Thiamine Deficiency--Induced Partial Necrosis and Mitochondrial Uncoupling in Neuroblastoma Cells Are Rapidly Reversed by Addition of Thiamine

Journal o.fNeurochernistry

Lί ppί n cott-R aven Pu blis h e r s, Ph iladelph ia 1995

10 International SocietyforNeuroch emistry

Th

ia

min

e Deficie

n

c

y

-

Indu

ce

d P

a

r

tia

l N

ec

r

os

i

s a

nd

M

itoc

h

o

ndri

a

l Un

co

uplin

g i

n N

e

ur

ob

l

asto

m

a Ce

ll

s

Ar

e

R

a

p

i

dly R

e

v

e

r

se

d

b

y Add

it

i

o

n

o

f Th

ia

min

e

L.

B

ettendodf,

*

F

.

S

luse, t G. Goesse n s, Ρ.

W

ins, and Τ.

G

risar Laboratories of Neurochemistry and ϊCellularand Tissular Biology, University (Y'Liège,

L

iège;

and

*

Laboratory ι#'Bioenergetics, University ofLiège, Sart-Tilman, Belgium

Abstract : Culture of neuroblastoma cells in α med ium of low-thiamine concentration (6 ηΜ) and in the presence of the transport inhibitor am prolium leads to the ap pear-ance of overt signs of necrosis; i.e., the chromatin con-denses in dark patches, the oxygen consumption de-creases, mitochondria are uncoupled, and

t

heir cristae are disorganized . Glutamate formed from glutamine is no longer oxidized and accumulates, suggesting that the thiamine diphosphate-dependent α-ketoglutarate deh y-d rogenase activity is impaired. When thiamine (10 μΜ) is added to the cells, the ΟΖ consumption increases, res-piratory control is restored , and normal cell and mito-chondrίal morphology is recovered within 1 h.Succinate, which is oxidized via the thiamine diphosph ate-indepen-dent succinate dehydrogenase, is also able to restore α normal ΟΖ consumption (with respiratory control) in dig-itonin- permeabilized thiamine-deficient cells. Our results therefore suggest that the slowing of the citric acid cycle is the main cause of the biochemical lesion induced by thiamine deficiency as observed inWernίcke's enceph a-lopathy. Key Words:Thiamine deficiency-

Necrosis-Neuroblastoma cells- Mitochond ria-Wernicke's en-cephalopathy.

J.

Neurochem. 65, 2178-2184 (1995) .

Thiamine diphosphate (TDP), which is the main

thiamine compound in the mammalian brain, is α r e-quired cofactor for pyruvate and u-ketoglutarate dehy-drogenases. Therefore, thiamine deficiencyshould lead to α decreased turnover

o

f the intermediates of the citric acid cycle and, thus, slow down oxidative metab-olism (for review, see Butterworth , 1993) . The

W

er-nicke-Korsakoff syndrome is α common consequence of thiamine deficiency linked to alcohol abuse in hu-mans (Katzman and Terry, 1983) .

W

ernicke's enceph-alopathy is characteri zed by ophthalmoplegia, ataxia, global confusional state, and delusions (Victor et α1 ., 1989) . Polyneuropathies are also observed and, at α later stage, there appears α dementia characterized by anterograde amnesia and disorientation ( Korsakoff 's

2178

psychosis ;

M

cEntee and Mair, 1990) . Although sev-eral symptoms of

W

ernicke's encephalopathy (oph-thalmoplegia) are al most fully reversible after adminis-tration of thiamine, others (ataxia) are only partially reversible and the

K

orsakoff psychosis generally does not respond to thiamine . These thiamine-nonrespon-sive symptoms probably involve irreversibleneuronal lesions . The reasons for theneuronal loss observed in severely thiamine-deficient (TD) brai n are probably multiple and involve impaired

e

nergy metabolism (Ai-kawa et α1 ., 1984), acidosis (Hakim and Pappius, 1983), and excitotoxic phenomena ( Langlais and Μαίr, 1990 ; Hazell et α1 ., 1993) .

Since the early work of Peters (for review, seeP e-ters, 1936), it is known that the early symptoms of thiamine deficie ncy in birds are rapidly reversible on thiamine administration.

M

oreover, the oxygen con-sumption was lower in minced brains prepared from avitaminous compared with controlpigeons and it was partially restored on addition ofthiamine in vitro. Oxi-dative decarboxy lation

o

f pyruvate and α-ketoglutarate was decreased in isolated mitochondria prepared from

the brains of pyrίthiam ine-treated rats compared with control animals (Gubler, 1961 ; B ennett et α1., 1966) .

