The G11778A LHON mutation does notenhance ethambutol cytotoxicity in a cybridmodel

R. Pommer1, S. Schoeler1, C. Mawrin2, R. Szibor3 and E. Kirches1

Institutes of 1Neuropathology and 3Forensic Medicine, Otto von Guericke University

Magdeburg, 2Institute of Neuropathology, Friedrich Schiller University Jena,

Germany

Ethambutol tox ic ity in LHON cybrids

Ab stract. Leber’s he red i tary op tic neu -rop a thy (LHON) is a ma ter nally in her ited mi -to chon drial dis or der, lead ing to a se lec tiveloss of ret i nal gan glion cells (RGC) and de -gen er a tion of the op tic nerve, which re sults inse vere vi sual im pair ment or even blind ness.The pri mary causes are point mu ta tions of themi to chon drial DNA (mtDNA), as so ci atedwith aminoacid ex changes in com plex I of the elec tron trans port chain (ETC), which arethought to dis turb ox i da tive ATP gen er a tionin the mi to chon dria. The ma jor side ef fect ofthe an ti bi otic ethambutol, com monly used intu ber cu lo sis ther apy, is a retinopathy, whichmay lead to se lec tive RGC loss, if not de -tected in an early stage. More over, LHONwas re ported to be elic ited by ethambutol insome mu ta tion car ri ers. Ob jec tive: The pres -ent study in tended to mea sure a pos si ble syn -er gism be tween mi to chon drial dys func tion,caused by the most com mon LHON mu ta tion(G11778A) and caused by ethambutol, which may lead to a higher cytotoxicity of the drugin LHON cells. Ma te rial: An NT2/D1 tera -toma-de rived LHON cybrid line and theparen tal cells. Method: De ter mi na tion ofethambutol tox ic ity in both lines, us ing amicrotiter tetrazolium as say, luminometricmea surement of ATP/ADP ra tios and de ter -mi na tion of mtDNA copy num bers by Re al -time PCR. Re sults: Short-term etham butoltoxi city oc curred only at mmo lar concentra -tions, far be yond the es ti mated plasma peakconcentrations of pa tients un der antibioticther apy. No sig nif i cant dif fer ence oc curred be -tween both cell lines. The ATP/ADP ra tios inthe cybrids were sur pris ingly low, but showed no cor re la tion with the mutational sta tus ofdrug-treated cells. The mtDNA copy num berof treated LHON and pa ren tal cells did notdif fer sig nif i cantly. Con clu sions: Ethambutol shows no syn er gism with the most com monpri mary LHON mu ta tion with re spect to mi -to chon drial en ergy pro duc tion or mtDNArep li ca tion in cybrid cells, al though the is sueof ATP de cline should be fur ther ad dressed in

neuronally dif fer en ti ated cybrids with com -plete OXPHOS de pend ency.

Introduction

Leber’s he red i tary op tic neu rop a thy

(LHON), a neuroophthalmologic dis ease ori -

ginally dis cov ered by the Ger man oph thal -

mol o gists von Graefe and Leber in the mid dle

of the 19th cen tury [Leber 1871, von Graefe

1858], is mean while known to be a ma ter nally

in her ited mi to chon drial dis or der [Wallace et

al. 1988]. Its pri mary cause are point mu ta -

tions of the mi to chon drial DNA (mtDNA),

which lead to the ex change of a sin gle amino

acid of NADH : ubiquinone oxidoreductase, a

large L-shaped pro tein com plex (com plex I)

in the in ner mi to chon drial mem brane, do -

nated to in tro duce NADH de rived elec trons

into the elec tron trans port chain (ETC). The

lo ca tion of the mu ta tions in com plex I and

early bio chem i cal stud ies sug gested an al -

tered ubiquinol site and en hanced prod uct in -

hi bi tion of the mu tated com plex [Carelli et al.

1999, Majander et al. 1996]. While a de creased

en zyme ac tiv ity in mu tant lymphoblasts, skin

fibroblasts or cybrid cells re mains contro ver -

sal for the most com mon mu ta tion G11778A,

com plex I res pi ra tion was found to be de -

creased in sev eral stud ies [Brown et al. 2000,

Floreani et al. 2005]. More over, the neu rol o -

gist Valerio Carelli and co work ers dem on -

strated a clear de fect in ox i da tive ATP syn the -

sis [Baracca et al. 2005] as well as in cel lu lar

ATP con tent un der met a bolic stress con di -

tions in a cybrid model [Zanna et al. 2003,

2005], re sult ing in en hanced apoptotic sen si -

tiv ity of the LHON cybrids. These and other

NP - 7079/ 18.06.2008

Key words

LHON – ethambutol –

cybrid – ATP – mtDNA

Re ceived

De cem ber 11, 2007;

ac cepted in re vised form

May 5, 2008

Cor re spon dence to

Dr. E. Kirches

Institut für Neuropatho -

logie Otto-von - Guericke -

Universität, Leipziger

Straße 44, 39120

Magdeburg, Ger many

Elmar.kirches@

medizin.uni-magdeburg.de

Clin i cal Neuropathology, Vol. 27 – No. n/2008 (nnn-nnn)

©2008 Dustri-Verlag Dr. K. FeistleISSN 0722-5091

ob ser va tions fa vored strongly the hy poth e sis

of ret i nal gan glion cell (RGC) death, me di -

ated by a com bi na tion of neuronal ATP de fi -

ciency and en hanced superoxide gen er a tion

in com plex I [Carelli et al. 2004a]. Ac cord ing

to this hy poth e sis, the se lec tive vul ner a bil ity

of RGC may be caused by their high ox i da tive

en ergy de mand. The axon sec tions in the ret i -

nal nerve fi ber layer are unmyelinated, thus

requiering more mi to chon drial ATP to save

the plasma mem brane po ten tial. This ar gu -

ment is un der lined by the stron ger histo -

chemical cytochrome c oxidase stain ing of

these intraocular por tions of the op ti cal ax ons

[Carelli et al. 2004b]. Al though the pri mary

causes of LHON are known, ad di tional fac -

tors in flu ence the like li hood of dis ease in mu -

ta tion car ri ers. This can be con cluded from

the in com plete penetrance, rang ing be tween

25% and 80% in dif fer ent pop u la tions and

stud ies [Mackey and But tery 1992, Man et al.

