Article

ISSN 0006-2979, Biochemistry (Moscow), 2024, Vol. 89, No. 10, pp. 1692-1710 © Pleiades Publishing, Ltd., 2024.

1692

Combined Administration of Metformin and Amprolium

to Rats Affects Metabolism of Free Amino Acids

in the Brain, Altering Behavior, and Heart Rate

Anastasia V. Graf

1,2

, Artem V. Artiukhov

1,3

, Olga N. Solovjeva

Alexander L. Ksenofontov

, and Victoria I.Bunik

1,3,4,a

Belozersky Institute of Physicochemical Biology, Lomonosov Moscow State University, 119234 Moscow, Russia

Faculty of Biology, Lomonosov Moscow State University, 119234 Moscow, Russia

Department of Biochemistry, Sechenov Medical University, 105043 Moscow, Russia

Faculty of Bioengineering and Bioinformatics, Lomonosov Moscow State University, 119991 Moscow, Russia

e-mail: bunik@belozersky.msu.ru

Received May 27, 2024

Revised July 30, 2024

Accepted August 6, 2024

Abstract— The risk of developing diabetes and cardiometabolic disorders is associated with increased levels

of alpha-aminoadipic acid and disturbances in the metabolism of branched-chain amino acids. The side effects

of the widely used antidiabetic drug metformin include impaired degradation of branched-chain amino acids

and inhibition of intracellular thiamin transport. These effects may be interconnected, as thiamine deficiency

impairs the functioning of thiamine diphosphate (ThDP)-dependent dehydrogenases of 2-oxo acids involved in

amino acids degradation, while diabetes is often associated with perturbed thiamine status. In this work, we

investigate the action of metformin in rats with impaired thiamine availability. The reduction in the thiamine

influx is induced by simultaneous administration of the thiamine transporters inhibitors metformin and ampro-

lium. After 24 days of combined metformin/amprolium administration, no significant changes in the total brain

levels of ThDP or activities of ThDP-dependent enzymes of central metabolism are observed, but the affinities

of transketolase and 2-oxoglutarate dehydrogenase to ThDP increase. The treatment also significantly elevates

the brain levels of free amino acids and ammonia, reduces the antioxidant defense, and alters the sympathetic/

parasympathetic regulation, which is evident from changes in the ECG and behavioral parameters. Strong pos-

itive correlations between brain ThDP levels and contents of ammonia, glutathione disulfide, alpha-aminoad-

ipate, glycine, citrulline, and ethanolamine are observed in the metformin/amprolium-treated rats, but not in

the control animals. Analysis of the obtained data points to a switch in the metabolic impact of ThDP from the

antioxidant and nitrogen-sparing in the control rats to the pro-oxidant and hyperammonemic in the metformin/

amprolium-treated rats. As a result, metformin administration along with the amprolium-reduced thiamine sup-

ply significantly perturb the metabolism of amino acids in the rat brain, altering behavioral and ECG parameters.

DOI: 10.1134/S0006297924100043

Keywords: behavior, metabolism of brain amino acids, ECG, 2-oxoglutaratre dehydrogenase, 2-oxoadipate dehy-

drogenase, pyruvate dehydrogenase, tricarboxylic acid cycle, thiamine diphosphate, transketolase, vitaminB1

* To whom correspondence should be addressed.

INTRODUCTION

Obesity and insulin resistance have long been

known to be accompanied by disturbances in the

amino acid metabolism and changes in the circulat-

ing levels of amino acids [1, 2]. In particular, diabetes

is associated with the elevated plasma content of al-

pha-aminoadipate [3]. The therapeutic potential of the

antidiabetic drug metformin can be increased by over-

coming the metformin-induced impairments in the deg-

radation of branched-chain amino acids (BCAAs) medi-

ated by the AMP-activated protein kinase (AMPK) [4, 5].

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1693

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Abbreviations: OGDC, 2-oxoglutarate dehydrogenase com-

plex; PDC, pyruvate dehydrogenase complex; ThDP, thia-

mine diphosphate; TKT,transketolase.

Amino acids are degraded through the mitochon-

drial tricarboxylic acid (TCA) cycle, mostly after

their transamination to pyruvate, 2-oxoglutarate, or

branched-chain 2-oxo acids and further oxidation

by the corresponding thiamine diphosphate (ThDP)-

dependent dehydrogenase complexes. An insufficiency

of glucose oxidation typically causes upregulation of

the amino acid degradation as an alternative energy

source.

ThDP is a diphosphorylated derivative of thiamine

(vitaminB1). Because ThDP acts as a coenzyme in the

amino acid degradation pathways, thiamine deficiency

perturbs the metabolism of amino acids. In particu-

lar, the blockade of ThDP biosynthesis and, therefore,

impaired catalysis by ThDP-dependent enzymes, are

known to cause changes in the levels of major ami-

no acids in the brain of experimental animals [6, 7].

Onthe other hand, improvement of brain metabolism

as a result of thiamine administration is accompanied

by decreased oxidation of amino acids [8].

Although it has long been believed that thiamine

deficiency is no more a problem in developed coun-

tries, the number of cases in which pathogenesis of

neurodegenerative disorders is associated with over-

looking the patient’s thiamine status is increasing [9].

Thiamine consumption is not recommended togeth-

er with administration of the antidiabetic drug met-

formin because thiamine and metformin compete for

the same intracellular transporters. Owing to this,

thiamine may decrease pharmacological effect of met-

formin. However, the opposite possibility, i.e., that the

competition of metformin and thiamine for the same

transporters can also decrease intracellular levels

of thiamine, is considered surprisingly rarely when

metformin is prescribed. This is even more surpris-

ing because the patients with type 2 diabetes mellitus

repeatedly exhibit an impaired thiamine status asso-

ciated with the reduced absorption and intracellular

transport of thiamine [10]. Nevertheless, to our best

knowledge, the association between the thiamine de-

ficiency and metformin administration has not been

studied. That said, such well-known effect of thiamine

deficiency as lactic acidosis is also a common side ef-

fect of metformin therapy. Moreover, some case stud-

ies have demonstrated that thiamine alleviates the

metformin-induced lactic acidosis [11, 12]. Thiamine

deficiency is also known to impair oxidation of the

branched-chain 2-oxoacids [6], that may underlie the

impairment of the branched-chain amino acids (BCAA)

degradation by metformin [4]. However, similar to lac-

tic acidosis, this side effect of metformin has not been

linked to the drug-induced thiamine deficiency either.

Here, we study the effects of chronic metformin

administration to rats whose thiamine status is simul-

taneously challenged by administration of amprolium.