N

ormal activity could berapidly restored after addition of TDP to the mitochondrial preparation (Gubler, 1961) . Parker et α1. ( 1984) reported similar observa-tions in mitochondria isolated

f

rom TDrat brain. State 3 respiration as well as respiratory control was de-creased with substrates such as glutaιnate, α-ketogluta-rate, or

c

itrate, butnot withsuccinate, which is metabo-lized independently of TDP-dependent enzymes. Par-ReceivedFebruary 6, 1995; revised manuscript received May 15, 1995 ; accepted May 15, 1995.

Address correspondence and reprintrequests toDr.L.Bettendorff at University of Liege, Laboratory of Neurnchemί stry, 17place De-cour, Β-4020 Liege,Belgium.

Abbreviations used: CCCP, carbonyl cyanideιη -chlornphenylh y-drazone; TD, thiaιnine-deficient ; TDA, thiamίne-deficient and treated withamprnlium; TDP, thiamine dίphosphate.

THIAMINE DEFICIENCY AND BIOCHEMICAL

L

ESION ker et α1 . ( 1984) considered

t

hat this situation could

lead to abnormal glutamate levels, which might favor the appearance of excitotoxic lesions.

In cultured neuroblastoma cells, α

r

eduction of the thiamine concentration in the extracellular τnedium leads, within

d

ays, to an important loss of intracell ular

thiamine compounds, but the cells

k

eep α

n

early ηοr-ιηαΙ oxygen consumption ( Bettendorff et α1., 1995) . A ddition of amproli um (α

t

hiamine transport inhibitor ) to the culture

m

edium leads to α further

d

ecrease in intracell ular

t

hiaιnine, decreased respiration, mito-chοndrial swelling and uncoupling, lactate production, and cell death . The aim

o

f

t

he present study was to investigate the reversibility of those changes after resti-tution of thiamine to the culture medium.

M

ATERIALS AND METHODS

Chemicals

Thiamine, carbonyl cyanide

m

-chlorophenylhydιazone (CCCP), amprolium,

r

otenone, digitonin, and oligomycin

were purchased from Sigma. Cell culture

Νeιιrοb1 αstπniα cells were cultured as previously de-scribed ( Beuendorff and Wins, 1994) in

D

ιιlbecco's modi-Γied Eagle's medium (GIBCO, Ghent,

B

elgium) containing ΙΟ μΜthiamine and supplemented with 5%ο fetal calf serum (GIBCO) . TD cells were produced by growing them for at least 2 weeks in α

D

ulbecco's modified medium devoid of thiamine . Under these conditions, the only thiamine source was the fetal calf serum and its concentration in the medi um was -7 ηΜ. To further increase thiamine deprivation, am-ρrπΙίιιm (20 μΜ) was added to the Culture medium of TD cells 4 days before the experiment ( "TDA cells" ), as

de-by

B

ettendorff et α1. (1995) .

Thiamine derivatives were determi ned by an HPLC proce-dure exactly as previously described (Bettendorff et α1., 1991 ) . Protein concentrations were determined by the method of Peterson ( 1977) .

Amino acids were extracted

f

romthe cells as

d

escribed by Patel and Hunt ( 1985) and determined by HPLC ( Tapuhi et al ., 1981 ) .

Oxygen consumption

Oxygen uptake was measured polarographicαlly in α 2-ml cc[] at 37°C as described by Vayssière et al. ( 1986), con-taining - ΙΟ-20 Χ 10` cells in their respective culture me-dia. Digitnnin pernιeabilization was

p

erformed as described

by Vercesi ct α1 . ( 1991 ) .

E

lectron microscopy

M

onolayer cell cultures were scraped offthe dishes and centrifuged at 350 g for 3 min. Small fragments of the pellet were fixed at 4°C in glutαraldehyde (2 .5% in cacodylate hυ ffer) and then pοstfixedin 1 %ο osmium tetrnxide solution. The cells were included in Εροη .

U

ltrathin sections mounted on copper grids were stained with uranyl acetate and lead citrate before examination under αJeol CX 100 11 electron

microscope at 60 kV .