2003, www.ncbi.nlm.nih.gov/omim], and

from the strong male pre dom i nance. While

the lat ter sug gests the par tic i pa tion of a yet un -

known, recessive X-al lele [Hud son et al.

2005], the dis ease-pro mot ing role of al co hol

abuse [Sadun et al. 2002, 2003, 2004] un der -

lines the im por tance of epigenetic fac tors, tar -

get ing the ret ina.

One such fac tor may be the an ti bi otic

ethambutol (EMB), which has com monly

been used since 1961 to treat tu ber cu lo sis

[Fraunfelder et al. 2006]. The most prom i nent

EMB side ef fect is a retinopathy, which re sem -

bles LHON in mul ti ple fea tures and was called

ethambutol op ti cal neu rop a thy (EBON) by

Ikeda and col leagues [Ikeda et al. 2006]. This

type of retinopathy is also char ac ter ized by a

se vere loss of vi sual acu ity, dis turbed color

vi sion and cen tral scotoma. Like LHON, the

dis ease pre dom i nantly af fects the smaller

RGC of the macula and the cor re spond ing

smaller ax ons of the papillomacular bun dle.

In con trast to LHON, the dis ease is most of ten

re vers ible af ter chang ing the an ti bi otic re -

gime, if de tected at an early stage by the oph -

thal mol o gist. It may, how ever, procede to se -

vere vi sual im pair ment or blind ness, in some

cases even af ter early de tec tion of the symp -

toms and drug ommission. In re cent years,

some cases of G11778A mu ta tion car ri ers

have been de scribed, which de vel oped

LHON dur ing an ti bi otic EMB treat ment and

were orig i nally rec og nized in this way as mu -

ta tion car ri ers (sum ma rized by [Ikeda et al.

2006]).

This sit u a tion led us to ex am ine the cyto -

toxicity of EMB in a cybrid cell cul ture mo -

del. For this pur pose, NT2/D1 teratoma- de -

rived transmitochondrial cybrid cells, in

which the tu mor mtDNA had been com -

pletely re placed by the mtDNA of a G11778A

LHON pa tient [Wong et al. 2002], were com -

pared with the pa ren tal cell line. This of fers

the pos si bil ity to com pare cells with and with -

out the LHON mu ta tion un der con di tions of

an iden ti cal chro mo somal back ground. Af ter

es tab lish ing EMB dose re sponse curves in a

cytotoxicity as say, ATP/ADP ra tios were

meas ured as a di rect in di ca tor of al tered ox i da -

tive phosphorylation (OXPHOS). Further -

more mtDNA copy num bers were mea sured

with re spect to the hy poth e sis, that EMB may

in hibit mtDNA rep li ca tion and, thus, in di -

rectly im pair OXPHOS.

Materials and methods

Cell lines and genetic

confirmation

The Ntera-2 clone D1 (NT2/D1) teratoma

cell line was pur chased from the Amer i can

Type Cul ture Col lec tion (ATCC, Manassas,

VA, USA). The NT2/D1-de rived LHON cy -

brid clone was a kind gift from Prof. Dr. Gino

Cortopassi and Prof. Dr. Al ice Wong from the

Uni ver sity of Cal i for nia (Da vis, CA, USA).

The nu clear ge netic iden tity of both cell lines

was proven by iden ti cal allelic pat terns of a

col lec tion of microsatellite mark ers and ame -

logenin, us ing the MenType Nonaplex II PCR

kit (BioType, Dresden, Ger many) as de -

scribed else where [Schoeler et al. 2007]. In

brief, the poly mor phic marker re gions were

PCR-am pli fied with flu o res cence-la beled

pri mers and re ac tion prod ucts un der went

elec tro pho re sis on an ABI310C cap il lary se -

quencer (Ap plied Biosystems, Fos ter City,

CA, USA), fol lowed by au to matic al lele iden -

ti fi ca tion by the GeneScan soft ware (Ap plied

Biosystems). No allelic dif fer ences were ob -

served. The pres ence of 100% mu tant G -

11778A mtDNA in the LHON cybrid clone

was proven by di rect se quenc ing and a spe -

cific PCR-RFLP as say, as de scribed in de tail

else where [Schoeler et al. 2007]. In brief, the

Pommer, Schoeler, Mawrin et al. 2

NP - 7127/ 18.06.2008

lat ter as say con sisted of a PCR of the ND4 re -

gion of the mtDNA, fol lowed by a di gest with

the re stric tion endonuclease SfaNI. While

the mu ta tion con fers re sis tance to the di gest,

the wild type is cut. 6-FAM-la beled PCR

frag ments were sep a rated on an ABI310C

cap il lary se quencer, fol lowed by au to matic

frag ment iden ti fi ca tion by the GeneScan soft -

ware. The sta bil ity of the LHON mu ta tion

was checked over the whole time pe riod of

the work. No re main ing wild type DNA could

be mea sured within the sen si tiv ity limit of the

as say (< 1%) in all sub se quent rounds of cell

cul ture. The ab sence of the three pri mary

LHON mutations at nu cle o tide po si tions

14484, 3460 and 11778 in the NT2/D1 line

was checked by di rect se quenc ing and for the

lat ter two mu ta tions with an PCR-RFLP as -

say, as de scribed else where [Schoeler et al.

2007]. Fur ther more, se quenc ing re vealed no

other patho genic mu ta tions which dis tin -

guished the mtDNAs of the two clones. The

only ad di tional dif fer ences ob served were

known poly morphisms.

Cell culture and EMB treatment

All cell lines were cul tured in high glu -

cose (4.5 g/l) DMEM with a low ered bi car -

bon ate con tent (1.5 g/l) from ATCC, sup ple -

mented with 1 mM so dium pyruvate, 4 mM

L-glutamine, 10% fe tal calf se rum (EUGold,

PAA, Linz, Austia) and 1% penicil line/strep -

tomycine (PAA). Cells were cul tured in T75

flasks at 37 °C in 5% CO2 at mo sphere, un til

80% confluency was reached. They were split

by trypsinization in a ra tio of 1 : 4 or higher.