Amprolium is a coccidiostat used in poultry produc-

tion. It blocks thiamine absorption in the intestine,

aswell as reduces its flux across the blood-brain bar-

rier and intracellular transport [13-15]. Due to these

properties, the drug is also used in basic research to

create animal models of thiamine deficiency (reviewed

in [16]). Amprolium administration can induce thia-

mine-responsive cerebrocortical necrosis and polioen-

cephalomalacia in farm and domestic animals [17-21],

thus questioning the safety of using amprolium-treated

animals for human consumption [22]. Chronic amproli-

um administration to mice induces behavioral changes

along with impaired cellular functioning [23]. There-

fore, by using a combined administration of metformin

and amprolium to rats, we model the comorbidities

typical for diabetic patients, such as side effects of

biguanide antidiabetics and undiagnosed thiamine

deficiency. Taking into account that metformin and

amprolium affect intracellular thiamine transporters

of the SLC19A and OCT families [24-26], we compare

the thiamine-dependent metabolism in the control and

treated rats via (i) assessment of the levels of ThDP (the

coenzyme form of thiamine), activities of ThDP-depen-

dent enzymes, and saturation of these enzymes with

ThDP in the brain homogenates, (ii) quantification of

amino acids and related compounds in brain extracts,

and (iii) behavioral tests and electrocardiogram (ECG)

recordings in the course of the experiment. We show

that by the end of the experiment, the control and ex-

perimental animals differ in the saturation of ThDP-de-

pendent enzymes with ThDP and in the metabolic im-

pact of ThDP and ThDP-dependent enzymes, but not in

the total levels of ThDP. The drugs increase the content

and strengthen the metabolic interactions of free ami-

no acids and ammonia in the rat cerebral cortex, thus

affecting animal behavior and ECG parameters.

MATERIALS AND METHODS

Animal experiments. All animal experiments

were carried out in accordance with the Guide for the

Care and Use of Laboratory Animals published by the

European Union Directives 86/609/EEC and 2010/63/EU).

Twenty-six Wistar male rats were randomly assigned

to the control and experimental groups. The 5 to 6

weeks old rats entered the experiment with the body

weight of 155.3 ± 2.7 and 155.4 ± 3.3 g in the control

and experimental groups, respectively. At the end of

the experiment, the 9 to 10 weeks old rats weighted

271.0 ± 5.1 and 278.0.3 ± 5.4 g in the control and ex-

perimental groups, respectively. Amprolium (40mg/kg

body weight) and metformin (200mg/kg body weight

GRAF et al.1694

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Fig. 1. The flowchart of the animal experiment. Over the period of 24 days, the drugs metformin(M) and amprolium(A) were

injected on the days indicated. The first three injections of 200mg/kg body weight metforminwere followed by 15 injections

of 70mg/kg body weight metformin. Each of the metformin injections was accompanied by the second injection of 40mg/kg

body weight amprolium. Physiological and biochemical tests were performed at the indicated days as described in Materials

and Methods.

for the first three injections and 70 mg/kg body weight

for the next fifteen injections) were administered to

rats as two separate injections (n = 14; two rats died

on day 3), while the control group (n = 12) received

two separate injections of physiological saline. The

animals were injected in the morning (ZT 2 ± 1). To-

tal 18 double injections were administered during 24

days according to the scheme shown in Fig. 1, i.e., 5

days with the injections were followed by the two

days pause until the last three days with the injections.

Thedoses were selected based on the published data

[23, 27]. Animal weight and consumption of food and

water were monitored daily. No significant differenc-

es between the experimental and control groups were

observed. The open field test (OpenScience, Moscow,

Russia) was used to assess the spontaneous activity of

animals in an unfamiliar environment [28]. The test

was conducted for three minutes in complete silence;

the area was illuminated with a 15-W red lamp as

described before [29, 30]. The following parameters

were estimated: locomotor activity (the number of line

crossings); number of entries to the central zone (the

number of movements to the central zone intersecting

the outer and inner circles); number of rearing (the

number of stands on the hind limbs); the time and

number of grooming acts, freezing time, defecation

acts. Exploratory activity and anxiety were also quan-

tified by cumulative indexes. The index of exploratory

activity summarized the numbers of rearings and en-

tries to the central zone. The index of anxiety sum-

marized the acts of grooming and defecation and the

grooming and freezing times.

The ECG was recorded for 3min using non-inva-

sive electrodes as described earlier [29]. The autonom-

ic regulation of the heart was assessed from the fol-

lowing ECG parameters: an average R-R interval(ms);

the range of R-R interval values, i.e., the difference be-

tween the maximal and minimal R-R values (dX, ms);

root mean square of successive differences in the R-R

intervals (RMSSD, ms); and stress index(SI). Physiolog-

ical monitoring was carried out on days 8, 15, and 25

of the experiment.

After physiological tests had been conducted on

day 25, the animals were sacrificed by decapitation.

The brains were excised and transferred on ice; cere-

bral cortices were separated and frozen in liquid nitro-

gen within 60-90s after decapitation.

Preparation of brain homogenates. Frozen cor-

tex tissue was homogenized in 50 mM MOPS buffer

(pH 7.0) containing 0.2mM EGTA, 1 mM DTT, 20% glyc-

erol, and the protease inhibitors (1 mM AEBSF, 0.8 μM

aprotinin, 50 μM bestatin, 10 μM pepstatin A, 15 μM

E-64, and 20 μM leupeptin) using a T10 Basic Ultra-Tur-

rax disperser (IKA, Germany) as described before [29].

Oneml of the buffer was used per 0.4 g of the tissue

fresh weight (gFW). The tissue was further disrupted

by sonication (7 cycles of 30-s sonication in a low-in-

tensity mode with 30-s pauses) with a Bioruptor soni-

cator (Diagenode, Belgium) in an ice-cold water bath.

The resulting homogenates were mixed at a 3 : 1 (v/v)

ratio with the solubilization buffer containing 40 mM

Tris-HCl (pH 7.4), 600 mM NaCl, 4 mM EDTA, 1% sodium

deoxycholate and 4% NP-40, and incubated for atleast

30min before the assays.

Enzyme assays. The activity of transketolase (TKT)

was assessed spectrophotometrically from the rate of

NADH oxidation in the triosephosphate isomerase/

glycerol-3-phosphate dehydrogenase coupled system

by the established method [31] modified to use the

microplate format [32]. The linear part of the product

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1695

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

accumulation curve (30min) was used to calculate the

reaction rate followed by subtraction of the background

reaction rate assayed without pentose phosphates.

Theactivity of pyruvate dehydrogenase complex (PDC)

was determined calorimetrically on a CLARIOstar Plus

microplate reader (BMG Labtech, Germany) from the

NADH production coupled to iodonitrotetrazolium re-

duction to formazan [33] with modifications described

earlier [34, 35]. The linear part of the product accu-

mulation curve from the 1st to 10thmin was used to

calculate the reaction rate. The activity of the 2-oxog-

lutarate dehydrogenase complex (OGDC) was assessed

from the absorbance of produced NADH at 340nm as

described before [8] using a Sunrise microplate reader

(Tecan, Austria). A steady-state reaction rate from the

5th to 10thmin, i.e., after completion of a lag period,

was used to calculate the activity. To assess the activ-

ities of endogenous holoenzymes of the ThDP-depen-

dent dehydrogenases, the assays were conducted in

the absence of MgCl

and ThDP in the reaction media.

Theendogenous holotransketolase was determined in

the medium lacking ThDP.