RESULTS AND DISCUSSION

Figure 1 shows the

r

ate of oxygen consumption

b

y neιιroblasto ιna cells under different experimental

con-2179

FIG. 1. Polarographic recording of oxygen consumption by neu-roblastoma cells in whole culture media containing various amounts of thiamine (Α), TD medium (Β), TDA medium (C), and TDA cells after addition of 10 μΜ thiamine for 60 min (D) . In the case of TDA cells, amprolium (20 μΜ) was added to the culture days before the experiment . The arrows correspond to the addi-tions of various compounds: 1, oligomycin (16 μg/ml) ; 2, CCCP

(5 μΜ) ; 3, KCN (1 mM).

d

itions. The comτnercial

m

edium alway s contained 30 mM gl ucose, 1 ιηΜ pyruvate, and 4 ιηΜ glutamine. 1η each case, we first

m

easured

t

he

b

asal oxygen con-sumption . Then, the rate of Ο, concon-sumption was esti-mated i n the presence of added oligomycin, followed by the

u

ncoupler CCCP. In the case of cells grown at α high thiamine concentration ( ΙΟ μ,M), a reasonably

h

igh 02 consumption was

m

easured. As expected, oli-gomycin decreased oxygen consumption, whereas in the presence of CCCP it was increased about threefold above

b

asal level . Oxygen consumption was com-pletely inhibited after

a

ddition of 1 mM

K

CN. Α simi-lar pattern was observed with cells grown at α lower thiamine concentration (7 ηΜ ) . Under

t

hose condi-tions, total intracellular thiamine content was

d

e-creased from 210 to 13 ρmο1 /mg of protein ( Betten-dorff et α1., 1995) . To further decrease intracellular thiamine concentration, amprolίum (20 μM), α com-petitive inhibitor of

t

hiaιnine transport, was added to the TD culture

m

edi um ( TDA) for 4

d

ays. Pyrithi-amίne is the

m

ost

p

otent

k

nown th iamine antίmetabo-lite in vivo;

b

ut in this study we

p

referred to use am-ρrοΙίυm, though -30-fold higher

a

mounts of th is com-pound are

r

equired to block thiamine transport ( Bettendorff and Wi ns, 1994) . In contrast to

p

yrίthia-mine, amprolium cannot be phosphoryl αted. Thus, no i nterference with TDP-dependent enzy mes is to

b

e ex-pected with

a

mprolium.

U

nder these conditions, the

b

asal

r

espi ration was considerably

d

ecreased compared with the untreated cells . Furthermore, oligomycin and CCCP were without effect on

r

espiration , suggesting uncoupling of mitochondria.

T

hat

K

CN (as well as roteno ne; not shown ) still inhibited oxygen consump-tion

s

uggests that, even under these severe

T

D condi-tions, part of

t

he

r

espiration persists and the most con-spicuous effect of thiamine

d

eficiency is the apparent

m

itochondrί al

u

ncoupling . When thiamine was added to the TDA cells 60

m

in

b

efore the

p

olarographic deter-.1.Νιυrηι hon- V ιι1. 05, No. 5. 1995

2180

FIG. 2. Time-dependent recovery of oxygen consumption in TDA cells after addition of 10 μΜ thiamine. Basal oxygen con-sumption ( ι ) ; oxygen concon-sumption in the presence of oligomy-cin (Ο) or CCCP (Ο) . Each point represents the mean ± SD

value for th ree experiments except for the points in the presence of oligomycin, which are the mean values of two experiments.

m

ination

o

f oxygen consumption, α nearly complete

r

ecovery of mitochondrial function was observed; i.e., the basal oxygen consumption increased

a

nd

t

he usual effects of oligomycin

a

nd CCCP

r

eturned, suggesting α recoupling of the

m

itochondria.

F

igure 2 shows the time scale of the

r

ecovery of mitochondrί al

r

espiration. It can be seen that, already 5 min

a

fter

a

ddition of thiamine, α significant increase in the

b

asal and uncoupled

r

espiration was observed and

t

he effect was complete after 60

m

in. The

r

espira-tion in the presence of oligomycin

r

emained

un-changed.

Thiamine by itself

h

as no known

e

ffect on

m

itochon-dria)

r

espiration and

t

he important compound is the cofactor TDP. Thiami ne is,

h

owever, actively trans-ported into

n

euroblastoma cells and pyrophosphory-lated in the cytoplasm ( Bettendorff and Wins, 1994) . TDP is then transported into mitochondria ( Barile et al ., 1990) where it binds to the pyruvate and α-ketoglu-tarate dehydrogenases. Α

n

early 10-fold increase in intracellular TDP is observed within 1 h after

a

ddition of thiamine to the cells (Fig. 3) . The lag

p

eriod may

b

e explained by the fact that thiamine pyrophosphoki-nase is dependent on the intracellular ATP concentra-tion with an excepconcentra-tionally

h

igh Κ~, of 7 mM ( Betten-dorff and Win s, 1994) . In TDA cells, intracellular ATP concentrations are lowered by -50%ο compared with normal cells ( Bettendorff et α1., 1995) . Thus, the

r

ate of thiamine phosphorylation is slow

b

ut increases as mitochondria are producing ATP.