For ex per i ments in 96-, 48- or 24-well plates,

an ad e quate cell num ber per well (see re sults

sec tion) was seeded and in cu bated for 24 h.

Af ter this preincubation, var i ous con cen tra -

tions of EMB (Riemer Arzneimittel AG,

Greifs wald, Ger many) were ap plied for the

in di cated time pe ri ods, fol lowed by micro -

titer-tetrazolium (MTT) cell sur vival as say.

For ATP/ADP and mtDNA copy num ber as -

says, the cells were seeded in the in di cated

den sity (see re sults) in 6-well plates, treated

with var i ous EMB con cen tra tions and sub -

jected to lysis in ATP buffer or DNA ex trac -

tion, re spec tively. Since the dif fer en ti a tion

sta tus of the cybrid lines was sus pected to in -

flu ence their OXPHOS de pend ency, cell sur -

vival and ATP/ADP ra tios were also meas -

ured, fol low ing a short-term dif fer en ti a tion

with retinoic acid (RA), which stopped cell

pro lif er a tion. This was achieved by seed ing

cells as usual, but re plac ing DMEM af ter a

24-h in cu ba tion by se rum-free Neurobasal

(Invitrogen, Karlsruhe, Ger many), en riched

with 5% B27 sup ple ment (Invitrogen) and

10 mM all-trans retinoic acid (RA) from Sigma

(St. Louis, USA). Cells were in cu bated in this

me dium for 4 days, prior to the in di cated

ethambutol treat ment.

MTT assay

For the microtiter-tetrazolium as say one

vol ume of cell cul ture me dium, con tain ing

1.5 mg/ml 3-(4,5-dimethylthiazol- 2-yl)- 2,5-

di phenyl ter tra zolium (MTT, Sigma, St.

Loius, USA) was added to the cul ture wells

and in cu bated at 37 °C for fur ther 2 h. There -

af ter, the me dium was re placed by DMSO,

which lysed the cells and dis solved the gen er -

ated formazan dye. Op ti cal den si ties (OD) at

562 nm (ref er ence wave length 630 nm) were

mea sured in a microplate reader, and cell sur -

vival was ex pressed as the OD per cent age,

nor mal ized to untreated controls.

ATP/ADP ratio

Af ter trypsinization, cells were shortly

washed 2 times in phos phate-buf fered sa line

(PBS), re sus pend ed in 300 ml boil ing ATP

buffer (100 mM Tris pH 7.75, 10 mM KCl, 4

mM MgCl2) and boiled for fur ther 2 min. Af -

ter a short high-speed centrifugation, the su -

pernatant was dis trib uted into two fresh

tubes. Af ter cool ing to room tem per a ture, 1.5

mM phosphoenolpyruvate (Sigma) and 6

units/ml pyruvate kinase (Sigma) were added

to 1 of the 2 aliquots and in cu bated for 30 min

at room tem per a ture, to phosphorylate all

ADP in the sam ple to ATP. Af ter this step, the

ali quot was boiled for 1 min, to de stroy the

pyruvate kinase ac tiv ity. Then 50 ml of an ap -

pro pri ate di lu tion of both lysates (1 : 20 – 1 :

50) was mixed with 50 ml of lu ci fer ase re ac -

tion mix (CLS II ATP kit, Roche, Mannheim,

Ger many). The re sult ing lu mi nes cence was

im me di ately mea sured in a TD20/20 lumino -

meter (Turner De sign In stru ments, Sunny vale,

Ethambutol tox ic ity in LHON cybrids 3

NP - 7079/ 18.06.2008

CA, USA). The ATP con cen tra tions of the

sam ples were cal cu lated ac cord ing to a di lu -

tion se ries of a stan dard in the range be tween

62.5 nM and 500 nM. While the pyruvate

kinase-treated lysate rep re sented the sum of

ATP and ADP con cen tra tions, the un treated

lysate rep re sented only ATP, al low ing the cal -

cu la tion of ATP/ADP ra tios.

mtDNA copy number

To de ter mine the mtDNA copy num ber,

nor mal ized to a nu clear sin gle copy gene,

SYBR green Realtime PCR was used. To tal

DNA was ex tracted from the EMB-treated

cells in 6-well plates, us ing the QIAamp DNA

Mini kit (Qiagen, Va len cia, CA, USA). Mi to -

chon drial DNA was am pli fied from these

samples by the hypervariable re gion 2 (HVR2)

primers: 5’-CTATCACCC TATT AACC ACT-

3’ (for ward) and 5’-GTTAAAAG TGCA -

TACC GCCA-3’ (re verse), while exon 8 of

the p53 gene was the nu clear ref er ence am -

plicon: 5’-ACTGCCTCTTGCTTCTCTT-3’

(forward) and 5’-AAGTGAATCTGA GGCA -

TAAC-3’ (re verse). The ref er ence amplicon

had been found ear lier in our lab o ra tory to

show no sig nif i cant dif fer ences be tween the

two cell lines, if nor mal ized to var i ous other

nu clear genes. It was cho sen to faciliate PCR

by al low ing equal ther mal con di tions for both

amplicons. PCR am pli fi ca tion was per -

formed from 50 ng tem plate DNA with the

QuantiTect kit (Qiagen, Va len cia, CA, USA)

ac cord ing to the man u fac tur ers in struc tions

on an ABI Prism 7000SDS cycler (Ap plied

Biosystems). The ther mal con di tions were

2 min 50 °C, 10 min 95 °C, fol lowed by 40 cy -

cles with 95 °C de na tur ation, 55 °C an neal ing

and 72 °C ex ten sion, 30 sec each step. Copy

num bers of both amplicons were au to mat i -

cally cal cu lated by the cycler soft ware, us ing

a di lu tion se ries of HVR2 and p53exon8 stan -

dards in the same run. The pu rity of re ac tion

prod ucts was iden ti fied by their melt ing tem -

per a ture.

