Determination of Acute Lethal and Chronic Lethal Thresholds of … · bioreactors (MC2 Biotek, (Hørsholm, Denmark) Cat. no. 010). These bioreactors are specially constructed to be

University of Southern Denmark

Determination of Acute Lethal and Chronic Lethal Thresholds of Valproic Acid using 3DSpheroids Constructed from the Immortal Human Hepatocyte Cell Line Hepg2/C3A

Fey, S. J.; Wrzesinski, Krzysztof

Published in:Valproic Acid

Publication date:2013

Document versionAccepted manuscript

Citation for pulished version (APA):Fey, S. J., & Wrzesinski, K. (2013). Determination of Acute Lethal and Chronic Lethal Thresholds of ValproicAcid using 3D Spheroids Constructed from the Immortal Human Hepatocyte Cell Line Hepg2/C3A. In A. Boucher(Ed.), Valproic Acid (pp. 141-165). Nova Science Publishers.https://www.novapublishers.com/catalog/product_info.php?products_id=42166

Terms of useThis work is brought to you by the University of Southern Denmark through the SDU Research Portal.Unless otherwise specified it has been shared according to the terms for self-archiving.If no other license is stated, these terms apply:

• You may download this work for personal use only. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying this open access versionIf you believe that this document breaches copyright please contact us providing details and we will investigate your claim.Please direct all enquiries to [email protected]

Download date: 21. Jan. 2021

https://www.novapublishers.com/catalog/product_info.php?products_id=42166

In: Valproic Acid ISBN: 978-1-62417-952-5

Editor: Alexis Boucher © 2013 Nova Science Publishers, Inc.

Chapter V

Determination of Acute Lethal

and Chronic Lethal Dose

Thresholds of Valproic Acid

Using 3D Spheroids Constructed

from the Immortal Human

Hepatocyte Cell Line HepG2/C3A

Stephen J. Fey,† and Krzysztof Wrzesinski†

Department of Biochemistry and Molecular Biology,

University of Southern Denmark, Denmark

Abstract

Valproic acid (VPA) is a broad-spectrum antiepileptic drug that is

now used commonly for several other neurological and psychiatric

indications. While VPA is usually well tolerated, on rare occasions, it has

To whom correspondence should be addressed at Department of Biochemistry and Molecular

Biology, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark.

Tel: +45 6550 2440; Fax: +45 6550 2467; Email: [email protected]. † These authors contributed equally to the work presented here.

No part of this digital document may be reproduced, stored in a retrieval system or transmitted commercially in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.

Stephen J. Fey and Krzysztof Wrzesinski 142

been associated with severe, and sometimes, fatal liver injuries. These

complications may also arise due to acute VPA overdose.

As a branched chain carboxylic acid, VPA is readily metabolized in

the liver via glucuronic acid conjugation, mitochondrial β- and cytosolic

ω-oxidation to produce multiple metabolites, and it is probably some of

these metabolites that are involved in its toxicity. While the actual

mechanism of VPA hepatotoxicity is still not understood, two pathways

have been in focus: the first involves the formation of reactive

metabolites of VPA (and their subsequent covalent binding to cellular

proteins) and the second is the development of oxidative stress in the cell.

We describe here a culture system based on 3D spheroid culture of

immortal hepatocytes which can determine the toxicity of valproic acid

(or structurally or functionally related molecules) in vitro. The spheroids

were used to follow changes in ATP production, glucose uptake and

adenylate kinase following treatment and can be treated repeatedly with

VPA to determine both acutely-lethal and chronically-lethal doses. This

system is homeostatic and has been shown to be metabolically stable over

at least 24 days and can therefore be used to follow recovery after

treatment.

Keywords: 3D Spheroid culture; drug toxicity; valproic acid; acute lethal

threshold; chronic lethal threshold

Introduction

When cells are grown in three dimensional (3D) environments (i.e. as

multilayers or clusters in suspension, on scaffolds etc.) they have been shown

to possess a number of physiological characteristics which makes them

resemble more closely the native tissue from which they originated than the

same cells grown in classical two dimensional (2D – i.e. as monolayers)

conditions in culture flasks or microtitre plates.

This may be due to the fact that cells grown in 3D cell culture are allowed

to develop a more elaborate extracellular matrix and better intercellular

communication [1], [2], [3], and this leads to a recovery or maintenance of in

vivo function [4].

For this reason, we recently investigated how long it takes cells to recover

from trypsinisation. Our studies revealed that the hepatocyte cell line HepG2-

C3A went through some form of transition after 18 days, after which several

key metabolic traits (growth rate, ATP levels, urea and cholesterol production)

all stabilised[5]. A review of the literature showed that this was not a unique

Determination of Acute Lethal and Chronic Lethal Dose … 143

phenomenon and there were several well-known cell lines (including the cell

lines Caco-2, HT 29, MDCK, MCF-10A and HepG2 used to model the small

and large intestine, kidney, breast acini and liver respectively) which all

needed similar periods of time to fully establish themselves. We therefore

suggested that these changes represent a ubiquitous recovery process after

trypsinisation rather than differentiation as is often stated. This would help to

explain why three dimensional structure has been proposed as the missing link

which will provide in vitro models that sufficiently mimic in vivo conditions

[6], [7] and [8].

In this study we have exploited the potential offered by 3D spheroid

culture to investigate drug treatment over extended periods of time. After the

18 days needed to recover from trypsin, 3D spheroid cultures are essentially

homeostatically stable from 18 days to at least 42 days of culture [9]. They

grow slowly, have a high survival percentage; their physiological capabilities

(glycogenesis, ATP, cholesterol and urea synthesis and drug metabolism) are

constant; and they have a stable gene expression of several key genes related

to the main liver pathways. Furthermore, they appear to more predicitive than

primary hepatocytes or similar cells when grown in classical 2D cultures of the

in vivo toxicity in man, at least for a small number of drugs [10]. All in all, this

makes 3D spheroids ideal for long term studies – for example to investigate

drug induced liver injury (DILI).