We

h

ave previously reported the existence of abnor-mal

m

itochondria in TDA cells (Bettendorff et α1 ., 1995) . With

o

ngoing thiamine

d

eficiency, the mito-chondria) matrix became disorgani zed and

e

lectron

t

ranslucent; no intact cristae

r

emai ned visible and some mitochondria

b

ecame abnormally large.

H

owever, we

d

i d not observe any significant increase in the number of mitochondria in

T

DA cells compared

w

ith control cells .

M

itochondria)

r

espiration in

T

DA cells was es-sentially uncoupled

a

nd, as α preincubation with

t

hia-mine leads to α

r

ecoupling of respiration, we wanted to / . Νe ιι rιιι h e ιιτ ., Vπ /. 65, No. 5, 1995

L.

BE

ΤT

E

NDOR

FF

ET

AL.

know whether

t

his was accompanied

b

y morphological modifications . Figure 4α and b

s

how, with different

m

agnifications, α

t

ypical TDA cell with largely

e

lec-tron-translucent mitochondria.

T

he nucleus shows evi-dent signs of chromatin condensation

t

ypical of

t

he early abnormalities of necrosis (Wyllie et al ., 1980) . After 1 h in

t

he presence of 10 μΜ thiamine these abnormalities were essentially reversed ( Fig. 4c and d) ; the

m

itochondria became electron dense and

c

ristae were

r

eformed. The

m

itochondria

r

esembled those of control cells ( Bettendorff et al., 1995) . Note that the cell cycle in neuroblastoma cel ls lasts -υ 24 h

a

nd, thus, normalization in cell morphology, within Ι

h,

cannot be explained

b

y the generation

o

f new cell s through mitotic

d

ivision.

The

r

easons

r

espiratory control is lost in TDA cells are not clear. The link between uncoupling and the di sorganization

o

f

c

ristae may

a

ppear obv ious, as the electrochemical proton gradient tends to dissipate when the inner membranes are

d

amaged;

h

owever, the

r

easons the cristae

b

ecome

d

isorganized are unclear. We

m

ay consider

t

he possibility that the phenomenon is

r

elated to the lack

o

f

s

ubstrates able to donate elec-trons to the

r

espiratory chain. If this

h

ypothesis is true, the direct addition of α permeant substrate should

r

e-verse

u

ncoupling in α

m

anner

a

nalogous to thiamine addition,

m

aybe even faster. As shown in

F

ig. 5, the addition

o

f

s

uccinate to cells

p

ermeabili zed with dig-itonin

i

ndeed restored coupled

r

espiration . After

a

dd i-tion of digitonin to

T

DA cells, the oxygen consumption

g

radually

d

ecreased, as α

r

esult of dilution of the

r

e-maining substrates . Addition of succinate increased the

r

espiration up to sixfold

i

n some experiments. This ΟΖ consumption was slightly

i

nhibited

b

y oligomycin and increased

b

y CCCP as expected. The

u

ncoupled

r

espi-ratory control, i .e., the

r

atio of Ο, con sumption in

t

he

p

resence of CCCP to the Ο, consumption in the pres-ence of

o

ligomycin, was 1 .8 -+- 0.5 . This uncoupled

r

espiratory control value was not significantly

h

igher when normal (instead of TDA) cells were

u

sed.

R

ote-none, an inhibitor of respiratory chain complex Ι, was without effect on ΟΖ consumption after succinate addi-tion, but antimycin, an inhibitor

o

f complex III, nearly completely inhibited oxygen consumption. This

sug-FIG. 3. Time-dependent recovery of intracellular TDP content in TDA cells after addition of 10 μΜthiamine.Each point represents

F

IG. 4. Electron micrographs of TDA cells (α, b)and TDA cells exposed to 10 IVthia-mine for 1 h (ε, d) . Magnification bars: 1 μm in (α) and (c) ; 0.2 μm in (b) and (d) .

THIAMINE

D

EFICIENCY AND BIOCHEMICAL LESION

gests that, under

t

hese conditions, the respiration is indeed sustained by succinate. In

T

DA cells, at least some of

t

he

m

itochondria

r

emain

f

unctional, and

t

he apparent uncoupling ( Fig. 1C) appears to

b

e linked to t he lack of oxidί zable substrates. A direct consequence of th is lack of

r

educing substrates is that at least com-plexes III and IV will

r

emain completel y oxidized for α long time.

T

his may somehow lead to disorganization of cristae,

b

ut the

m

olecular mechanisms involved, if any, remain unknown.