Statistical analysis

Com par i sons of sur viv ing cell frac tions,

ATP/ADP ra tios and mtDNA copy num bers

un der EMB treat ment with the un treated con -

trols were per formed us ing a 1-way ANOVA,

to elu ci date whether the drug gen er ally had an

ef fect on these pa ram e ters. The spe cific drug

con cen tra tions, which ex hib ited a sta tis ti cally

sig nif i cant dif fer ence to the con trols, were

iden ti fied by a Dunnett post-hoc test. Sig nif i -

cant dif fer ences be tween the two cell lines for

a given EMB con cen tra tion were de ter mined

by a t-test. All cal cu la tions were per formed

with the soft ware pack age SPSS 13 (SPSS,

Chi cago, USA), as sum ing sig nif i cance for

p < 0.05. All di a grams show means ± SEM.

Results

Cell survival under EMB

treatment

30.000 cells were seeded in 24-well plates

in DMEM, in cu bated for 24 h and treated for

an other 24 h with the EMB con cen tra tions in -

di cated in Fig ure 1. The toxic ef fects, which

could be mea sured in the MTT cell sur vival

as say, were sur pris ingly low, even if mmo lar

con cen tra tions of the drug were ap plied for

this short time pe riod. Less than 30% of cells

died at 2.7 mM, while the rel e vant con cen tra -

tions in vivo are thought to be in the range be -

tween 20 and 80 mM (see Dis cus sion). How -

ever, tox ic ity be came sig nif i cant (p < 0.01) in

both cell lines for con cen tra tions above 1.2

mM in the Dunnett test and was clearly dose-

dependent. Al though the dose re sponse curve

for the LHON mu tant was be neath that one of

the wild type over the whole con cen tra tion

range, the dif fer ence be tween the lines only

be came sig nif i cant (p < 0.01) for the high est

con cen tra tion tested (2.7 mM), which is

thought to be ir rel e vant in vivo.

To elu ci date pos si ble ef fects of the plate

type used, of cybrid dif fer en ti a tion and of

EMB ex po sure times, sev eral sim i lar ex per i -

ments were per formed. They all had es sen -

tially the same re sults, and are, there fore, not

shown in de tail in seperate Fig ures. All ex per -

i ments re vealed a clear dos age-de pend ent

tox ic ity of the drug, but no proof of a higher

sen si tiv ity of the LHON mu tant as com pared

to the wild type.

In one se ries of ex per i ments, 3,000 cells

per well were seeded into 96-well plates and

the same EMB con cen tra tions, as shown in

Fig ure 1, were ap plied for 24 h. The tox ic ity

Pommer, Schoeler, Mawrin et al. 4

NP - 7127/ 18.06.2008

of the drug seemed to be some what higher in

this as say, since the max i mal frac tion of dead

cells reached about 40% for 2.7 mM EMB.

The slight dif fer ence be tween both plate

types may be due to the higher per cent age of

cells, which are lost dur ing suck ing the cell

cul ture liq uid out off the wells, since EMB

seemed to re duce the ad he sion of the cells to

plas tic sur faces. The drain ing of cells with

low ered ad he sion be comes more prom i nent

in smaller wells. In this as say vari ant, the dose

re sponse curve of wild type NT2/D1 cells was

usu ally be low that one of the LHON mu tant,

al though this dif fer ence again be came sig nif i -

cant only for the highest concentrations

above 2.4 mM.

Even short-term dif fer en ti a tion with reti -

noic acid (RA) did not lead to an es sen tially

dif fer ent out come. For this pur pose, 5,000

cells were seeded into 48-well plates, fol -

lowed by 4 days RA dif fer en ti a tion, prior to

the 24 h EMB treat ment in the same con cen -

tra tion range, as shown in Fig ure 1. Again, a

clearly dos age de pend ent tox ic ity of the drug

oc curred, which was very sim i lar to that of

un dif fer en ti ated cells. No el e vated sen si tiv ity

of the mutant clone was found.

In vivo the drug in flu ences the ret i nal gan -

glion cells (RGC) over long-time pe ri ods of

usu ally sev eral monthes. Al though a chronic

treat ment could not be sim u lated in the MTT

as say, the time pe riod of EMB in cu ba tion

of the un dif fer en ti ated cells was ex tended in

one experimental se ries to the pos si ble max i -

mum of 4 days. This limit re sulted from the

growth of the un treated con trols which died

about 24 h af ter reach ing confluency. In all,

2,000 cells were seeded in 96-well plates. Af -

ter 24 h of preincubation in DMEM, they

were treated with the es tab lished EMB con -

cen tra tions for a time pe riod of 4 days, prior to

MTT as say. Af ter this pro longed in cu ba tion,

tox ic ity reached a max i mum around 45% cell

death for 2.7 mM EMB, very sim i lar to the re -

sults ob tained for the 24-h ex po sure. Again,

the dos age-de pend ent ef fect was sta tis ti cally

sig ni ficant, without increased response of the

mutant.

Due to the lim ited EMB tox ic ity in the

MTT as say, we in tended to find out whether

ex tremely high con cen tra tions of the drug

may cause a more pro nounced cell death in

the LHON cybrids. Ex tend ing the con cen tra -

tion range to 12 mM lead to death of more

than 80% of the cells, with out el e vated sen si -

tiv ity of the mu tant clone. While 50% death

had not been reached in the pre vi ous ex per i -

ments, the LD50 could now be de ter mined to

be around 4 mM in both cell lines. Taken to -

gether, all types of MTT as says did not re veal

a higher sen si tiv ity of the mu tant against the

cytotoxic effect of EMB.

For con cen tra tions above 0.5 mM, a clear

morphologic change oc curred in both lines.

Small vac u oles ap peared in phase-con trast

mi cro graphs. Num ber and mean di am e ter of

these vac u oles in creased with EMB con cen -

tra tion, but no ob vi ous dif fer ence was vis i ble

in the mi cro scope be tween both cell lines. Af -

ter 24 h in 2.7 mM of the drug, the cy to plasm

was filled with large num bers of these vac u -

oles (Fig ure 2). Their or i gin and im pact for

the cells are yet unknown.