Valproic acid (VPA) has become one of the most widely prescribed

anticonvulsant and mood-stabilizing drug worldwide. Despite the

pharmacological importance and the effectiveness of VPA, its potential

hepatotoxicity is still a major concern. The recommended VPA starting dose is

between 750 and 1500 mg daily taken in divided doses and then dosages are

adjusted based on response and blood level. The usual effective dose range is

between 1000 and 2000 mg daily.

VPA may have its beneficial neurological effects by increasing the

concentration of the inhibitory neurotransmitter Gamma Amino Butyrate

(GABA). VPA acts as a double edged sword causing partial creatine

deficiency in the brain by stimulating the GABA-forming enzyme glutamate

decarboxylase (GAD65/67); and inhibiting 2-oxoglutarate dehydrogenase (a

rate-limiting mitochondrial Kreb’s cycle enzyme) and the GABA-catabolizing

enzymes GABA transaminase (GABA T) and succinic semialdehyde

dehydrogenase (SSADH)[11][12]. This becomes toxic if GABA levels are

raised too far.

However, treatment with VPA can also lead to liver toxicity, or more

specifically drug-induced liver injury (DILI). This is an unresolved problem


which has significant clinical and economic impacts [13][14]. While the exact

mechanisms of DILI remain largely unknown [15] there are several areas

under investigation [16]. In the liver, the disturbance of mitochondrial function

(including the inhibition of mitochondrial β-oxidation) leads to

microvesicularfatty liver [17] and the impairment of oxidative phosphorylation

and reduction of ATP synthesis [18][19][20].

Finally, valproic acid can also cause epigenetic effects due to a histone

deacetylase inhibitor activity (which would then stimulate gene expression in a

more or less specific manner) [21].

Using the two key advantages of hepatic spheroids, namely their long term

stability and the fact that they are homeostatic and metabolically ‘in

equilibrium’ we describe here the effects of VPA treatment and define two

different lethal treatment thresholds, (one for a single treatment (acute) and

one for multiple treatments (chronic)) by measuring basic physiological

parameters of glucose uptake, cellular ATP levels and the release of adenylate

kinase. The results suggest that cell death by single high doses of VPA (acute)

are different to cell death caused by multiple low doses of VPA (chronic).

Materials and Methods

Standard 2D Cell Culture Conditions

The immortal human hepatocyte cell line, HepG2/C3A (ATCC CRL-

10741, third passage after receipt from ATCC, Manassas, Virginia.), were

thawed from liquid nitrogen storage and cultured in standard tissue culture

conditions in a customised D-MEM growth medium (87.5% D-MEM (1 g

Glucose/L) (Gibco (Carlsbad, California) Cat. no. 31885-023); 1% Non

Essential Amino Acids (Gibco Cat. no. 11140-035); 10% FCS (Foetal calf

serum) (Sigma (St. Louis) Cat. no. F 7524); 0.5% Penicillin / Streptomycin

(Gibco Cat. no. 15140-122) 1% GlutaMAX (Gibco Cat. no. 35050-038), 37⁰C,

5% CO2 95% air).

In order to generate cells for spheroid construction, the C3A cells were

initially propagated using traditional 2D tissue culture. Cells were trypsinised

(0.05% Trypsin / EDTA Cat. no.: Gibco 15400-054) for 3 minutes and sown

out into falcon flasks/microtitre plates and cultivated at 37⁰C, 5% CO2 95% air

in a humidified incubator.


To eliminate any effects of storage in liquid nitrogen or thawing, cells

were grown for at least three passages before starting the experiments,

exchanging the medium every two to three days. Cells were used between

passages 4 and 15. The doubling time for C3A cells grown under these

conditions was 76 hours.

3D Spheroid Culture Conditions

Preparation of Spheroids Using AggreWell™Plates

The C3A cell spheroids have been prepared with use of AggreWellTM

400

plates (Stemcel Technologies (Grenoble, France) Cat. no. 27845). Each

AggreWell™ plate has 6 wells and each of these contains approximately 4,700

microwells measuring 400 µm in diameter. These plates are used to form cell

aggregates. Before use, the plates were washed twice with growth medium

(customised D-MEM). In order to remove all residual air bubbles from the

well surface the plates were prefilled with 0.5 ml of growth medium and

centrifuged (3 min with 3,000 x g). Cells (1.2 x 106) were added to each well

and the plates were centrifuged (3 min with 100 x g) and left in the

AggreWellTM

plate overnight to form spheroids.

Spheroid Culture in ProtoTissue™ Bioreactors

The spheroids were detached from the AggreWellTM

plates by gently

washing the wells with prewarmed growth medium. The detached spheroids

were collected in to a Petri dish and the quality of the spheroids checked by

microscopy. Compact spheroids were introduced into ProtoTissue™

bioreactors (MC2 Biotek, (Hørsholm, Denmark) Cat. no. 010). These

bioreactors are specially constructed to be easy to open and close and designed

to maintain 100% humidity around the growth chamber. The spheroids

(approximately 300 per bioreactor) were then cultivated at 37⁰C, 5% CO2 95%

air in a humidified incubator for a minimum of 21 days, exchanging the

medium every two to three days [22].

The day when the cells were transferred into the bioreactor is defined as

day 0. An estimated 90% of the medium was changed on day 1 and thereafter

three times a week during the recoveryperiod. To achieve a stable suspension

of spheroids, the rotation speed of the bioreactors was initially set between 21-

23 rpm and it was adjusted to compensate for the growth in size of the


spheroids (and reached 27-30 rpm on day 21). Speed adjustments were made

so that inter-spheroid contact was minimised and these adjustments were made

more frequently at the beginning of the culture than at the end. Because the

spheroids get larger with time, the population density of the spheroids was

regulated by opening the bioreactor and ‘splitting the population’ byremoving

excess spheroids (this usually needed to be performed between day 11 and 14).