Quantification of brain metabolites. Free ami-

no and related amino compounds were quantified by

ion-exchange chromatography with ninhydrin derivat-

ization in the methanol-acetic acid extracts of the brain

cortex as described before [36]. Taurine was quantified

within a non-resolved peak with phosphoethanolamine

(PEA). In several studies on mammalian brain, relative

abundances of taurine and PEA vary from compara-

ble to up to 10 times more abundant taurine than PEA

[37-40]. Oxidized glutathione and NAD

were measured

with a fluorometric assay [41, 42] in black 96-well mi-

croplates using a CLARIOstar Plus microplate reader.

ThDP was quantified with the apo-TKT activation assay

[43] modified for the microplate format [44].

Statistical analysis was performed with the STA-

TISTICA (version 6.0), GraphPad Prism (version 8.4),

and R (version 4.3) software packages. The outliers

were identified with the iterative Grubbs’ test at Alpha

set to 0.01, and excluded from statistical analysis.

Thedifferences in the content of metabolites between

two groups were analyzed with the t-test in view of

the normal data distribution according to D’Agostino–

Pearson test. The differences in cumulative param-

eters of correlations between metabolites (sum and

mean correlation coefficients as well as the numbers

of significant positive and negative correlations) were

analyzed by Mann–Whitney test due to non-normal

distribution. When more than two groups were an-

alyzed, factorial analysis of variance (ANOVA) was

used together with the post-hoc comparison of group

averages by the Šídák test. The data were presented

as mean ±standard error of mean (SEM). The correla-

tions between the assayed parameters were analyzed

using the Spearman’s rank correlation coefficient, as

not all physiological parameter values were normally

distributed. The differences between the groups and

correlations were considered significant at p ≤ 0.05; the

values of p ≤ 0.1 were considered as trends.

RESULTS

Levels of ThDP, activities of ThDP-dependent

enzymes and redox-related compounds as indica-

tors of thiamine-dependent metabolism in the rat

cerebral cortex. The content of ThDP may character-

ize potential changes in the thiamine-dependent me-

tabolism after combined chronic administration of the

thiamine transport inhibitors metformin and amproli-

um, as ThDP is a coenzyme form and major derivative

of thiamine in the brain. In view of the essential role

of ThDP in the central redox metabolism, the levels of

NAD

, as well as those of the antioxidant peptides car-

nosine and glutathione, are other important indicators

of the thiamine-dependent metabolic changes.

As seen from Fig. 2, a minor decrease in the to-

tal ThDP content in the brain after the treatment does

not reach the level of statistical significance (p=0.28).

However, administration of thiamine transport inhibi-

tors causes a certain level of oxidative stress, which is

a well-known feature of thiamine deficiency [45, 46].

The oxidative stress is evidenced by a decrease in

the content of carnosine, which protects the brain

from the damage by peroxynitrite [47], and a trend

(p = 0.09) to elevation of oxidized glutathione (GSSG)

(Fig.2).

The total activities of ThDP-dependent enzymes in

the brain cortex, which are determined in the pres-

ence of ThDP in the assay media, do not differ in the

control and treated groups. However, the extent of en-

zyme activation by ThDP added to the assay medium

decreases after the treatment. For TKT, this is mani-

fested as a disappearance in the treated samples of

statistically significant, although small, activation of

TKT by ThDP that is observed in the control samples

(Fig. 3a, upper panel). For OGDC, there is a statistically

significant decrease in the apo-OGDC fraction in the

treated vs. control samples (Fig. 3c, bottom panel).

Activation of PDC by ThDP is ~100% in both the con-

trol and treated groups, which is in agreement with

the known requirement for ThDP addition to the PDC

assay medium [34, 48]. In contrast to TKT and OGDC,

the dissociation of PDC holoenzyme under the assay

conditions is complete in both the control and treated

groups, which does not allow one to detect changes

in the enzyme saturation with ThDP, which may have

been induced by the treatment. However, for both TKT

and OGDC, the levels of enzyme saturation with ThDP

increase after the treatment with metformin/amproli-

um, compared to the control group.

GRAF et al.1696

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Fig. 2. The levels of ThDP, NAD

, antioxidant peptides carnosine and glutathione (GSH), oxidized glutathione (GSSG) and the

ratio of the reduced and oxidized glutathione in the cerebral cortex of rats treated with metformin/amprolium(M+A) and

control animals. gFW,g of fresh weight.

a b c

Fig. 3. Activation of TKT(a), PDC(b), and OGDC(c) by ThDP addition to the assay medium in the brain cortices of the control

rats and rats treated with metformin/amprolium(M+A). In the upper panel, open and solid points indicate activities measured

in the absence and presence of ThDP, respectively (analysis by repeated measures ANOVA). Bottom panel compares the frac-

tions of endogenous TKT, PDC, and OGDC apoenzymes in the treated and control samples, calculated as [1–(Activity without

ThDP)/(Activity with ThDP)]× 100%. The outlier excluded from analysis is indicated with “x”.

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1697

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Fig. 4. Changes (%) in the content of amino acids and related compounds in the cerebral cortex of rats treated with metformin/

amprolium vs. control animals (n =12 in each group). The outliers outside of y-axis range (not shown) are excluded from

statistical analysis; *significant difference from the control values set as 100%.

If changes in the ThDP-induced activation of TKT

and OGDC were due to different ThDP contents in the

rat brains, the points of the endogenous holoenzyme

activities or the enzyme activation by ThDP vs. ThDP

level in the brains of the treated and non-treated ani-

mals would have occupied different regions of the XY

space. If changes in the enzyme activation by ThDP

were due to changes in the enzyme properties, the

points would have occupied the same XY space. The

correlations between the ThDP level and contents of

endogenous holo- or apoenzymes (Fig.S1 in the Online

Resource 1) favor the second rather than the first as-

sumption, as the data for both animal groups occupy

the same XY space in the graphs. Remarkably, apo-TKT

in the control rats demonstrates an expected statisti-

cally significant negative correlation with the ThDP

content in the brain. Alteration of this correlation in

the treated animals provides additional support for the

treatment-changed affinity of TKT to ThDP. The acti-

vation of mitochondrial OGDC by ThDP added to the

reaction medium shows no significant correlation with

the total tissue level of ThDP in either treated or con-

trol animals.

As a result, the treatment with metformin/ampro-

lium does not significantly decrease the total brain

levels of ThDP, but increases the affinities of TKT and

OGDC to the coenzyme ThDP in the rat brain.

Changes in the profiles of free amino acids and

related compounds in the rat brain after chronic

administration of metformin and amprolium. Com-

pared to the control group, chronic administration of

metformin/amprolium induces statistically significant

increases in the content of methyllysine, tryptophan,

serine, glutamate, aspartate, and beta-alanine, decreas-

ing the content of carnosine (Fig. 4). There are also

trends (0.05 < p ≤ 0.1) toward increases in the levels

of cystathionine, leucine, beta-aminoisobutyrate and

alanine. The content of free brain amino acids, includ-

ing BCAAs, mostly increases. This is accompanied by

a statistically significant elevation in the brain ammo-

nia level (Fig.4), which suggests hyperammonemia re-

sulting from the elevated degradation of accumulating

amino acids. Therefore, in addition to the perturbed

levels of the peptides involved in cell antioxidant de-

fense, i.e., carnosine and oxidized glutathione (Fig.2),

the overall perturbation of amino acid metabolism, as-

sociated with hyperammonemia, represents a marker

of pathological changes in the brain of the metformin/

amprolium-treated rats.