ATP/ADP ratio

Here 100,000 cells per well were seeded

into 6-well plates, in cu bated 24 h in DMEM

and ex posed to 0 (con trol), 0.6 or 2.7 mM

EMB for 24 h, prior to the luminometric

ATP/ADP as say. For ex per i ments with dif fer -

en ti ated cy brids, only 50,000 cells were ini -

tially plated, but a 4-day dif fer en ti a tion was

Ethambutol tox ic ity in LHON cybrids 5

NP - 7079/ 18.06.2008

Fig ure 1. Sur viv ing cell frac tion (%) in re sponse to

in creas ing EMB con cen tra tions (mM) for NT2/D1

pa ren tal cells (NT2) and cybrids with the G11778A

mtDNA mu ta tion. * = sig nif i cant dif fer ence to un -

treated con trol cells (C) of the same line (p < 0.01).

The star sym bol re fers to both lines, since a sig nif i -

cant ef fect oc curred at iden ti cal con cen tra tions in

both clones. # = sig nif i cant dif fer ence be tween the 2

lines for 2.7 mM EMB (p < 0.01).

in cluded in the pro to col, prior to EMB

treatment.

Gen er ally, the en er getic sta tus of the cy -

brids was sur pris ingly low. The untreated

LHON mu tants in DMEM did not show per se

a sig nif i cantly lower ATP/ADP ra tio, when

com pared with un treated NT2/D1 cells. This

would have been in con cor dance with a hy -

poth e sis of the lit er a ture, as sum ing that pro -

lif er at ing cybrid cells in DMEM pos ses a

largely nonoxidative (glycolytic) en ergy me -

tab o lism, which masks the ef fects of the mu -

ta tions. How ever, the LHON cells even ex -

hib ited a higher ra tio, when com pared with

the wild type (p < 0.05), a rather un ex pected

re sult (Fig ure 3). EMB se lec tively de creased

the ra tio in the mu tant (p < 0.05), thus, lead ing

to iden ti cal ATP/ADP ra tios in mu tant and

wild type. This ex per i ment was not fi nally

con clu sive. It sug gested no gen eral im pact of

EMB on OXPHOS, since the en ergy sta tus of

the wild type did not worsen un der the in flu -

ence of the drug. On the other hand, a pos si ble

in flu ence se lec tively in the mu tant clone was

Pommer, Schoeler, Mawrin et al. 6

NP - 7127/ 18.06.2008

Fig ure 2. Phase con trast mi cro graph show ing rel a tively nor mal mor phol ogy, but ac cu mu la tion of large

amounts of ves i cles in cyrid cells with the G11778A LHON mu ta tion, as well as in pa ren tal cells, in cu bated

with 2.7 mM EMB. Un treated cells are shown in A (NT2/D1) and B (LHON cybrids), while mi cro graphs C

(NT2/D1) and D, E (LHON cybrids) show cells af ter EMB ex po sure. The bars rep re sent 50 mm.

in con clu sive, since the sta tis ti cally sig nif i -

cant ef fect of EMB in this clone de pended

com pletely on the ab er rant be hav ior of the

un treated LHON cells (Fig ure 3).

For this rea son and due to the sug gested

glycolytic en ergy me tab o lism of the un dif fer -

en ti ated cybrids in DMEM, the ex per i ment

was re peated with dif fer en ti ated cells (Fig ure

4), which were thought to ex hibit en hanced

OXPHOS de pend ency. In this case, no signi -

ficant dif fer ences at all oc curred, nei ther be -

tween the cell lines, nor be tween treated and

un treated groups. Taken to gether, these ex -

peri ments sug gest no ma jor in flu ence of EMB

on OXPHOS.

mtDNA copy number

150,000 cells were seeded into each well

of a 6-well plate. Af ter 24 h in DMEM, 0(con -

trol), 0.6 or 2.7 mM EMB were added. The

cells were in cu bated in this me dium for 5

days, to al low some mtDNA turn over. The

copy num ber of mtDNA, nor mal ized to a nu -

clear sin gle copy gene (p53), was as sayed by

SYBR green Realtime PCR in the ex tracted

DNA sam ples (Fig ure 5). The cells con tained

a mean of about 600 mtDNA cop ies per p53

copy. Al though, the mtDNA con tent in the pa -

ren tal cell line seemed to be slightly higher as

com pared with the LHON cells, this dif fer -

ence did not reach sta tis ti cal sig nif i cance.

EMB had no in flu ence on the mtDNA con -

tent, nei ther in the wild type, nor in the LHON

cells, sug gest ing no influence on mtDNA re -

plication.

Discussion

In a se ries of 24 pa tients with EMB-in -

duced retinopathy, Hwang and col leagues

[Hwang et al. 2003] did not find any of the

known pri mary LHON mu ta tions. This sug -

gested that the mu ta tion in ci dence in such

patients is not in creased. How ever, this re sult

does not con clu sively ex clude a syn er gism

be tween ge netic sta tus and drug, since mu ta -

tion car ri ers are rel a tively rare in the gen eral

population. On the other hand, sev eral car -

riers of the G11778A tran si tion had been

iden ti fied by the de vel op ment of LHON dur -

ing tu ber cu lo sis ther apy with EMB, sug gest -

ing that the drug fa vors dis ease out break. This

sug ges tion was un der lined by four cases (re -

viewed by [Ikeda et al. 2006]), in which the

age of LHON on set in two males and two fe -

males was be tween 53 and 70 years, while

spo radic or fa mil ial LHON dis ease usu ally

af fects young men in their first three de cades

of life.

While the neurotoxic side ef fects of EMB

are well-doc u mented in tu ber cu lo sis pa tients,

the tox ic ity has not been well-de fined in RGC

cul tures. In the last years, two groups in ves ti -

gated these toxic ef fects in dis so ci ated rat ret -

Ethambutol tox ic ity in LHON cybrids 7

NP - 7079/ 18.06.2008

Fig ure 4. ADTP/ADP ra tios of retinoic acid dif fer -

en ti ated NT2/D1 pa ren tal cells (NT2) and LHON

cybrids (11778), treated for 24 h with the in di cated

EMB con cen tra tions. n = 5 in de pend ent ex per i -

ments.