Before use, the spheroid batch quality was assessed by staining for 3 min. with

0.4% Trypan Blue (Gibco Cat. no. 15250-061). Batches showing greater than

90% viability were accepted.

Microscopy and Planimetry

Photomicrographs were taken using an Olympus IX81 motorised

microscope, and an Olympus DP71 camera. Images were transferred to the

Olympus AnalySiS®Docu program (Soft Imaging System) and the ‘shadow’

area of spheroids measured using the ‘Fitted Polygon Area’ function which

calculates the planar surface of the spheroids in µm².

Drug Treatment

Valproic acid (VPA) was purchased from Sigma (Cat no. P4543). Stock

solutions were prepared just before use by dissolving the compound in either

growth medium or DMSO (Sigma Cat. no. D2650) and diluted into growth

media. A control media was prepared which contained the same concentration

of DMSO vehicle without any compound. The maximum final DMSO

concentration was 0.01%.

Three hundred 21-day old spheroids (corresponding to approximately 25

million cells or 3.5 mg protein) were gently pipetted (using a cut-tip) into each

10ml bioreactor before treatment to provide ample material for analysis. To

initiate drug treatment, the bioreactor rotation was stopped for 30 s to allow

the spheroids to settle to the bottom of the bioreactor. Ninety percent of the

media volume was exchanged with media containing the compound at

concentrations which gave the final treatment concentrations given in the

figures. Experiments to determine the LD50 were carried out initially using a

broad range of drug concentrations (e.g. varying by a factor of ten) followed

by two experiments using a narrow concentration range (varying typically by a

factor of four).


The data from the broad range of VPA concentrations were used to

identify the final narrow range used, and so these data are not presented. Each

narrow range experiment was carried out in duplicate and ATP was measured

in technical triplicates for each sample. Media was exchanged at 48, 96, 168,

216 and 264 h after the start of the experiment.

ATP Assay

Cell viability was measured based on their ability to produce ATP [23].

Samples of the hepatocytes grown as spheroids were collected (usually 2 - 6

spheroids per assay point depending on the size of the spheroids) at

appropriate times and transferred to white opaque microtitre plates (Nunc,

(Roskilde, Denmark) Cat. no. 165306), the volume of growth medium was

adjusted to 100 µl.

The cells were then lysed with 100 µl of lysis buffer (CellTiter-Glo

luminescent cell viability assay, Promega (Fitchburg, WI) Cat. no. G7571) and

shaken in the dark for 20 min. before the luminescence was measured in a

FluoStar OPTIMA (BMG Labtech, Ortenberg, Germany) using the following

parameters: one kinetic window, 10 measurement cycles with 0.3 s of

measurements interval time, 2 s delay per measurement, additional 0.5 s delay

per position change, (repeated twice for each measured plate). Replicates were

performed as described under ‘Drug Treatment’.

The data was normalised with reference to a standard curve for ATP and

to the untreated control.

Adenylate Kinase Assay for Cell Viability

Adenylate kinase was measured using the bioluminescent cytotoxicity

assay kit (LonzaToxilight assay kit, cat. no. LT07-117). Spent media from

spheroid cultures were collected in triplicate in connection with refreshing the

growth media (typically at 48 h intervals), diluted with five volumes of

adenlyate kinase detection reagent, allowed to incubate for five minutes and

then read the bioluminescence in a Fluostar Omega® (BMG Labtech) using 1

second measurement interval time per each measured well. A standard curve

was prepared in the same way for each assay plate using a cell standard (4.29

million C3A cells per mL lysed in lysis buffer). Gain was adjusted based on

the highest standard curve values.


Protein Determination

In order to correlate the shadow area of the spheroid to its protein content,

the latter was determined using the fluorescence based ProStain Protein

Qantification Kit (Active Motif, Inc (La Hulpe, Belgium) Cat. no. 15001)

according to the manufacturer’s instructions. All data was expressed as a

function of the amount of protein present.

Glucose Determination

Samples for glucose determination were collected from every culture at: 0,

4, 24, 48, 52, 72, 96, 100, 168, 174, 192, 216, 222, 240, 264 and 270 h and

frozen until needed. Glucose was measured using a ‘Onetouch Vita glucose

meter’ and test strips, (MediqDanmark, Cat No. 6407078 Cat. No. 6407079

respectively).

The instrument was calibrated using a 1 mg/ml D-glucose standard

solution, (Sigma-Aldrich Cat. No. G3285-5ML) and Onetouch Vita control

solution (MediqDanmark Cat.No. 64-07-081). Standards and samples were

warmed to room temperature and shaken on a rotary shaker at 450rpm for 1 h.

For each measurement a test strip was inserted into the instrument and 2-3 µl

of each sample were spotted onto it. The glucose concentration is read

immediately.

Results

This study was set up to investigate the effects of repeatedly treating

hepatocyte spheroids grown in rotating bioreactors with various doses of VPA,

to mimic the repeated treatment of patients in vivo and thus to investigate the

molecular changes occurring during the development of drug-induced liver

injury (DILI).

When spheroids are formed using Aggrewell™ plates and grown in the

bioreactors (figure 1) using the procedures described here, they are quite

uniform in size (± 21% [10]). A few spheroids are shown in figure 2 to

illustrate this (a lower power image of many spheroids can be seen in [10]).

They can be expected to respond more uniformly than spheroids of different

sizes (for example in large spheroids the outer cells might metabolise the drug


and thus inner cells would be shielded from the full drug concentration,

whereas cells in a small spheroid will all receive the full dose).

Figure 1. A bioreactor used in this study. Spheroids are cultivated in the 10ml chamber in

front of the white membrane and the air in the chamber behind the membrane is humidified.