Analysis of correlations between the levels

of ThDP or activities of ThDP-dependent enzymes

and the content of brain metabolites. The average

tissue levels of ThDP or activities of ThDP-dependent

enzymes (Figs. 2 and 3) provide a rough measure of

changes in the thiamine status, but are not suitable to

resolve the differences in metabolic fluxes in the con-

trol and treated states. In addition to the mean values

of the parameters, pairwise correlations between the

thiamine status parameters and the content of metab-

olites of the ThDP-dependent network are useful to

characterize the treatment-induced metabolic chang-

es. As shown in Table1, the treatment strongly affects

significance of the correlations between the ThDP con-

tent or activities of ThDP-dependent enzymes and the

levels of redox indicators or amino acids. In particular,

the treatment induces significant positive correlations

between the levels of ThDP and pathological markers,

such as levels of glutathione disulfide, α-aminoadi-

pate and ammonia. Besides, ThDP levels in the brains

of the treated rats become positively correlated with

the contents of citrulline, ethanolamine, glycine, and

combined level of taurine and phosphoethanolamine.

GRAF et al.1698

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Table 1. The correlations between the levels of ThDP or activities of ThDP-dependent enzymes and the contents

of NAD

, antioxidant peptides, free amino acids or related metabolites in the cerebral cortex of the control rats

(Ctrl) and rats treated with metformin/amprolium (M+A)

Parameter

ThDP

PDC

–ThDP

PDC

+ThDP

OGDC

–ThDP

OGDC

+ThDP

–ThDP

+ThDP

Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A

NAD

0.34 –0.47 –0.21 –0.11 –0.05 0.10 0.66 –0.70 0.77 –0.60 0.36 0.30 –0.15 0.55

Carnosine 0.29 –0.33 –0.31 0.33 0.12 0.50 –0.08 0.08 0.21 0.38

0.27 –0.62 –0.07 –0.43

GSH 0.06 –0.07 –0.16 –0.01

–0.73 0.10 0.27 –0.13 –0.03 –0.01 –0.18 0.02 –0.45 0.15

GSSG

–0.05 0.76 0.38 –0.15 0.25 –0.60 –0.02 0.38 0.08 0.26 –0.03 0.31 0.11 0.11

GSH/GSSG –0.14 –0.54 –0.17 –0.16 –0.04 0.46

–0.11 –0.70 –0.11 –0.61 0.00 0.19 –0.23 0.38

GSH+2*GSSG –0.04 0.36 0.14 –0.52 0.02 –0.28 0.04 –0.27 0.16 –0.24

–0.16 0.70 –0.39 0.69

α-Aminoadipate –0.18 0.75 0.08 –0.01 0.10 –0.61 0.36 0.61 0.16 0.45 –0.07 0.15 –0.12 –0.01

α-Aminobutyrate 0.15 0.01

–0.08 –0.57 –0.08 –0.28 0.34 –0.31 0.20 –0.45 0.09 0.59 –0.03 0.71

Arg –0.45 0.26 0.31 0.37 0.29 0.11 –0.65 –0.05 –0.45 0.04 0.17 –0.10 0.34 0.01

β-Aminoisobutyrate

0.07 0.37 –0.12 –0.57 –0.69 –0.58 0.10 0.32 –0.30 0.06 –0.28 0.34 –0.36 0.21

Citrulline

–0.16 0.60 0.21 –0.10 –0.07 –0.36 0.50 –0.03 0.52 –0.20 –0.04 0.31 –0.28 0.22

Cystathionine 0.31 –0.17

–0.10 0.75 –0.15 0.45 –0.21 –0.27 –0.08 –0.09 0.47 –0.36 0.29 –0.26

Ethanolamine

0.52 0.60 –0.22 0.06 0.00 –0.31 0.38 0.26 0.51 0.41 0.35 0.23 –0.03 0.22

Gly

–0.24 0.57 0.01 0.07 0.42 –0.06 –0.29 0.01 –0.11 0.24 0.37 0.08 0.55 0.15

Hydroxylysine 0.11 0.30 –0.08 0.26 0.11 0.40

–0.60 0.06 –0.17 0.27 0.48 –0.49 0.37 –0.48

Ile 0.02 0.57 –0.03 0.15 0.35 –0.24 –0.54 0.01 –0.03 0.11

0.73 0.14 0.73 0.21

Leu –0.13 0.52 0.07 0.21 0.52 –0.20 –0.61 0.04 –0.13 0.19 0.58 0.10 0.69 0.18

Lys 0.76 0.53 –0.65 0.02 –0.21 –0.11 –0.04 0.11 0.27 0.32 0.52 0.03 0.10 0.06

Met –0.19 0.47 0.08 0.20 0.43 0.10

–0.76 –0.22 –0.23 –0.03 0.59 –0.02 0.65 –0.02

NH3 0.19 0.66 –0.01 –0.01 –0.26 –0.26 –0.37 0.08 –0.49 0.31 0.20 0.30 0.40 0.29

Phe –0.47 0.43 0.44 0.27

0.57 0.21 –0.57 0.01 –0.32 0.30 0.15 –0.24 0.50 –0.19

Phosphoserine 0.10 0.31 –0.15 –0.55 0.29 –0.22

0.33 0.59 0.61 0.76 0.32 0.08 0.07 –0.03

Taurine+PEA

0.19 0.79 –0.22 –0.31 –0.29 –0.48 0.34 0.42 0.33 0.38 0.30 0.34 –0.13 0.00

Thr

0.57 0.49 –0.38 –0.13 0.01 0.06 –0.24 –0.19 0.03 0.18 0.88 0.20 0.64 0.29

Trp –0.79 –0.33 0.73 0.56 0.20 0.25 –0.49 –0.59 –0.66 –0.36 –0.08 0.03 0.35 0.29

Urea 0.33 0.43 –0.15 0.17 0.24 –0.13 0.07 –0.25 0.35 –0.02

0.63 0.46 0.28 0.54

Val –0.21 0.44 0.28 0.27

0.70 –0.02 –0.20 –0.19 0.17 0.12 0.56 0.15 0.67 0.29

Note. The table shows metabolites whose contents demonstrate at least one significant correlation with the levels of ThDP or

activities of ThDP-dependent enzymes. Each cell in the table shows the Spearman’s rank correlation coefficient for a given pair

of parameters. Coefficients of the correlations with changed level of statistical significance in the control and treated states

areshown in red on grey background. Coefficients of significant correlations (p≤0.05) are given in bold.