Fig ure 3. ATP/ADP ra tios of NT2/D1 pa ren tal cells

(NT2) and LHON cybrids (11778) treated for 24 h with

the in di cated EMB con cen tra tions. * = sig nif i cant dif -

fer ence to the cor re spond ing EMB-free con trol, # =

sig nif i cant dif fer ence be tween the 2 lines, grow ing in

DMEM with out the drug. n = 4 in de pend ent ex per i -

ments.

i nal cell cul tures. Both groups ap plied the

drug for 24 h di rectly to the me dium. Heng et

al. [1999] used an MTT as say, but quan ti fied

the formazan de po si tion se lec tively in RGC

with a mi cro scop i cal tech nique. In their ex -

per i ments, the most pro nounced de crease in

vi a ble cell num ber oc curred for con cen tra -

tions be tween 1 mM and 100 mM, the lat ter

con cen tra tion re sult ing in about 50% cell

death. Sur pris ingly, even con cen tra tions in

the mM range did not fur ther en hance cell

death (Fig ure 1A of their pub li ca tion). In con -

trast, Yoon et al. [2000] mea sured a more lin -

ear de crease of the vi a ble cell num ber for

EMB con cen tra tions be tween 200 mM and

1 mM, us ing di rect cell count ing of mor pho -

log i cally in tact MAP-2-pos i tive neu rons.

Their dose re sponse curves sug gested an

LD50 around 800 mM (Fig ure 3A of their pub -

li ca tion). Since the plasma peak con cen tra -

tions in EMB treated tu ber cu lo sis pa tients are

in the range be tween 20 and 80 mM [Lee et al.

1977, Peets et al. 1965], the first study better

re flects an im me di ate ef fect of these rel e vant

con cen tra tions on RGC vi a bil ity, while the

sec ond does not. How ever, cell cul ture stud -

ies can not de ter mine the long term pro -

apoptotic ef fects of the drug, which might oc -

cur at much lower con cen tra tions. Since the

teratoma cybrids, used in the pres ent study,

are known to be less sen si tive against var i ous

neurotoxic and pro apop totic stim uli as com -

pared to pri mary neu ron cul tures, the lower

sen si tiv ity to EMB treat ment was not sur pris -

ing. The LD50 was around 4 mM for a 24-h

ap pli ca tion. A se lec tively higher sen si tiv ity

of the LHON cybrids was not ob served. In ter -

est ingly, at drug con cen tra tions of 500 mM or

higher, many vac u oles ac cu mu lated in the cy -

to plasm. With re spect to this fea ture, the

cybrids be haved largely like cul tured RGC,

which ac cu mu lated these vac u oles at a con -

cen tra tion range be tween 700 and 1,000 mM

[Yoon et al. 2000]. The na ture of these vac u -

oles re mains un known. They seemed to be

sur rounded by only a sin gle mem brane and,

thus, not to be of mi to chon drial, but even tu -

ally of ER or Golgi or i gin. Al though the cy to -

plasm may be come densely filled with vac u -

oles, their ap pear ance is re vers ible, since they

dis ap peared af ter re mov ing EMB from RGC

cul tures.

The mech a nism of EMB neurotoxicity re -

mains un clear up to now, but a di rect at tack of

OXPHOS com plexes had been hy poth e sized

as one pos si bil ity. The known cop per che lat -

ing ef fect of the drug was sug gested to lead to

in hi bi tion of cytochrome c oxidase (COX) ac -

tiv ity [Ikeda et al. 2006]. This in hi bi tion may

cause a sig nif i cant re duc tion of elec tron flux

through the ETC, thus, di min ish ing ox i da tive

ATP syn the sis. In the highly ATP de mand ing,

unmyelinated intraocular por tions of RGC

ax ons, this OXPHOS re duc tion may lead to

ATP de fi ciency and cell death. A direct

OXPHOS in ter fer ence would be as sumed to

be fast and, thus, to be mea sur able in LHON

cybrid cul tures. COX ac tiv i ties or res pi ra tion

rates of EMB treated cells could not be meas -

ured di rectly in the pres ent study with re li able

bio chem i cal meth ods. This ap proach re quires

rel a tively high cell num bers, which were not

avail able, es pe cially not fol low ing RA treat -

ment. In con trast, a re sult ing ATP de cline can

eas ily be mon i tored us ing min i mal amounts

of cells, but it must be kept in mind that cybrid

cells may rely mainly on glyco lytic ATP gen -

er a tion, un less they are forced to switch to the

ox i da tive path way. Al though a short-term RA

dif fer en ti a tion of the teratoma based cybrids

was ap plied in the pres ent stu dy, to en hance

their OXPHOS de pend ency, even these cul -

tures may still rely mainly upon glycolytic

ATP gen er a tion. This may ham per the de tec -

tion of pos si ble OXPHOS ef fects of EMB.

More over, the lim ited mor pho log i cal ef fect

of EMB sug gests a low ef fi cacy in the model.

The prob lem may be solved by ap ply ing the

drug to cybrids af ter ter mi nal neuronal dif fer -

Pommer, Schoeler, Mawrin et al. 8

NP - 7127/ 18.06.2008

Fig ure 5. mtDNA copy num ber nor mal ized to the

nu clear p53 gene in pa ren tal teratoma cells (NT2)

and LHON cybrids (11778), treated for 5 days with

the in di cated EMB con cen tra tions. n = 2 in de pend -

ent ex per i ments.

en ti a tion. Dif fer en ti a tion pro to cols for the pa -

ren tal cell line have been pub lished [Plea sure

et al. 1992], but the cybrids may re quire

astrocytic feader lay ers and are ex pected to

of fer only low cell yield. More over, the re -

quired long-term RA treat ment seemed to be

rather toxic by it self in our hands. The

mtDNA copy num bers, de ter mined by Real -

time PCR in the pres ent study, were com pa ra -

ble to those found in the same two cell lines

ear lier, us ing other mi to chon drial and nu clear

amplicons [Wong et al. 2002], but did not re -

act to EMB in both cell lines.