M

itochondrial swelling with rupture of cristae

h

as long

b

een

k

nown to occur after

2181

treatment with ascorbate,

f

errous ions ( Hunter et α1 ., 1963 ), or

a

fter treatments that

f

avor lipid peroxidation (Shigenaga et α1 ., Ι 994) . Swelling and inner mem-brane damage were also observed after treatment of mitochondria by thiol-blocking alkylati ng agents ( Lê Qunc and

L

ê Qu6c, 1985), glutathione deficiency (.lain et α1., 1991), and deficiency in an enzyme in-volved in cαrdίο1ίρίη synthesis (Ohtsuka et α1 ., 1993) . 1 η all cases described so far,

h

owever, there is no evi-dence that the respiratory chain is involved .

As thiamine

d

eficiency

h

as

b

een reported to affect

2182

FIG. 5. Respiration in digitonin-permeabilized TDA cells. The cells were sedimentedandthe culture medium replaced by the test medium containing 125 mM sucrose, 65 mM KCI, 10 mM Tris-HCI (pH 7.2), 1 mM MgC12 , 0.33 mM EGTA, 2.5 mM ΚΗΖΡΟ4 , and 2.5 mM ADP. Trace (α) shows the oxygen con-sumption without addition. In traces (b) and (c), 50μΜdigitonin (DIG) was added. Other additions (trace b) are 16 μ,g/ml oligo-mycin (Ο), 50 ηΜ CCCP, 2MM rotenone (ROT), and 2 μ.g/ml antimycin (Α) .

amino acid metabolism in rat brain (Butterworth and H6roux, 1989 ;

P

age et al., 1989; Butterworth, 1993), we

d

etermined the concentrations of

i

ntracellular

amino acids in our cells (Table 1) . In control cells, the amino acid concentrations were close to those pre-viously

r

eported in cultured

n

eurons and astrocytes (Patel and Hunt, 1985) . As expected, glutamate was the

m

ost abundant amino acid.

W

e were

u

nable to detect any GABA in our cells, an observation that is important for the interpretation of our results as this means that they lack glutamate

d

ecarboxγlase, an en

-zyme specific

f

orGABAergie

n

eurons (Martin, 1986) .

This implies that, in our cells, no succinate can

b

e formedthrough the GABA shuntwhen α-ketoglutarate dehydrogenase is inactive due to lack

_o

f TDP. Th

ia-mine

d

eficiency increases the

m

etabolic flux through

this

p

athway (Page et α1., 1989) and it can thus be expected that cells with α functionally intact GABA shunt would be

m

ore

r

esistant to thiamine deficiency thancells that do not

p

ossess this

p

athway.

H

owever, it

h

as been shown that glutamate decarboxylase activ -ity and GABA levels

r

eversibly decrease in the brain

of

T

D

r

ats (H6roux and Butterworth, 1988),

w

hich

may cause disturbances in GABAergic neurotransmis-sion and add to the

r

eversible symptoms caused by

t

hiamine deficiency in brain.

Aspartate, glycine, and alanine levels were sign ifi-cantly increased in TD compared with control cells. That α1αηίηe is increasedsuggests that

p

yruvate deh

y-drogenase is already partially inhibited in these cells; i.e ., pyruvate, instead ofentering the citric acidcycle, is

p

artially transaminated to alanine. Additionof am

-prolium leads to α large increase in intracellular

gluta-mine and glutamate concentrations. In

n

eurons, α-k e-tnglutarate andglutamate

r

apidly and reversibly equ ili-brate in

p

arallel withoxaloacetate and aspartate, as α !.Νeυrπιheιπ., Vol. 65, No. 5, /995

L.

B

ETTENDORFF ET AL.

r

esult of very fast transamination (Erecifiska et α1., 1993) . In braincells, glutamate

d

ehydrogenase activity is probably less important than transamination in the

f

ormation

o

fα-ketoglutaratefrom glutamate (McCar -thy and

T

ipton, 1983) . Indeed, glutamate

d

ehydr

oge-nase is

m

ore likely to catalyze the

r

everse

r

eaction, especiallyin thepresence of high _{NH4 ι co}ncentrations. In intact brain, glucose ispracticallythe only substrate crossing the blood-brainbarrier and,

h

ence, glutamate is essentially formed from α-ketoglutarate

d

erived from glucose. This

m

ight explain thatin the

T

D brain

g

lutamate levels are

d

ecreased (Plaitakis et al., 1979;

Butterworthand H6roux, 1989) . In ourcells,

h

owever, glutamine is taken up from the culture medium and is thedirect source of glutamate. In TDA cells, where

α-ketoglutarate

d

ehydrogenase is probably strongly in-hibited, α-ketoglutarate will accumulate (especially in the absence of the GABA shunt) and the glutamate

f

ormed from glutamine is no longeroxidized via the citric acid circle.There is no

p

arallel increasein aspar -tate concentration presumably because no oxaloacetate can be

f

ormed.