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1699

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

In contrast, significances of the positive correlation be-

tween the levels of ThDP and lysine, and the negative

correlation between the levels of ThDP and tryptophan

in the control animals are reduced by the treatment.

Taken together, these data point to changed metabolic

contribution of ThDP to the cellular redox state and

metabolism of amino acids after chronic administra-

tion of thiamine transport inhibitors.

Similarly, significances of the correlations be-

tween the brain activities of ThDP-dependent enzymes

and the levels of redox indicators or amino acids are

strongly changed by the treatment (Table1). In partic-

ular, after the treatment, the OGDC activity correlates

negatively with the levels of NAD

and glutathione

redox state (GSH/GSSG), while in the control animals

the OGDC activity exhibits a strong positive correlation

with the NAD

content. Unlike the brain TKT activity of

the control rats, the brain TKT activity of the treated

rats correlates positively with the levels of total gluta-

thione (GSH+2*GSSG) and α-aminobutyrate, and nega-

tively with the level of carnosine. In addition, signifi-

cant positive correlations of the TKT activity with the

BCAA levels, observed in the control group, disappear

after the treatment. Overall, correlation analysis indi-

cates that the metabolic relationships between ThDP

or ThDP-dependent enzymes and free amino acids or

related compounds are different in the treated and

control rat brain.

The treatment effects on the rat behavior

in the open field test and ECG. Physiological tests

during the experiment help understanding how the

observed changes in the metabolism of brain amino

acids affect animal behavior and ECG. Unlike the bio-

chemical characterization of the brain tissue, physio-

logical testing may be repeated in the course of exper-

iment, thus providing insights into the development

of treatment-induced changes. According to ANOVA,

the factor “treatment” by metformin/amprolium is sig-

nificant for the number of entries to the central zone

(Fig. 5a) and associated parameter of locomotor activ-

ity (Fig. 5b). The cumulative indexes of exploratory

activity and anxiety also show the significance of the

“treatment” factor (Fig. 5b). Significant interaction be-

tween the number of entries to the central zone and

days of treatment (Fig. 5a) manifests in an increase in

this parameter after the day 15 in the treated rats only.

As a result, on the day 25, the number of entries to

the central zone for the treated animals is significantly

higher than that for the control ones. Obviously, this

difference contributes the most to a higher cumulative

index of exploratory activity in the treated vs. control

rats. The overall significances of the treatment for the

locomotor activity (p = 0.046) and cumulative anxiety

index (p = 0.03) do not reveal particular differences be-

tween the groups tested on specific days of the treat-

ment (Fig.5b).

In view of the fact that the treatment does not af-

fect the number of rearings, but is significant for the

number of entries to the central zone (p = 0.02) and the

locomotor activity (p = 0.046), the observed increase

in the number of entries to the central zone and ac-

companying increase in exploratory activity (p = 0.02)

seem to be linked to the overall elevation of locomotor

activity, induced by the treatment (Fig.5, a, b). These

behavioral changes in the treated rats are reciprocated

by their increased heart rate, as the treatment induces

a decrease in the R-R interval of ECG in the treated

rats only (Fig.5c).

Therefore, chronic administration of metformin

and amprolium affects several behavioral and ECG

parameters in a time-dependent manner. The most sig-

nificant differences between the control and treated

animals are observed by the end of experiment, man-

ifested in the increased number of entries to the cen-

tral zone and related cumulative index of exploratory

activity in the treated vs. control rats.

Analysis of correlations between the levels of

ThDP or activities of ThDP-dependent enzymes and

the behavioral or ECG parameters. In order to un-

derstand how much the thiamine status and its inter-

play with the content of brain metabolites (Table 1)

contribute to the metformin/amprolium-induced

changes in rat behavior and ECG, the correlations of

thiamine-dependent biochemical parameters with the

physiological ones are analyzed. The results of analy-

sis (Table 2) reveal an interesting time dependence of

the treatment effects on the correlations. In the rats

sacrificed on the day 25, the ThDP levels at this final

day of experiment correlate more significantly with

the behavioral parameters determined on the day 15

than with those determined on the day 25. That is, in

the control animals the brain ThDP levels on the day

25 correlate positively with the cumulative index of

anxiety on the day 15 and negatively– with the cumu-

lative index of exploratory activity on the day 15, with

both the correlations disappearing in the treated rats.

Besides, in the treated rats, the brain ThDP level on the

day 25 correlates negatively with locomotion on the

day 15, while the correlation is absent in the control

rats. Remarkably, these correlations are in accordance

with physiological considerations. In the control ani-

mals, the content of ThDP is positively associated with

anxiety and negatively with the exploratory activity, as

exploratory activity is known to be higher when anxi-

ety is low. Similarly, in the treated animals, the content

of ThDP correlates negatively with both the locomotor

activity and the number of entries to the central zone,

since the two physiological parameters are known

to change in a similar direction. The graphs in Fig.5

show that on day 15, the trends observed for many

assessed physiological parameters switch to the oppo-

site: if a parameter is mostly decreasing from day 8

GRAF et al.1700

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Fig. 5. Time-dependent changes in rat behavior in the open field test(a, b) and ECG parameters(c) during chronic admin-

istration of metformin/amprolium(M+A) in comparison with the control rats. Statistically significant differences between

the groups are shown on the graphs; statistically significant factors, such as “treatment” by metformin/amprolium and the

“treatment time”, along with their interaction, are shown under the graphs (according to the repeated measures ANOVA with

Šídák post-hoc test).

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1701

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

to day 15, it starts to increase from day 15 to day 25,

and vice versa. This switch might be of either adaptive

or maladaptive nature. The importance of ThDP levels

in the observed responses is emphasized by the fact that

the number of entries to the central zone is the only be-

havioral parameter that differs significantly on day25

Table 2. Correlations of the brain levels of ThDP or activities of ThDP-dependent enzymes in the presence(+)

orabsence(–) of ThDP in the assay media with ECG of behavioral parameters in the control (Ctrl) and metformin/

amprolium-treated(M+A) rats

Parameter

ThDP

PDC

–ThDP

PDC +ThDP

OGDC

–ThDP

OGDC

+ThDP

TKT

–ThDP

TKT

+ThDP

Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A Ctrl M+A

R-R interval –0.24 –0.19 0.03 –0.08

0.65 0.06 0.02 –0.19 –0.01 –0.15 –0.22 0.13 –0.09 0.24

dX –0.37 0.05 0.27 –0.23 0.04 –0.08 0.18 0.02 –0.10 –0.02

–0.75 0.48 –0.50 0.42

RMSSD –0.10 –0.18 0.02 0.09 0.11 –0.21 0.13 0.29 0.03 0.20 –0.34 0.04 –0.35 –0.01

SI 0.36 –0.06 –0.08 0.34 –0.27 0.21 –0.06 0.03 0.21 0.03

0.60 –0.40 0.38 –0.43

Anxiety, day 8 0.04 –0.02 0.20 0.41 –0.52 0.02 –0.14 –0.20 –0.45 –0.30 –0.37 –0.05 –0.10 –0.06