The spheroid cultivation chamber can be opened to give a petri-dish like access to the

spheroids. The two round plugs measure 1.5 cm and give access to either the growth

chamber or to refill the humidification reservoir. The ‘tall’ white plug is removed when

changing the media.

Figure 2. 21 day-old spheroids. Photograph of a small number of 21 day-old spheroids to

illustrate their uniformity. Note that there is very little cellular contamination in the media.

Scale bar = 1mm.



Figure 3. Relationship between spheroid shadow area, protein content and number of cells.

The spheroid shadow area was measured in four separate experiments and the data

combined, averaged and smoothed. A) Averaged spheroid shadow area (on average 180

spheroids were measured per data point). B) the averaged protein content per mm2 of the

shadow area of the spheroid. Error bars in A) and B) represent the standard deviation of the

unsmoothed data. C) Calculated protein content of a spheroid depending on it’s age; D)

calculated number of cells in a spheroid depending on it’s age. The data have been

smoothed so that the graphs can be used as look up tables for the actual experiments.

The relationship between the size of a spheroid (it’s shadow area), their

protein content and the calculated number of cells contains is shown in figure

3 for spheroids from their day of formation (day 0) to day 42.

In order to determine the acute and chronic lethal concentrations at which

valproic acid kills human C3A spheroids, twenty one day old spheroids were

treated with different concentrations of valproic acid. The spheroids were

repeatedly treated with VPA by exchanging ninety percent of the growth

media (containing the appropriate VPA concentration) after 48, 96, 168, 216

and 264 h (2, 4, 7, 9 and 11 days). Only ninety percent of the media was

exchanged to reduce the risk that the spheroids agglomerated.

Measurements of cellular ATP levels have been suggested to be one of the

best indicators of cellular viability [23]. A small number of spheroids (so as

not to substantially affect the total number in the bioreactor) were collected

from each culture during each media refreshment, in order to determine ATP

levels. The actual number collected depended on the size of the spheroids and

was typically between six and two depending on the age of the culture. All

data have been normalised with respect to the amount of ATP present in the

control, mock treated culture at the same age.


The results (figure 4) show that a dose of 20 milligram VPA per milligram

total cellular protein (20 mg VPA/mg TCP) was sufficient to reduce cellular

ATP levels to lower than 30% after 24 hours and lower than 11% after 48

hours. The second treatment killed the remaining few cells.

Treatment of the spheroids with the first dose of 4 mg VPA/mg TCP

resulted in a fall of ATP levels to about 75% of the control levels (figure 4).

After the second, fourth and fifth doses cells appeared to be able to recover

their ATP levels to normal. However there was a trend that, at each subsequent

dose, the ATP levels are supressed lower and lower and by extrapolation, one

could anticipate that if treatment was repeated a few times more that the cells

would die (after the sixth treatment the ATP levels were down to that seen

with one treatment of 20 mg VPA/mg TCP.

Lower doses (0.8 and 0.16 mg VPA/mg TCP) perturbed the ATP levels

but each time the ATP levels recovered to essentially the same level as seen in

the untreated controls. ATP levels above that seen in the controls were

frequently measured. In order to investigate the system further, we

investigated glucose consumption of the cultures because VPA has been

shown to have a profound effect on glucose metabolism [24]. Fresh growth

media contains 1 mg/mL glucose (ca. 5.5 mM). The glucose levels of the

cultures were measured each time the media was changed (at 48, 96, 168, 216

and 264 h) and again 4 or 6 h after the media change. Thus glucose levels were

measured at 0, 4, 24, 48, 52, 72, 96, 100, 168, 174, 192, 216, 222, 240, 264

and 270 h for every culture. Figure 5 reflects the actual amount of glucose in

the growth media, and shows for example typically that most of the glucose in

the media had be consumed 24 h after its refreshment even at moderate VPA

doses. Interpreting the data becomes easier if one views selected

measurements. Figure 6A shows only the glucose measurements made 4 or 6 h

after refreshing the growth media and figure 6B shows the glucose

concentrations in the spent media before refreshment. These selected data

show that there is a clear correlation between the VPA dose and ability to

remove glucose from the media.

At high VPA doses the spheroids rapidly lose the ability to import glucose

while spheroids treated with low doses and the mock-treated control became

increasing proficient at doing so. figure 3Exactly the same interpretation can

be reached by viewing the glucose concentrations in the spent media (figure

6B). Spheroid cultures treated with the highest VPA (20 mg VPA/mg TCP)

managed to remove about 70% of the glucose from the media in the first 48

hrs (long dashes in figure 6B). Thereafter they showed no ability to remove

glucose and were essentially dead. Again spheroids treated with the 4 mg


VPA/mg TCP dose initially can cope with the treatment but at later stages

(after 168 hrs) have almost lost the ability to import glucose (medium length

dashed line in figure 6B).

Figure 4. ATP content of cells in VPA treated spheroids. The spheroids were treated with 0

(____ solid line), 0.16 (……. Dotted line), 0.8 (--- short dash), 4 (- - - medium dash) and 20 (__

__ __ long dash) mg VPA/mg total cellular protein (TCP) at the timepoints indicated by the

hollow arrows on the abscissa. The data has then been normalised to the mock-treated

control at the same timepoint. Error bars (only positive bar shown for clarity) show the

standard deviation calculated from the average of triplicate technical replicates for two

independent biological replicates.

The final question we investigated was whether the changes in cell

reactivity was related to apoptosis or necrosis [25][26]. In apoptosis, cell death

is an ATP-driven, carefully programmed process whereas in necrosis, cell

death is a result of a cellular trauma (e.g. infection or poisoning) that the cell

cannot cope with. Necrosis is an ATP independent event which results in the

massive release of cellular contents, including the enzyme adenylate kinase

and this results in inflammation of the surrounding cells). Measurements of

adenylate kinase release reveal an interesting aspect to what is happening in

the cultures (figure 7). Treatment with the high dose of VPA (20 mg VPA/mg

TCP) caused a clear release of adenylate kinase, peaking at 48hrs.