If OXPHOS plays no role, other mecha -

nisms may be in volved, which could also bring

mi to chon dria onto the scene, since these

organelles con trol the in trin sic apoptotic path -

way. One of these mech a nisms is glu ta mate

excitotoxicity. EMB seems to re quire glu ta -

mate to reach its maximal tox ic ity in RGC

cul tures, which could be di min ished by glu -

ta mate an tag o nists [Heng et al. 1999]. Fur -

ther more, the an ti bi otic en hances the cyto so -

lic and the mi to chon drial cal cium load [Heng

et al. 1999], which is in con cor dance with the

sug gested mech a nism of glu ta mate excito -

toxicity, in clud ing mi to chon drial cal cium

over load, per me abil ity tran si tion and apopto -

tic neuronal death. On the other hand, LHON

mu ta tions can be sug gested to faciliate glu ta -

mate excitotoxi city. The pro posed en ergy cri -

sis [Carelli et al. 2004a, Zanna et al. 2003]

would di min ish the avail able cy to plas mic

ATP, which is used by the Na/K-ATPase to

gen er ate the plasma mem branes rest ing po -

ten tial. This would faciliate excitation of

NMDA re cep tors and cy to plas mic cal cium

in crease, which has to be buf fered par tially by

the mitochonrial cal cium store. The mu tant

mi to chon dria also does not buffer cal cium as

well as com pared to the wild type [Haroon et

al. 2007]. In ad di tion, LHON mu ta tions may

ham per ret i nal glu ta mate clearance by Mul ler

cells, since the mu ta tions di min ish glu ta mate

trans port ac tiv ity in cell cul tures [Beretta et

al. 2004]. It should, thus, be in ter est ing to

ana lyse a pos si ble syn er gism be tween EMB,

cal cium and glu ta mate in a cybrid model.

Conclusion

No syn er gism be tween pri mary LHON

mu ta tions and EMB with re spect to tox ic ity,

mor pho log i cal al ter ations, cel lu lar ATP con -

tent and mtDNA copy num ber could be de -

tected in a teratoma-based LHON cybrid

model. Al though EMB-in duced ac cu mu la -

tion of cy to plas mic vac u oles was com pa ra ble

to ear lier ob ser va tions in RGC cul tures, it

can not be ruled out, that the NT2/D1 model

may not be ideal to mea sure EMB in duced

mi to chon drial dys func tion, due to a mainly

glycolytic en ergy me tab o lism of cybrids. At

least it can be stated that EMB is not likely to

in duce mi to chon drial dys func tion in a way

which is in de pend ent of ETC us age. More -

over, it is un likely to dis turb mi to chon drial

func tion and main te nance by in hib it ing

mtDNA rep li ca tion.

Acknowledgment

We thank Prof. Gino Cortopassi and Prof.

Al ice Wong for kindly pro vid ing the LHON

cybrid clone and grate fully ac knowl edge the

ex cel lent technical as sis tance of Mrs Ines

Schellhase. We ac knowl edge fi nan cial sup -

port by the NBL grant of the Ger man Min is try

for Ed u ca tion and Sci ence (grant number:

01ZZ0407).

References

Baracca A, Solaini G, Sgarbi G, Lenaz G, Baruzzi A,

Schapira AHV, Martinuzzi A, Carelli V. Se vere im pair -

ment of com plex I-driven adenosine triphosphate syn -

the sis in Leber he red i tary op tic neu rop a thy cybrids.

Arch Neurol. 2005; 62: 730-736.

Beretta S, Mattavelli L, Sala G, Tremolizzo L, Schapira

AHV, Martinuzzi A, Carelli V, Ferrarese C. Leber he -

red i tary op tic neu rop a thy mtDNA mu ta tions dis rupt

glu ta mate trans port in cybrid cell lines. Brain. 2004;

127: 2183-2192.

Brown MD, Trounce IA, Jun AS, Akken JC, Wallace DC.

Func tional anal y sis of lymphoblasts and cybrid mi to -

chon dria con tain ing the 3460, 11778 or 14484 Leber’s

he red i tary op tic neu rop a thy mi to chon drial DNA mu -

ta tions. J Biol Chem. 2000; 275: 39831-39836.

Carelli V, Ghelli A, Bucchi L, Montagna P, De Negri A,

Leuzzi V, Carducci C, Lenaz G, Lugaresi E, Degli

Eposti M. Bio chem i cal fea tures of mtDNA 14484

(ND6/M64V) point mu ta tion as so ci ated with Leber’s

he red i tary op tic neu rop a thy. Ann Neurol. 1999; 45:

320-328.

Carelli V, Rugolo M, Sgarbi G, Ghelli A, Zanna C,

Baracca A, Lenaz G, Napoli E, Martinuzzi A, Solaini

G. Bioenergetics shapes cel lu lar death path ways in

Leber’s he red i tary op tic neu rop a thy: a model of mi to -

chon drial neurodegeneration. Biochem Biophys Acta.

2004a; 1658: 172-179.

Ethambutol tox ic ity in LHON cybrids 9

NP - 7079/ 18.06.2008

Carelli V, Ross-Cisneros FN, Sadun AA. Mi to chon drial

dys func tion as a cause of op tic neuropathies. Prog Re -

tin Eye Res. 2004b; 23: 53-89.

Floreani M, Napoli E, Martinuzzi A, Pantano G, De Riva

V, Trevisan R, Bisetto E, Valente L, Carellio V,

Dabbeni-Sala F. An ti ox i dant de fence in cybrids har -

bor ing mtDNA mu ta tions as so ci ated with Leber’s he -

red i tary op tic neu rop a thy. FEBS J. 2005; 272: 1124-

1135.

Fraunfelder FW, Sadun AA, Wood T. Up date of ethambu -

tol op tic neu rop a thy. Exp Opin Drug Saf. 2006; 5:

615-618.

von Graefe A. Ein ungewöhnlicher Fall von hereditärer

Amaurose. Arch Ophthalmol. 1858; 266-268.

Haroon MF, Fatima A, Scholer S, Gieseler A, Horn TF,

Kirches E, Wolf G, Kreutzmann P. Minocycline, a pos -

si ble neuroprotective agent in Leber’s he red i tary op tic

neu rop a thy (LHON): stud ies of cybrid cells bear ing

11778 mu ta tion. Neurobiol Dis. 2007; 28: 237-250.