Anxiety, day 15

0.57 0.02 –0.53 –0.10 0.05 0.27 0.30 –0.05 0.45 –0.01 0.29 –0.38 0.01 –0.24

Anxiety, day 25 –0.16 0.08 0.01 –0.11 0.36 0.14 0.44 0.08 0.14 0.02

–0.59 0.21 –0.39 0.03

Grooming time –0.04 0.52 –0.15 –0.09 –0.14 –0.40 0.00 0.21 0.10 –0.04 –0.28 0.42 –0.30 0.01

Freezing time –0.31 –0.19 0.28 –0.14 0.38 0.36 0.47 –0.19 0.15 –0.14

–0.65 0.09 –0.37 0.23

Defecation acts –0.44 –0.11 0.14 –0.37 0.28 0.23 0.05 –0.43 –0.07 –0.01 –0.45 0.03 –0.07 0.24

Grooming acts –0.25 –0.17 0.22 0.04 0.05 –0.05 –0.24 0.18 –0.06 –0.05 –0.27 0.23 –0.19 0.09

Exploratory

activity, day 8

–0.33 –0.46 0.04 –0.28

0.61 0.13 –0.06 0.00 0.15 0.23 0.39 0.07 0.32 0.22

Exploratory

activity, day15

–0.60 –0.25 0.43 –0.22 0.19 –0.13 –0.34 –0.23 –0.47 –0.22 –0.10 0.22 0.03 0.22

Exploratory

activity, day 25

–0.10 –0.33 0.11 0.21 0.07 0.02 –0.43 –0.25 –0.01 –0.32

0.58 –0.29 0.49 –0.15

Number

of rearings

–0.06 0.03 0.25 0.13 0.18 –0.27 –0.45 –0.27 0.01 –0.32

0.63 0.03 0.57 0.09

Number of entries

to the central zone

–0.38 –0.80 0.09 0.39 0.03 0.55 –0.43 –0.10 –0.34 –0.08 0.12 –0.69 0.20 –0.39

Locomotor

activity, day 8

–0.40 –0.52 0.05 0.20 0.41 0.31 0.08 0.04 0.22 0.16 0.15 –0.32 0.06 –0.06

Locomotor

activity, day 15

–0.26 –0.58 0.21 0.22 –0.36 –0.02 0.02 –0.18 –0.25 –0.34 –0.18 0.16 –0.30 0.23

Locomotor

activity, day 25

–0.05 0.10 0.07 0.00 –0.12 –0.30

–0.61 0.12 –0.24 0.06 0.60 –0.15 0.52 –0.18

Note. Each cell shows the Spearman’s rank correlation coefficient for a given pair of parameters. Correlations whose statistical

significance (p≤0.05) differ between the control and treated animals, are shown in red on grey background. Statistically sig-

nificant correlation coefficients are given in bold. The correlations with “Locomotor activity” and cumulative indexes “Anxiety”

or “Exploratory activity”, use the data on days 8, 15, and 25; the correlations with other physiological parameters are shown

only for the data on day 25.

GRAF et al.1702

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

between the treated and control animals, and at the

same time, the only parameter in the treated group

that correlates with the ThDP levels on day 25. Signifi-

cant correlations of both the brain ThDP levels or activ-

ity of ThDP-dependent TKT with the number of entries

to the central zone on day 25 (Table2) indicate that the

physiological switch on day 15, when the difference be-

tween the groups has not been yet manifested (Fig.5),

is linked to the ThDP-dependent brain metabolism.

At the end of experiment, the activities of

ThDP-dependent enzymes mostly exhibit significant

correlations with behavioral or ECG parameters in the

control animals, with the correlations disappearing in

the treated animals (Table2). The same is observed for

the correlation between the PDC activity and explor-

atory activity on day8. Only the negative correlation

between the TKT activity and the number of entries to

the central zone arises in the treated animals, absent

in the control ones.

Overall, the activity of endogenous holoenzyme of

TKT shows the highest number of significant correla-

tions, either positive or negative, with the behavior-

al or ECG parameters, mostly in the control animals

(Table2).

Changes in the rat brain, characterized by the

treatment-modified correlations of biochemical

and physiological parameters. The patterns of inter-

action between the contents of metabolites, enzyme

activities, and physiological parameters in the control

and treated rats are shown in Figs. 6 and 7, respectively.

Fig. 6. Correlation matrix for physiological and biochemical parameters in the control rats. Thick lines divide the parameters

into five groups: (1) ECG parameters, (2) anxiety parameters, (3) parameters of exploratory and locomotor activities, (4) ac-

tivities of ThDP-dependent enzymes together with ThDP levels, and (5) levels of NAD

, amino acids, and related compounds.

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1703

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Fig. 7. Correlation matrix for physiological and biochemical parameters after chronic administration of metformin/amproli-

um(M+A). Thick lines divide the parameters into five groups: (1) ECG parameters, (2) anxiety parameters, (3) parameters of

exploratory and locomotor activities, (4) activities of ThDP-dependent enzymes together with ThDP levels, and (5)levels of

NAD

, amino acids, and related compounds.

Overall, the correlation matrices for the control and

treated animals reveal a significantly higher positive

interdependence between the contents of brain ami-

no acids and related compounds (group 5 in Figs. 6

and7) after the treatment with metformin/amprolium.

Compared to the control animals, a 2-fold increase in

both the number of positive correlations and values

of corresponding correlation coefficients is observed

(TableS1 in the Online Resource 1). A simultaneous in-

crease in the content of free amino acids in the brain

(Fig. 4) and in the number of positive correlations be-

tween the levels of different amino acids manifests a

common cause for the elevation in content of different

amino acids, most probably protein degradation.

Important shifts in the correlations between the

levels of metabolites and activities of ThDP-depen-

dent enzymes, caused by the treatment (Table 1 and

discussion above), are accompanied by changes in

the correlations between the ThDP-linked markers of

pathological states. In particular, the change in the sign

of the correlation between the NAD

content and the

OGDC activity is accompanied by changes in the signif-

icances of other NAD

correlations after the treatment.

In the control animals, the NAD

levels are positively

correlated to the contents of urea and combined lev-

els of taurine (antioxidant) and phosphoethanolamine.

After the treatment, these correlations are substituted

by the positive correlations between the levelsofNAD

GRAF et al.1704

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

and tryptophan or glutathione antioxidant potential

(GSH/GSSG), added by a significant negative correla-

tion between the levels

of NAD

and α-aminoadipate.

Remarkably, in the treated animals, the levels of α-ami-

noadipate and GSH/GSSG correlate negatively, while

the levels of α-aminoadipate and GSSG correlate posi-

tively, with all these correlations absent in the control

animals. Significant correlations between the levels

of antioxidant carnosine and ammonia (negative), al-

anine (positive), or histidine (positive) are observed

in the control animals. In the treated animals, these

correlations are substituted by significant negative

correlations between the levels of carnosine and to-

tal glutathione or activity of TKT, and by significant

positive correlations between the levels of carnosine

and proline or cystine. Thus, simultaneously with mod-

ified interactions of the levels of ThDP or activities of

ThDP-dependent enzymes with other parameters, ad-

ministration of metformin/amprolium changes the re-

lationships involving different pathological markers,

such as GSSG which is a marker of oxidative stress,

or α-aminoadipate which is a marker of diabetes [3].