Thereafter the amount falls as there are fewer living cells left which can

release the enzyme. No other VPA dose gave rise to a similar reaction pattern.


When compared to the mock treated control, all three of the other VPA doses

appear to repress adenylate kinase release in a dose dependent manner,

especially at later timepoints.

Since low ATP levels do not lead to mitochondrial membrane

depolarisation, the spheroid cells are probably dying via necrosis. This is

consistent with breach of the plasma membrane and release of adenylate

kinase at the high doses of VPA (20 mg VPA/mg TCP) [29].

However, the moderate doses of VPA (i.e. 4 mg VPA/mg TCP) also lead

towards a slow ATP depletion, and since apoptosis is ATP dependent then

these moderate doses would also be expected to be driving the hepatocytes

also towards an ATP-independent necrotic death despite the absence of

adenylate kinase release. Thus it appears that there could be differences in the

necrotic pathways. This needs to be explored further, for example by

measuring cytochrome c release, glutathione, Bcl-2 and caspase 3 activation

during these processes [30][31] and [32].

Figure 5. Glucose clearance from the media. The amount of glucose remaining in the

growth media was measured during the VPA treatment. All other details as in figure 4,

except that the data were not normalised to the control.


Figure 6. Glucose clearance from the media. Subsets of data shown in figure 5. Glucose

levels remaining in the growth media A) four or six hours after refreshing the media and B)

in the spent media immediately before refreshment.

Apoptosis is an energy dependent process and is usually associated with

mitochondrial membrane depolarisation and release of cytochrome C [27]and

[28].


Discussion

The goal of this study was to investigate the suitability of 3D spheroids for

long term drug treatment using VPA as the test compound, with the long term

aim of using the system to investigate drug-induced liver injury. Our approach

was to treat mature 3D spheroids with different concentrations of VPA

repeatedly at two or three day intervals. Cell viability and cytotoxicity was

then measured by following the total cellular ATP content, the glucose

concentration of the media and the release of adenylate kinase.

Figure 7. Adenylate kinase levels in the media. The amount of adenylate kinase released

into the growth media during the VPA treatment. All other details as in figure 4.

The data presented here showed that spheroids treated with high levels of

VPA (20 mg VPA/mg TCP) rapidly lost their ability to synthesise ATP,

import glucose and they released large amounts of adenylate kinase. Moderate

levels of VPA (4 mg VPA/mg TCP) caused a slow loss in the ability to

synthesise ATP (over a 10 day period) and a corresponding slow loss in the

ability to import glucose. Low levels of VPA (0.8 and 0.16 mg VPA/mg TCP)

induced an initial increase in ATP levels, despite a lowered glucose

consumption. Both moderate and low levels of VPA treatment reduced the

amount of adenylate kinase released to below control levels. This is

presumably because the number of cells in the spheroids is increasing. From

this data it is clear that spheroids treated with the 4 mg VPA/mg TCP dose


initially can cope with the treatment but at later stages (after 168 h) are losing

the ability to import glucose.

Molecular Basis for VPA Toxicity

Measurement of glucose levels in the 3D cultures were quite revealing

because they showed that mock treated cultures cleared the glucose from the

media very quickly, probably in about 12 hours and certainly by 24 h.

Presumably, the cells are importing the glucose and synthesising glycogen

(glycogen granules have been described in 3D spheroid cultured cells [5]). At

high oxygen and glucose levels the C3A cells are probably functioning

‘inefficiently’ and exhibit the Crabtree effect and reduce Kreb’s cycle activity

[33][34]. When the glucose levels in the media fall, the cells will activate the

glycolytic pathway and the theKreb’s cycle so that they can hydrolyse

glycogen and metabolise glucose efficiently in order to ‘survive’ until the next

media change.

It has been shown that VPA causes the inhibition of glycogen synthase

kinase-3 (GSK-3) [25], [35] (in a manner analogous to lithium, another mood

controlling compound [36]). This would be expected to lead to an increase in

glycogen synthesis which in turn would to cause an increase in glucose uptake

from the media. This effect was never seen in the studies presented here even

at the lowest VPA doses investigated and therefore if VPA is acting through

GSK-3 must be occurring via one of GSK-3’s other roles. Another possibility

is via a signalling pathway [36].

GSK-3 is a part of the canonical β-catenin/Wnt pathway, which signals the

cell to divide and proliferate [37] and thererfore inhibition of GSK-3 would be

expected to slow these cellular activities. However, in spheroids, cells are

growing very slowly and therefore GSK-3’s inhibition would be expected also

to have a minimal effect and so once again, VPA treatment is probably not

operating via this pathway.

A more probable explanation is that VPA is a simple fatty acid and as

such can be a substrate for the fatty acid b-oxidation pathway. This occurs

primarily in the mitochondria and is known to result in a cumulative inhibitory

effect on long-chain fatty acid b-oxidation [20] [38][39]. Simultaneous

inhibition of the mitochondrial pyruvate uptake will impair the pyruvate-

driven oxidative phosphorylation and together with the inhibition of 2-

oxoglutarate dehydrogenase reduce ATP synthesis [11], [12], [18], [19], [20]

and [40]. This is potentially the primal cause of both the observed reduction of


glucose uptake and the fall in cellular ATP levels (at moderate or high VPA

doses).

Acute and Chronic Lethal Thresholds

The treatment of spheroids revealed that there were two different ‘lethal

thresholds’. At high VPA doses (20 mg VPA/mg TCP), the cells were killed

essentially immediately by one dose and we defined this as the ‘acute lethal

threshold’. Above this dose a single dose is sufficient to kill the cells whereas

below this dose, the single dose is not sufficient to kill.