Addition ofthiamine to the cells leads, within 1

h,

to α significant

d

ecrease in intracellular glutamate level. This would be in agreement with α

r

ecovery

o

f

α-ketoglutarate dehydrogenase activity and oxidative

m

etabolism. The respiration in thepresence of thi

a-mine is

n

early completely blocked

b

y rotenone (an inhibitor

o

f NADH dehydrogenase) in agreement with this observation.

One ofthe

m

ostpuzzling observations in thiamine deficiency, in animal

m

odels as well as in human pa-thology, is that the early symptoms are so

r

apidly r e-versed on thiamine administration (Butterworth, 1993) . This is the "biochemical lesion"

d

escribed by

P

eters (1936) . Prolonged thiamine

d

eficiency,

h

ow-ever, leads to irreversible

h

istological lesions with ne u-ronal death. In this and

o

ur previous work (Bettendorff

et α1., 1995), we show that, in cultured

n

eιιroblastoma cells, severe thiamine

d

eficiencyleads to α decreasein

the

r

ate of respiration, to

m

itochondrial uncoupling,

TABLE 1 . Amino acid contents of ηcυrπh/αsυηηα cells

under various conditions of thiamine deficiency

Amino acids were determined by HPLC according to the method

ofTapuhi et α1. (1981). The last group

r

epresents TDA cells incu-bated in the presence of thiamine ( ΙΟμΜ)for Ι h. In each case, the culture medium contained 0.4 mΜ glycine and 4mM glutamίne.

M

ean - SD values

f

or three experiments. ηιηο1/ιηg of protein Gin Asp Gin Gly Αlα Control 4.2 0.9 5.2 ± 0.4 65 ± 8 43 ± 5 23 ± 1 TD 3.4±2.0 Ι 2±3 56±13 66±2 41±4 TDA 21 ±8 9.4± 1 165±42 83±7 22±3 ΤDλ + thiamine 58 ± 13 10 ± 3 118 ± 15 6 1 ± 1 22 ± 3

REFERENCES

THIAMINE DEFICIENCY AND BIOCHEMICA L LESION and to morphological abnorm alities corresponding to

the early sympto ms of nec rosi s (Wyllie et α1., 1980) . ATP concentrations aredec reased, lactate pr odu ction increases, the ce ll sbecomedepolarized, and cell mor-tality increases ( Bettendorff et α1., 1995) . The most remarkable pr operty shown in the pr esent st udy is that those cells that sur vive long enough un de r seve red efi-ciencyrespond rapidly to thia mine treatment; i.e ., nor-ιηαΙ respiration as we ll as norm al cell and mitochon-drial morphology are recovered within 1 h. Th eser e-sults suggest that the biochemical lesion obse rved in thiamine deficiency is theres ult of an inabili ty of the cells to oxidi ze substrates . En ergy failure lea ds to α cellular co llapse, whi ch, if not treated, res ul ts in necro-sis and cell death.

In α recentreport, Zhang et α1 . (1995)demonstrated

the existence of disintegrating mitochon dr ia and chr o-ιηαtίη clumping in degenerating neurons of dience-phalic n uclei in pyrίthiamine-treatedrats. Th eseres ults suggest that, in brain, mec hanis ms similar to those described here

f

or cult ured neur οblasto ma cells may

be operating, except that additional phenomena such as excitotoxicity mig ht contribute to the selective vul-nerability of certain brainregions in pyrithiaminerats (Langlais and Μαί r, 1990 ; Hazell et α1., 1993 ;

L

anglais and Zh ang, 1993) . So

f

ar, however, there is no evi-dence that excitotoxic phenomena are important during the ac ute, reversible ph ase of thiamine deficie ncy (Wernicke's enceph alopathy ) .

Ac kn owle dgment:We than k theBelgian Natio nal Fu nds for Scientific Research (FNR S)

f

or α grant to

L.Β.

P.W. is Research Associate at the FNRS .

Λίkτιωα Α . Η., Watanabe Ι. S., Furuse Τ., Iwasaki Υ., Satoyoshi Ε., Sιιιηί Τ., and Μπrοjίί Τ. ( 1984)Low energy levels in thiamine-deficient encephalopathy. ./ . Νeυrηηαthοί. Exp. Νeυrο1. 43,

276-287.

ΒαrίΙε Μ ., Ρassarella S., and Quagliariello Ε. (1990) Thiamine pyro-phosphate uptake into isolated rat liver ιτιitochondria. Arch. Bioc hem. Βiορ h,vs. 280, 352-357 .