Regarding changes in the relationships between

the biochemical and physiological parameters, the

treatment with metformin/amprolium switches the

sign of correlations between the contents of amino ac-

ids and behavioral parameters of the groups2 and3.

Predominantly negative correlations between the pa-

rameters of anxiety and the contents of amino acids in

the control animals become predominately positive in

the treated animals. At the same time, the opposite is

observed for the correlations between the amino acid

contents and parameters of exploratory and locomotor

activities (Figs.6 and 7). The fact that in the different

states of the rats, the relationships between the brain

amino acids and behavioral parameters are different,

corresponds to the non-causative nature of correla-

tions. Nevertheless, irrespectively of the treatment, the

opposite relationships between the contents of amino

acids and behavioral parameters in the groups 2 and3

are observed. This finding corresponds to a negative

relation between anxiety and exploratory or locomotor

activities: an anxious animal prefers neither to move,

nor to explore.

The negative correlations of SI (an indicator of

sympathetic activity) with the duration of the R-R in-

tervals and their variability (dX) manifest the basic

physiological relationship: The lower the sympathetic

activity, the longer the R-R interval and the higher its

variability. The significances of these correlations are

therefore preserved after the treatment. However, the

negative correlation of SI with RMSSD in the control

group disappears in the treated group. Simultaneously,

a positive correlation of dX with RMSSD in the con-

trol rats is substituted by the positive correlation of dX

with the R-R interval. Hence, chronic administration

of metformin/amprolium affects the balance between

the sympathetic and parasympathetic regulation man-

ifested in the correlations between the specific ECG

parameters. These changes are accompanied by al-

terations in the correlations between ECG parameters

and the contents of brain metabolites involved in the

cellular redox regulation (reduced, oxidized and total

glutathione, cystine, cystathionine, taurine+phosphoe-

thanolamine), or some other amino acids (isoleucine,

aspartate, and threonine) (Figs.6 and 7). The negative

correlation between the R-R interval and the levels of

reduced glutathione is not changed by the treatment,

representing the only significant correlation of the R-R

interval with the contents of studied metabolites.

Thus, administration of metformin/amprolium af-

fects the balance between the sympathetic and para-

sympathetic regulation, that is associated with changes

in the correlations between the ECG parameters and

levels of the brain amino acids. Simultaneously, the

regulation of this balance is manifested as changes

in the anxiety indicators. In rodents, freezing (a form

of behavioral inhibition that is also an indicator of

anxiety) is accompanied by a decrease in the heart

rate[49]. In our control animals, this is evident from

the positive correlations of freezing time with the du-

ration of R-R intervals and dX, as well as a negative

correlation of freezing time with SI. However, only the

positive correlation between the freezing time and dX

is significant after the treatment, further supporting

the treatment-induced changes in the balance of sym-

pathetic/parasympathetic regulation.

DISCUSSION

In this work, we show that chronic administration

of the thiamine influx inhibitors metformin and am-

prolium perturbs the amino acid metabolism in the rat

brain. Perturbed oxidation of BCAAs and other changes

in the brain amino acid metabolism are known to be

associated with thiamine deficiency [6, 7]. As discussed

in the Introduction, metformin is a commonly used an-

tidiabetic drug, whose well-known but underestimated

side effect is inhibition of thiamine transporters. Am-

prolium is a veterinary drug whose antiparasitic ac-

tion is based on the blockade of thiamine transport.

Although the parasites have a higher sensitivity to the

such blockade than their hosts [50], depending on a

dose, the hosts can also be affected. So amprolium is

used to model thiamine deficiency in animals and an-

imal cells [16, 25]. Because impaired thiamine status

is a common comorbidity in diabetes, we use chronic

combined administration of metformin and amproli-

um to model the effect of metformin in diabetic pa-

tients who have undiagnosed thiamine insufficiency.

High sensitivity of the nervous system to disturbances

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1705

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

in the thiamine status is in particular due to the thi-

amine dependence of the metabolism of amino acid

neurotransmitters or their precursors [30, 35, 41, 51].

Hence, disturbed thiamine status in the brain may also

influence behavioral and ECG parameters, thus justi-

fying our interest in the effects of the thiamine trans-

porters inhibitors on the brain metabolism.

We show that chronic administration of met-

formin/amprolium increases the levels of BCAAs in the

brain (Fig.4), which is in agreement with the results of

independent studies on the metformin-impaired BCAA

degradation [4]. The disappearance of significant cor-

relations between the contents of these amino acids

and activities of ThDP-dependent enzymes (Table 1)

in the treated animals links changes in the BCAA lev-

els with alterations in the thiamine-dependent me-

tabolism. The absence of the effect of metformin/

amprolium treatment on the content of total brain

ThDP (Fig. 2) may be partly due to the buffering role

of strong inhibition of thiamine diphosphokinase by

ThDP [52]. An increased saturation of TKT and OGDC

with the coenzyme by the end of treatment (Fig. 3)

suggests compensatory adaptations of ThDP-binding

enzymes in response to the chronic administration of

thiamine transport inhibitors. Such adaptations are

inferred from non-monotonous changes in the physio-

logical parameters during the experiment (Fig.5) and

correlations of these time-dependent parameters with

the end-stage levels of ThDP and ThDP-dependent en-

zymatic activities (Table2). Apart from the tissue ThDP

levels, the saturation of ThDP-dependent enzymes with

ThDP may also depend on the allosteric effects [53] or

posttranslational modifications of the enzymes [32, 44].

A higher extent of TKT saturation with ThDP has also

been observed in various pathological conditions in

independent studies (see [44] and references there-

in). Irrespectively of complex molecular mechanisms

involved, the increase in the saturation of brain TKT

and OGDC with ThDP after chronic administration of

metformin/amprolium indicates that the drugs affect

central glucose metabolism in both the cytoplasm and

mitochondria. Although the treatment-induced changes

in the average levels of redox indicators (Fig. 2) and

activities of ThDP-dependent enzymes (Fig. 3) appear

minor, the symptoms of oxidative stress (decrease in

the content of the antioxidant carnosine and a trend

toward increase in the content of oxidized glutathione,

Fig. 2) occur along with significant disturbances in

the amino acid metabolism and elevation of ammonia

(Fig. 4). These changes are accompanied by the treat-

ment-induced modifications of animal behavior and

ECG (Fig.5).

The effects of chronic administration of metformin

and amprolium on the content of brain amino acids

and relationship between their levels (Fig.4, TableS1

in the Online Resource 1) are strikingly similar to those

observed in the brain metabolism after acute hypox-

ia [54]. Both insults increase the levels of free amino

acids in the brain and strengthen their correlations

to each other. The similarity of effects of these very

different metabolic perturbations on the brain me-

tabolism of amino acids implies strong impairment of

oxidative energy metabolism not only in the case of

acute hypoxia, but also after chronic administration

of metformin and amprolium. As thiamine is essential

for oxidative metabolism, the similarity supports the

thiamine dependence of the perturbed metabolism of

amino acids after chronic administration of metformin

and amprolium inhibiting the intracellular thiamine

transport.