A VPA dose five times lower than the acute lethal dose resulted in a fall

of the ATP level to 75% of the control. At each subsequent dose, the ATP

levels were supressed lower and lower and the cells would be expected to die

if treatment was repeated a few times more. Exactly analogous results were

obtained measuring glucose clearance from the media: the spheroids were

increasingly inept following each successive treatment. Thus one can infer that

the chronic lethal threshold dose will be about 4 mg VPA/mg TCP. Based on

the ATP data, above this dose (but below 20 mg VPA/mg TCP) it would be

anticipated that cultures will die if repeatedly treated with VPA, while below

this dose they will not die.

Lower doses (0.8 and 0.16 mg VPA/mg TCP) perturbed the ATP levels

but each time the ATP levels recovered to essentially the same level as seen in

the untreated controls. ATP levels above that seen in the controls were

frequently measured: this is interpreted to indicate additional cellular activity

to repair VPA-induced cellular damage (e.g. the replacement of VPA

metabolite-cellular protein covalent adducts [41] or the repair of other damage

done to the cell and its membranes [42]. Extrapolation of these data would

suggest that these lower doses would be unable to kill the C3A spheroids

irrespective of how often the cells were treated.

From the data that we obtained, it appears that the acute lethal dose and a

chronic lethal dose kill the cells by different mechanisms: in the acute

situation, there is a sudden release of adenylate kinase; whereas in the chronic

situation (at a VPA dose only 5 times lower), there is an apparent ‘repression’

of adenylate kinase release. The explanation of this is not clear. One possible

explanation is that if cells are releasing adenlyate kinase immediately after

chronic VPA treatment, then because the enzyme has a short half-life in

solution (ca. 6 h) it cannot be detected when the assay is carried out 48hrs after

treatment. Another explanation could be that the VPA treatment is merely


paralysing cellular activity, including adenylate kinase release, and not

resulting in cell death.

The lower than control adenylate kinase levels (‘repression’) could then be

a reflection of the fact that the control cells are growing whereas the treated

cells are presumably not.

Modelling the Treatment Regimen

Currently, the FDA official recommendation for the initiation of VPA

treatment using sodium valproate injection is to start therapy at 10 to 15

mg/kg/day. The dosage should be increased by 5 to 10 mg/kg/week to achieve

optimal clinical response [43]. Since the liver is almost the exclusive site for

VPA metabolism, it could be expected that the apparent chronic ‘cellular

paralysis’ (failing glucose import and falling ATP levels) induced by low

levels of VPA treatment, might slow the metabolism and removal of VPA.

Subsequent doses would then result in a positive feedback situation increasing

the ‘effective dose’ because the liver’s capacity to respond is successively

impaired. This situation therefore underlines the importance of slowly

increasing the dose given to the patient.

Dose Frequency

In the model described here we have used a single constant dose given

each 48 or 72hrs. Since the half-life of VPA is 10 – 20 h [44] one could argue

that the VPA is given too infrequently. However, in vitro the VPA metabolites

are not removed as they are in vivo by the kidneys and therefore the biological

half life in vitro would be expected to be longer. 48hr treatment intervals were

used to minimise disturbances to the spheroids in the bioreactor cultures. If

VPA had been given more frequently, then one would expect that the chronic

threshold value identified would be lower. The ideal in vitro situation would

probably be a continuous perfusion supply and this would be expected to

reveal the lowest threshold values. However this would not mimic the

intermittent treatment effect when compounds are administered to patients.

The spheroid system described here, because of its stability (extending over

weeks) and the relative ease with which small samples can be removed

without perturbing the system (i.e. changing the drug cell number ratio) could


be used to provide in vitro data to match at least part of the theoretical

simulations of mathematical models [45].

Conclusion

The 3D spheroid model, because of its homeostatic ability to recover

following treatment and it stability over relatively long periods of time, offers

an excellent model for studying long term effects of compounds like VPA on

liver cells and for studying the effects of repeated insult leading for example to

the development of drug-induced liver injury.

Funding Information

This work was supported by MC2 Biotek, Hørsholm, Denmark and by the

University of Southern Denmark. The funders had no role in study design, data

collection and analysis, decision to publish or preparation of the manuscript.

The authors declare that there are no conflicts of interest.

Acknowledgments

The authors would like to thank Signe Marie Andersen, KiraEydJoensen

and Jacob BastholmOlesen for expert technical support in growing and

maintaining spheroid cultures and performing some of the assays.

References

[1] Lin RZ, Chou LF, Chien CC and Chang HY (2006) Dynamic analysis of

hepatoma spheroid formation: roles of E-cadherin and beta1-integrin.

Cell Tissue Res., 324:411-22. doi: 10.1007/s00441-005-0148-2.

[2] Loessner D, Stok KS, Lutolf MP, Hutmacher DW, Clements JA and

Rizzi SC (2010) Bioengineered 3D platform to explore cell-ECM

interactions and drug resistance of epithelial ovarian cancer cells.

Biomaterials, 31:8494-506. doi: 10.1016/j.biomaterials.2010.07.064.


[3] Daus AW, Goldhammer M, Layer PG and Thielemann C (2011)

Electromagnetic exposure of scaffold-free three-dimensional cell culture

systems. Bioelectromagnetics, 32:351-9. doi: 10.1002/bem.20649.

[4] Selden C, Khalil M and Hodgson H (2000) Three dimensional culture

upregulates extracellular matrix protein expression in human liver cell

lines--a step towards mimicking the liver in vivo? Int. J. Artif. Organs,

23:774-81.

[5] Wrzesinski K and Fey SJ (2012) After trypsinisation, 3D spheroids of

C3A hepatocytes need 18 days to re-establish similar levels of key

physiological functions to those seen in the liver. Toxicology Research,

in presshttp://xlink.rsc.org/?doi=C2TX20060K.

