expressed as MSR, MBR, or MNBR as % change relative to vehicle control;
mean ± SEM from n = 1–12; *p 0.05 (b, c)
Fig. 7 Heat map of the effects of strychnine and PTX (concentration in μM) on selected metric parameters on
culture model B. Color scaling is based on the magnitude of the % of change relative to the vehicle control. No
average could be calculated for white cells
43. 33
We have shown that hiPSC-derived neuronal co-cultures are
amenable to multiple real-time recording techniques, including
live cell imaging and electrophysiology. Our calcium imaging data
indicate that the co-culture models develop spontaneous calcium
oscillations and spontaneous changes in membrane potential
(Figs. 3 and 5). Since calcium oscillations and changes in mem-
brane potential occur in multiple cells at the same time (Figs. 3 and
5), it can be concluded that functional networks are formed.
Moreover, MEA recordings demonstrate that neuronal co-cultures
develop spontaneous network activity and (network) bursting
(Table 3 and Fig. 6).
Spontaneously active human iPSC-derived neuronal co-cul-
tures are suitable for (preliminary) neurotoxicity assessment [18,
25, 27, 31, 32] as is confirmed by our data (Figs. 5, 6, and 7).
However, it must be noted that model composition, e.g., the ratio
of GABAergic and glutamatergic neurons and the presence of astro-
cytes, greatly influences the model’s characteristics [25]. Therefore,
a careful model characterization must be performed prior to toxic-
ity testing.
The increasing availability of hiPSC from different donors
and/or patients differentiated in different cell types, e.g.,
GABAergic, glutamatergic, dopaminergic neurons and astrocytes
as well as peripheral neurons, holds great promise for future per-
sonalized toxicity and safety screening. Using hiPSC-derived neu-
rons in combination with the techniques described here will
provide a good starting point for neurotoxicity assessment.
Acknowledgments
We gratefully acknowledge members of the Neurotoxicology
Research Group for helpful discussions. This work was funded by
a grant from the National Centre for the Replacement, Refinement
and Reduction of Animals in Research (NC3Rs; project number
50308-372160), the Netherlands Organisation for Health
Research and Development (ZonMW; InnoSysTox project num-
ber 114027001), and the Faculty of Veterinary Medicine (Utrecht
University, The Netherlands).
References
1. Clapham DE (2007) Calcium signaling. Cell
131:1047–1058. https://doi.org/10.1016/J.
CELL.2007.11.028
2. Westerink R (2006) Targeting exocytosis: ins and
outs of the modulation of quantal dopamine
release. CNS Neurol Disord Drug Targets 5:
57–77. https://doi.org/10.2174/18715270
6784111597
3. Barclay JW, Morgan A, Burgoyne RD (2005)
Calcium-dependent regulation of exocytosis.
Cell Calcium 38:343–353
4. Südhof TC (2014) The molecular machinery
of neurotransmitter release (nobel lecture).
Angew Chem Int Ed 53:12696–12717.
https://doi.org/10.1002/anie.201406359
In Vitro Techniques for Assessing Neurotoxicity Using Human iPSC-Derived Neuronal…