Bc ιιnett C. D., Jones 1. Η., an d Nelson J. (1966) The effects of' thiamine deficiency on the metabolism of thebrain : 1. Oxidation of' various substrates in vitro by the liver and brain of normal an d pyr ίthia mίne-fed rats . J. Ne urochem. 13, 449-459. 13ettendorffL.and Wins Ρ. ( 1994)Mechanis m of thiamine transport

in ηeuroblastoma ce lls . Inhibition of α hig h affinity carrierby sodium channel activators anddependence of thiamineuptake on membrane potential and intracellular ATP. .Ι. Biol. Chem. 264, 14379- 14385 .

Iicι tcndorff L., Peeters Μ ., Jουαπ C., Wins Ρ., and Schnffenlels Ε. ( 1991 ) Determination of thia mi n and its phosphate esters in Cultured neurons and astrocytes using an ion-pair reversed phase high-performance liquid chromatog raphic method. Anal. Βίο-(hιιιι . 198, 52-59.

ΒctteηdnrffL.,Goessens G., Sl ιιse F., Wins Ρ., Bu reau Μ.,Laschet Ι ., and GrisarΤ. ( 1995 ) Thiamine deficiency in cultured neuro-hlastnma cells: effect on mitoc hondria] function andperipheral henzodia zepίne rece ptors . J. Ne uroc hem. 64, 2013-2021 .

ButterworthR. F .(1993) Pathophysiological mechanisms responsi-ble for the reversible (thia mine- res ponsive) and irreversible

2183 (thiami ne non-responsive) neurological sy mptoms of Wer-ηί ckε's encephalopathy . Dr ug AlcoholRev . 12, 315-322.

Butterworth R . Ε. an d Wrοιιχ Μ. (1989) Effect of pyrίthiami ne treatment and subsequent thiamine re habilitation on regional cerebral amino acids andthiamine- dependent enzymes . J.New

rοchem. 52, 1079- 1084 .

Erecifiska Μ., Pleas ure D., Nelson D., Nissl m Ι ., and Yudkoff Μ . (1993) Cerebral aspartateutili zation: near-equilibri um relation-ships in aspartate aminotransferasereaction. .Ι. Neurochem. 60,

1696-1706.

Goblet C. 1. (1961) Studies on the physiological functions of thia-mine, Ι.The effects of thiaminedeficiency andthiami ne antago-nists on the oxidation of α-keto acids by rat tissues. ./. Biol. Chem . 236, 3112-3120.

Haas R. Η . (1988) Thiamin and the brain. Anna. Rev. Nutr. 8, 483-515.

Ηαkίιη Α. Μ. and Ραρρίυ s Η. Μ. (1983) Sequence of metabolic, clinical andhί stnlogical events in experimenta l thiamine defi-ciency . Ann. Neurol. 13, 365-375.

Hazel] Α. S., ButterworthR. F., and Hakim Α. Μ. (1993) Cerebral vulnerability is associated with selective increase in extracellu-Iar gluta mate concentrations in experimental thiamine de fi-ciency .J. Neurochem. 61, 1155-1158 .

Héroux Β. and ButterworthR. F. (1988) Reversible alterations of cerebraly-aminobutyric acid in pyrithia mine-treated rats: impli-cations for the pathogenesis of Wern icke's encephalopathy . J. Neurochem. 51, 1221-1226.

Hunter F. Ε.Jr., Gebicki 1. Μ., Hoffsten Ρ. Ε., Wei nstei n 1 ., and Scott Α. ( 1963) Swelling and ly sis ofrat liver mitochondria induced byferrous ions . J. Biol. Chem. 238, 828-835. Jain Α., Mártensson J., Stole Ε., ΑυΙd Ρ. Α . Μ., and Meister Α.

(1991 ) Gl ιιtathione deficiency leads to mitochond ria] da mage in brain. Ρrηι. Nod. Αιαιί. Sci. USA 88, 1913-1917. Katzman R. andTerry R.D. (1983) The Neurology of Aging. F.Α .

Davis, Philadelphia .

Langlaίs Ρ. 1. andΜαίr R. G. (1990) Protective effects of the gluta-ιnate antagonists MK-801 on pyrίthiamine-induced lesions and amino acid changes in rat brain. J .Neurosci. 10, 1664-1674.

Langlaίs Ρ. 1. and Zhang S . Χ. (1993 ) Extracellular gl utamate is increased in thalamus du ring thiamine deficienc y-induced le-sions and is bloc ked by MK-801 . J. Ne υrοcheru. 61,

2175-2182,

Lê-Quóc Κ. and Lê-Quδc D. (1985) Crucial role of sulfhydry l groups in the ιnitochondrial inner membrane structure .J. Biol. Chem. 260, 7422-7428.