[6] Griffith LG and Swartz MA (2006) Capturing complex 3D tissue

physiology in vitro. Nature reviews. Molecular cell biology, 7:211-24.

doi: 10.1038/nrm1858.

[7] Pampaloni F, Reynaud EG and Stelzer EH (2007) The third dimension

bridges the gap between cell culture and live tissue. Nature reviews.

Molecular cell biology, 8:839-45. doi: 10.1038/nrm2236.

[8] Nelson LJ, Walker SW, Hayes PC and Plevris JN (2010) Low-shear

modelled microgravity environment maintains morphology and

differentiated functionality of primary porcine hepatocyte cultures.

Cells, tissues, organs, 192:125-40. doi: 10.1159/000308893.

[9] Wrzesinski K, Magnone MC, Visby Hansen L, Kruse ME, Bergauer T,

Bobadilla M, Gubler M, Mizrahi J, Zhang K, Andreasen CM, Eyd

Joensen K, Andersen SM, Bastholm Olesen J, Schaffalitzky de

Muckadell OB and Fey SJ (2013) HepG2/C3A 3D spheroids exhibit

stable physiological functionality for at least 24 days after recovering

from trypsinisation. Submitted.

[10] Fey SJ and Wrzesinski K (2012) Determination of Drug Toxicity Using

3D Spheroids Constructed From an Immortal Human Hepatocyte Cell

Line. Toxicological Sciences, 127:403-11. doi: 10.1093/toxsci/kfs122.

[11] Ghanizadeh A (2010) A novel explanation for potential toxic effects of

valproic acid on creatine: implications for autism. Mol. Genet. Metab.,

101:304. doi: 10.1016/j.ymgme.2010.07.019.

[12] Vamecq J, Joncquel-Chevalier Curt M, Mention-Mulliez K, Dobbelaere

D and Briand G (2011) Rise in brain GABA to further stress the

metabolic link between valproate and creatine. Mol. Genet. Metab.,

102:232-4. doi: 10.1016/j.ymgme.2010.10.006.


[13] Watkins PB and Seeff LB (2006) Drug-induced liver injury: summary of

a single topic clinical research conference. Hepatology, 43:618-31. doi:

10.1002/hep.21095.

[14] Holt M and Ju C (2010) Drug-induced liver injury. Handbook

ofexperimental pharmacology,:3-27. doi: 10.1007/978-3-642-00663-0_1

[15] Holt MP and Ju C (2006) Mechanisms of drug-induced liver injury.

AAPS J., 8:E48-54. doi: 10.1208/aapsj080106.

[16] Chateauvieux S, Morceau F, Dicato M and Diederich M (2010)

Molecular and therapeutic potential and toxicity of valproic acid.

Journalof biomedicine and biotechnology, 2010. doi: 10.1155/

2010/479364.

[17] Begriche K, Igoudjil A, Pessayre D and Fromenty B (2006)

Mitochondrial dysfunction in NASH: causes, consequences and possible

means to prevent it. Mitochondrion 6:1-28. doi: 10.1016/j.mito.

2005.10.004.

[18] Finsterer J and Zarrouk Mahjoub S (2012) Mitochondrial toxicity of

antiepileptic drugs and their tolerability in mitochondrial disorders.

Expert Opin. Drug Metab. Toxicol., 8:71-9. doi: 10.1517/17425255.

2012.644535.

[19] Luis PB, Ruiter JP, Aires CC, Soveral G, de Almeida IT, Duran M,

Wanders RJ and Silva MF (2007) Valproic acid metabolites inhibit

dihydrolipoyl dehydrogenase activity leading to impaired 2-

oxoglutarate-driven oxidative phosphorylation. Biochimica et biophysica

acta,1767:1126-33. doi: 10.1016/j.bbabio.2007.06.007.

[20] Silva MF, Aires CC, Luis PB, Ruiter JP, L IJ, Duran M, Wanders RJ and

Tavares de Almeida I (2008) Valproic acid metabolism and its effects on

mitochondrial fatty acid oxidation: a review. J. Inherit. Metab. Dis.,

31:205-16. doi: 10.1007/s10545-008-0841-x.

[21] Weiller M, Weiland T, Dunstl G, Sack U, Kunstle G and Wendel A

(2011) Differential immunotoxicity of histone deacetylase inhibitors on

malignant and naive hepatocytes. Exp. Toxicol. Pathol., 63:511-7. doi:

10.1016/j.etp.2010.04.001.

[22] Wrzesinski K (2009) Molecular Markers Associated With

Hepatotoxicity: Development Of In Vitro Test System Based On Human

Cells. Vdm Verlag, Saarbrücken, Germany.

[23] Riss TL, Moravec RA and Niles AL (2011) Cytotoxicity testing:

measuring viable cells, dead cells, and detecting mechanism of cell

death. Methods in molecular biology, 740:103-14. doi: 10.1007/978-1-

61779-108-6_12.


[24] Beger RD, Hansen DK, Schnackenberg LK, Cross BM, Fatollahi JJ,

Lagunero FT, Sarnyai Z and Boros LG (2009) Single valproic acid

treatment inhibits glycogen and RNA ribose turnover while disrupting

glucose-derived cholesterol synthesis in liver as revealed by the [U-

C(6)]-d-glucose tracer in mice. Metabolomics, 5:336-345. doi:

10.1007/s11306-009-0159-1.

[25] Kim AJ, Shi Y, Austin RC and Werstuck GH (2005) Valproate protects

cells from ER stress-induced lipid accumulation and apoptosis by

inhibiting glycogen synthase kinase-3. Journal of cell science,118:89-99.

doi: 10.1242/jcs.01562.

[26] Kim I, Xu W and Reed JC (2008) Cell death and endoplasmic reticulum

stress: disease relevance and therapeutic opportunities. Nature reviews.