Heng JE, Vorwerk CK, Lessell E, Zurakovsky D, Levin LA,

Dreyer EB. Ethambutol is toxic to ret i nal gan glion cells

via an excitotoxic path way. IOVS. 1999; 40: 190-196.

Hud son G, Keers S, Yu Wai Man P, Griffiths P, Huoponen

K, Savontaus ML, Nikoskelainen E, Zeviani M, Car ra -

ra F, Horvath R, Karcagi V, Spruijt L, de Coo IF,

Smeets HJ, Chinnery PF. Iden ti fi ca tion of an X-chro -

mo somal lo cus and haplotype mod u lat ing the phe no -

type of a mi to chon drial DNA dis or der. Am J Hum

Genet. 2005; 77: 1086-1091.

Hwang JM, Kim J, Park SS. Leber he red i tary op tic neu -

rop a thy mu ta tions in ethambutol op tic neu rop a thy. J

Neurol. 2003; 250: 87-89.

Ikeda A, Ikeda T, Ikeda N, Kawakami Y, Mimura O.

Leber’s he red i tary op tic neu rop a thy pre cip i tated by

ethambutol. Jpn J Ophthalmol. 2006; 50: 280-283.

Leber T. Über hereditäre und con gen i tal angelegte Seh -

ner venleiden. Arch Ophthalmol. 1871; 17: 249- 291.

Lee CS, Gambertoglio JG, Brater DC, Benet LZ. Ki net ics

of oral ethambutol in the nor mal sub ject. Clin

Pharmacol Ther. 1977; 22: 615-621.

Mackey DA, But tery RG. Leber he red i tary op tic neu rop a -

thy in Aus tra lia. Aust N Z J Ophthalmol. 1992; 20:

177-184.

Majander A, Finel M, Savontaus ML, Nikoskelainen E,

Wikstrom M. Cat a lytic ac tiv ity of com plex I in cell

lines that pos sess re place ment mu ta tions in the ND

genes in Leber’s he red i tary op tic neu rop a thy. Eur J

Biochem. 1996; 239: 201-207.

Man PY, Griffiths PG, Brown DT, Howell N, Turnbull

DM, Vhinnery PF. The ep i de mi ol ogy of Leber he red i -

tary op tic neu rop a thy in the North East of Eng land.

Am J Hum Genet. 2003; 72: 333-339.

Peets EA, Sweeny WM, Place VA, Buyske DA. The ab -

sorp tion, ex cre tion and met a bolic fate of ethambutol

in man. Am Rev Respir Dis. 1965; 91: 51-58.

Plea sure SJ, Page C, Lee VM. Pure, postmitotic, po lar -

ized hu man neu rons de rived from NTera 2 cells pro -

vide a sys tem for ex press ing exogeneous pro teins in

ter mi nally dif fer en ti ated neu rons. J Neurosci. 1992;

12: 1802-1815.

Sadun AA, Carelli V, Salomao SR, Berezosky A, Quiros P,

Sadun F, DeNegri AM, Andrade R, Schein S, Belfort R.

A very large Bra zil ian ped i gree with 11778 Leber’s

he red i tary op tic neu rop a thy. Trans Am Ophthalmol

Soc. 2002: 169-178.

Sadun AA, Carelli V, Salomao SR, Berezovsky A, Quiros

PA, Sadun F, DeNegri AM, Andrade R, Moraes M,

Passos A, Kjaer P, Pereira J, Valentino ML, Schein S,

Belfort R. Ex ten sive in ves ti ga tion of a large Brasilian

ped i gree pf 11778/haplogroup J Leber he red i tary op -

tic neu rop a thy. Am J Ophthalmol. 2003; 136: 231-

238.

Sadun F, DeNegri AM, Carelli V, Salomao SR, Bere -

zovsky A, Andrade R, Moraes M, Passos A, Belfort R,

da Rosa AB, Quiros P, Sadun AA. Ophthalmologic

find ings in a large ped i gree of 11778/Haplogroup J

Leber he red i tary op tic neu rop a thy. Am J Ophthalmol.

2004: 271-277.

Schoeler S, Winkler-Stuck K, Szibor R, Haroon MF,

Gellerich FN, Chamaon K, Mawrin C, Kirches E.

Glutathione de ple tion in an ti ox i dant de fense of dif fer -

en ti ated NT2-LHON cybrids. Neurobiol Dis. 2007;

25: 536-544.

Wallace DC, Singh G, Lott MT, Hodge JA, Schurr TG,

Lezza AM, Nikoskelainen EK. Mi to chon drial DNA

mu ta tion as so ci ated with Leber’s he red i tary op tic neu -

rop a thy. Sci ence. 1988; 242: 1427-1430.

Wong A, Cav a lier L, Col lins-Schramm HE, Seldin MF,

McGrogan M, Savontaus ML, Cortopassi GA. Dif fer -

en ti a tion-spe cific ef fects of LHON mu ta tions in tro -

duced into neuronal NT2 cells. Hum Mol Genet. 2002;

11: 431-438.

www.ncbi.nlm.nih.gov/omim nnn

Yoon YH, Jung KH, Sadun AA, Shin HC, Koh JY. Etham -

butol-induced vacuolar changes and neuronal loss in

rat gan glion cell cul ture: me di a tion by endo geneous

zinc. Toxicol Appl Pharmacol. 2000; 162: 107-114.

Zanna C. Ghelli A, Porcelli AM, Carelli V, Martinuzzi A,

Rugolo M. Apoptotic cell death of cybrid cells bear ing

Leber’s he red i tary op tic neu rop a thy mu ta tions is

caspase-in de pend ent. Ann N Y Acad Sci. 2003; 1010:

213-217.

Zanna C, Ghelli A, Porcelli AM, Martinuzzi A, Carelli V,

Rugolo M. Caspase-in de pend ent death of Leber’s he -

red i tary op tic neu rop a thy cybrids is driven by en er -

getic fail ure and me di ated by AIF and endonuclease G.

Apoptosis. 2005; 10: 997-1007.

Pommer, Schoeler, Mawrin et al. 10

NP - 7127/ 18.06.2008