ΜαιτίηD. L.(1986) Brain gluta matedecarbnx ylase, in Neurotrans-mitter Enzymes (Boulton Α. Α., Baker G. Β., and Υυ Ρ. Η., eds), pp. 261-388. ΗυιηαηαPress, Clifton,NewJersey. McCarthy Α.D.and TiptonΚ.F.(1983) Gl utamate dehydrogenase,

in Glυtιιυιίυ e, Glutamate, and GA BA in the Nervous Sι ste» ι (HertzL., Kva mme Ε.,McGeer Ε. G., and Sc housboe A., Jeds), pp . 19-32. AlanR. Liss, New York.

McEntee W. .Ι. and Μαίr R. G. ( 1990) The Korsakoff syndrome: α neurnc hemίcal perspective. Trends Νeurn.sci. 13, 340-344. Ohts uka Τ., Nlshijima Μ ., Su zuki Κ., and Akamatsu Υ . (1993)

Mitochon dr ia] dysfunction of a cultured Chinesehamster ovary cell mutantdeficient in cardio lipin . J . Biol . Chem. 268, 22914-22919.

Page Μ. G., Αηkοιηα -Sey V., Cou lson W. F., and Benders D. A. (1989) Brain glutamate andy-aιninnbutyrate (GABA)m

etabo-lism in thiamine-deficient rats . Br.J .Nutr. 62, 245-253. Parker W.D. Jr., Haas R., St um pf D. Α., Parks 1., Eguren L. Α.,

andJackson C. (1984) Brain ιnitochondrial metabolis m in

ex-peri mental thiamine deficiency. Neurology 34, 1477-1481 . Patel Α. 1. and Hunt Α. ( 1985 ) Concentration of free amino acids

in primary cultures of neurones andastrnc γtεs. J. Néurochem. 44, 1816-1821.

Peters R. Α. (1936) Th e biochemical lesion in vitamin Β 1 deficiency. Application of modernbiochemical analysis in its diagnosis . Lancet 1, 1161 -1164.

2184

Peterson G.

L.

(1977) Α simplification of the protein assay method of

L

ow ry et α1 . which is more generall y applicable. Anal. Βίο-chem. 83, 346-356.

Plaίta kis Α .,NicklasW. 1 .,andBerl S. (1979) Alterations inuptake and meta bolism of as partate and gl utamate inbrainofthi amine deficient animals. Brain Res. 171, 489-502.

Shigenaga Μ . Κ., Hagen Τ. Μ., and Ames Β. Ν. (1994) Oxidative damage and mitoc hondrial decay in aging. Pro( . Natl. Acad. Scί. USA 91 , 10771 -10778 .

Tapu hi V.,

M

iller Ν., an dKarge r Β.

L.

(1981)Practical considera-tions in the chiral separation of Dn s-amino-acidsby re versed-phase li quid chro matographyusi ng metalchelate ad ditives .

J.

Chromatogr. 205, 325-337.

Vayss ίereJ.

L., L

archerJ. C., BerthelotF., Benlot C.,Gros

F.,

and Croi zat Β. (1986)

E

ffects on mitochondrial metabolism of

.1. Νeιι r-υ(heηι ., Vol. (i5, No. 5, 1995

L.

BEΤTENDORFF ET AL.

CCA, one inducer of ne uroblasto ιna differentiation. Bίochem.

Bίophys . Res. Commun. 140,789-796.

Vercesi Α. Ε., Bernardes C.

F.,

Hoffm ann Μ. Ε., GadelhaF. R .,

an d DocampoR. (1991 ) Dίgίtπηίη perιneabilizat ίon doesnot affect mitoc hondrial function and allows thedetermination of the mitoc hondrial mem brane potential of Trypanosoma ιrυzί in situ . J. Biol. Chem. 266, 14431 -14434 .

Victor Μ ., AdamsR. D., and Collins G. Η. (1989) TheWernίck

e-Korsakoff Syndrome.

F.

Α. Davis, Philadelphia.

Wγllie Α. Η.,Kerr 1 .F.R.,andCurrie Α.R.(1980) Celldeath: the significance of apoptosis. Int. Rev. Cytol. 68, 251-306. Zh ang S. Χ.,Weilersbac her G. S.,

H

ende rso n 5 .W.,Co rso Τ., Olney

J. W., and

L

anglais Ρ. J.(1995)

E

xcitotoxic cytn pathology, prog ression, and reversibility of thiamine deficiency-induced diencephalic lesions. J . Neuropathol. Exp. Νeυrοί. 54, 255-267.