Drug discovery, 7:1013-30. doi: 10.1038/nrd2755.

[27] Tsujimoto Y (1997) Apoptosis and necrosis: intracellular ATP level as a

determinant for cell death modes. Cell Death Differ., 4:429-34. doi:

10.1038/sj.cdd.4400262.

[28] Single B, Leist M and Nicotera P (2001) Differential effects of bcl-2 on

cell death triggered under ATP-depleting conditions. Exp. Cell Res.,

262:8-16. doi: 10.1006/excr.2000.5059.

[29] Heinrich A, Balszuweit F, Thiermann H and Kehe K (2009) Rapid

simultaneous determination of apoptosis, necrosis, and viability in sulfur

mustard exposed HaCaT cell cultures. Toxicology letters, 191:260-7.

doi: 10.1016/j.toxlet.2009.09.008.

[30] Posadas I, Santos P, Blanco A, Munoz-Fernandez M and Cena V (2010)

Acetaminophen induces apoptosis in rat cortical neurons. PloS one,

5:e15360. doi: 10.1371/journal.pone.0015360.

[31] Li W, Ghosh M, Eftekhari S and Yuan XM (2011) Lipid accumulation

and lysosomal pathways contribute to dysfunction and apoptosis of

human endothelial cells caused by 7-oxysterols. Biochemical and

biophysical research communications,409:711-6. doi: 10.1016/j.bbrc.

2011.05.071.

[32] Finsterer J and Mahjoub SZ (2012) Mitochondrial toxicity of

antiepileptic drugs and their tolerability in mitochondrial disorders.

Expert Opin. Drug Metab. Toxicol., 8:71-79.

[33] Marroquin LD, Hynes J, Dykens JA, Jamieson JD and Will Y (2007)

Circumventing the Crabtree effect: replacing media glucose with

galactose increases susceptibility of HepG2 cells to mitochondrial

toxicants. Toxicological sciences, 97:539-47. doi: 10.1093/toxsci/

kfm052.


[34] Diaz-Ruiz R, Rigoulet M and Devin A (2011) The Warburg and

Crabtree effects: On the origin of cancer cell energy metabolism and of

yeast glucose repression. Biochimica et biophysica acta, 1807:568-76.

doi: 10.1016/j.bbabio.2010.08.010.

[35] Rayasam GV, Tulasi VK, Sodhi R, Davis JA and Ray A (2009)

Glycogen synthase kinase 3: more than a namesake. Br. J. Pharmacol.,

156:885-98. doi: 10.1111/j.1476-5381.2008.00085.x.

[36] Jope RS, Yuskaitis CJ and Beurel E (2007) Glycogen synthase kinase-3

(GSK3): inflammation, diseases, and therapeutics. Neurochem. Res.,

32:577-95. doi: 10.1007/s11064-006-9128-5.

[37] Jope RS and Johnson GV (2004) The glamour and gloom of glycogen

synthase kinase-3. Trends Biochem. Sci., 29:95-102. doi: 10.1016/j.

tibs.2003.12.004.

[38] Luis PB, Ruiter JP, Ijlst L, Tavares de Almeida I, Duran M, Mohsen

AW, Vockley J, Wanders RJ and Silva MF (2011) Role of isovaleryl-

CoA dehydrogenase and short branched-chain acyl-CoA dehydrogenase

in the metabolism of valproic acid: implications for the branched-chain

amino acid oxidation pathway. Drug metabolism and disposition,

39:1155-60. doi: 10.1124/dmd.110.037606.

[39] Lee MH, Kim M, Lee BH, Kim JH, Kang KS, Kim HL, Yoon BI, Chung

H, Kong G and Lee MO (2008) Subchronic effects of valproic acid on

gene expression profiles for lipid metabolism in mouse liver. Toxicology

and applied pharmacology,226:271-84. doi: 10.1016/j.taap.2007.

09.014.

[40] Aires CC, Soveral G, Luis PB, ten Brink HJ, de Almeida IT, Duran M,

Wanders RJ and Silva MF (2008) Pyruvate uptake is inhibited by

valproic acid and metabolites in mitochondrial membranes. FEBS letters

582:3359-66. doi: 10.1016/j.febslet.2008.08.028.

[41] Bailey MJ and Dickinson RG (1996) Chemical and immunochemical

comparison of protein adduct formation of four carboxylate drugs in rat

liver and plasma. Chemical research in toxicology, 9:659-66. doi:

10.1021/tx960017o.

[42] Cannell GR, Bailey MJ and Dickinson RG (2002) Inhibition of tubulin

assembly and covalent binding to microtubular protein by valproic acid

glucuronide in vitro. Life Sci., 71:2633-43.

[43] Anon (2012) Valproate Sodium Injection. In: FDA http://www.drugs.

com/pro/valproate.html.


[44] Schulz M and Schmoldt A (2003) Therapeutic and toxic blood

concentrations of more than 800 drugs and other xenobiotics. Die

Pharmazie,58:447-74.

[45] Grover A and Benet LZ (2011) Intermittent drug dosing intervals guided

by the operational multiple dosing half lives for predictable plasma

accumulation and fluctuation. J. Pharmacokinet Pharmacodyn, 38:369-

83. doi: 10.1007/s10928-011-9198-0.

Reviewed by: Dr Cliff Rowe, Research associate, MRC Centre for Drug Safety

Science, Department of Pharmacology and Therapeutics, Sherrington

Building, Ashton Street, The University of Liverpool, L69 3GE, U.K.,

Tel: +44 (0)151 794 6615, Fax: +44 (0)151 794 5540.

Documents

Determination of Acute Lethal and Chronic Lethal Thresholds of … · bioreactors (MC2 Biotek, (Hørsholm, Denmark) Cat. no. 010). These bioreactors are specially constructed to be