The elevated levels of free amino acids and am-

monia, with the strongest increase in the content of

free methyllysine (Fig.4), a product of proteolytic deg-

radation, point to perturbations in proteostasis at the

end of the treatment. A well-known mediator of the

metformin action, AMPK induces autophagy under nu-

trient stress [55-57]. The decrease in the intracellular

thiamine levels at the early stages of treatment may

represent a stress signal that activates autophagy by

the end of the treatment.

Hyperammonemia is often associated with per-

turbed amino acid metabolism, aberrant proteostasis,

and oxidative stress [58, 59]. Dysregulation of ThDP-de-

pendent enzymes induces hyperammonemia by caus-

ing insufficiency of the urea cycle due to the deficit of

acetyl-CoA and ATP [60]. The involvement of ThDP-de-

pendent metabolism in brain hyperammonemia caused

by chronic administration of metformin/amprolium, is

supported by a strong correlation between the brain

levels of ThDP and NH

at the end of the treatment,

which is not observed in the control animals (Table 1).

The increase in the OGDC saturation with ThDP as

an adaptation to reduced thiamine content may con-

tribute to hyperammonemia by activating amino acid

degradation in the TCA cycle, as OGDC catalyzes the

rate-limiting step of the cycle. Indeed, not only the lev-

els of ThDP, but also the activities of ThDP-dependent

enzymes in the brain strongly change their associa-

tions with the levels of amino acids (Table1).

In view of the significance of α-aminoadipate

as a marker of diabetes [3] and degradation of this

compound via metabolic pathway mediated by ThDP-

dependent 2-oxoadipate dehydrogenase [61], the treat-

ment-induced correlations between the levels of α-ami-

noadipate and ThDP or activities of ThDP-dependent

dehydrogenases (Table 1), are worth noting. The ac-

tivity of mitochondrial 2-oxoadipate dehydrogenase

complex cannot be assessed in brain homogenates be-

cause of a much higher expression of 2-oxoglutarate

dehydrogenase compared to 2-oxoadipate dehydroge-

nase and the fact that both complexes can utilize 2-ox-

oglutarate and 2-oxoadipate as substrates [8, 51, 62].

GRAF et al.1706

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Remarkably, however, in vivo inhibition of 2-oxoadi-

pate dehydrogenase by the active site-directed inhib-

itors decreases the carnosine content and increases

the β-alanine content in the rat brain [51]. The same

changes are observed after chronic administration

of metformin/amprolium (Figs. 1 and 4), suggesting

that the treatment-induced thiamine deficit inhib-

its 2-oxoadipate dehydrogenase due to the decreased

availability of ThDP for the enzyme saturation. This

treatment-induced increase in the OGDC affinity to

ThDP (Fig. 2) may contribute to the decreased avail-

ability of ThDP to 2-oxoadipate dehydrogenase. The in-

ability of 2-oxoadipate dehydrogenase to successfully

compete with OGDC for ThDP is supported by existence

of a positive correlation between the endogenous lev-

el of holoOGDC and the content of α-aminoadipate, an

upstream intermediate of the 2-oxoadipate dehydroge-

nase substrate, in the treated, but not control, animals

(Table 1). This correlation means that the higher the

activity of OGDC endogenously saturated with ThDP,

the higher the brain α-aminoadipate level. In other

words, the higher the OGDC activity, the less efficient

the degradation of the α-aminoadipate transamination

sibling 2-oxoadipate by 2-oxoadipate dehydrogenase.

The inhibition of 2-oxoadipate dehydrogenase by the

chronic administration of metformin/amprolium is fur-

ther supported by the positive correlation between the

levels of NAD

and tryptophan and negative correlation

between the NAD

content and α-aminoadipate content

in the treated, but not in control, rats (Figs. 6 and 7).

Asnoted elsewhere [62], inhibition of 2-oxoadipate de-

hydrogenase may promote NAD

biosynthesis from

tryptophan through quinolinic acid, a pathway alterna-

tive to the tryptophan degradation through α-aminoad-

ipate. Therefore, correlation analysis provides evidence

that chronic administration of metformin/amprolium

inhibits 2-oxoadipate dehydrogenase, thus increasing

the impact of tryptophan-dependent NAD

biosynthe-

sis in the rat brain. Induction of metabolic markers

of 2-oxoadipate dehydrogenase dysfunction and α-ami-

noadipate accumulation in the brain deserves special

attention in view of the known association between

the blood levels of α-aminoadipate or dysfunction

of2-oxoadipate dehydrogenase and diabetes, obesity,

and cardiovascular disorders (reviewed in[51]).

The metformin/amprolium-induced changes in

the rat brain biochemistry are associated with chang-

es in the rat behavior and ECG parameters. Adminis-

tration of metformin/amprolium perturbs the balance

between the sympathetic and parasympathetic regula-

tion, which is evident from the correlations between

the ECG parameters and freezing time.

It is known that amprolium reduces the explorato-

ry activity [23], while metformin demonstrates anxio-

lytic and antidepressive effects [63, 64]. In our study,

the combined administration of metformin and ampro-

lium produces a borderline anxiolytic action (p = 0.05

for the treatment factor, Fig. 5) accompanied by an

increase in the locomotor activity, which presumably

contributes to increases in the number of entries to

the central zone and cumulative index of exploratory

activity in the treated vs. control animals. These effects

are linked to the time-dependent increases in the num-

ber of entries to the central zone and the heart rate,

observed in the treated animals only (Fig.5). Theab-

sence of these changes in the control rats may manifest

their better accommodation to repeated testing, com-

pared to the treated rats.

Along with the inhibition of thiamine transport,

other mechanisms of the metformin action, in partic-

ular its activation of the AMPK-dependent pathways,

may underlie the complexity of biochemical and

physiological effects of the combined administration

of metformin and amprolium. Our study highlights a

conditional nature of metformin action. The benefits of

metformin administration can be reduced by exacer-

bation of thiamine insufficiency due to various comor-

bidity factors, including nutritional problems and/or

genetic variants of thiamine transporters in humans.

CONCLUSION

Our scheme of chronic administration of thia-

mine transport inhibitors metformin and amprolium

does not significantly change the final levels of ThDP

and total activities of ThDP-dependent enzymes in the

treated vs. control rats. However, the non-monotonous

changes in the behavior and ECG of the treated rats

in the course of experiment, and the relationships

of behavioral or ECG parameters with the levels of

ThDP or activities of ThDP-dependent enzymes at the

end of experiment suggest adaptation to the impaired

thiamine availability. The treatment with metformin/

amprolium switches the metabolic effects of ThDP

and ThDP-dependent enzymes from the antioxidant

and nitrogen-sparing to the pro-oxidant and hyperam-

monemic ones. This metabolic switch is accompanied

by an increase in the heart rate in the treated, but not

control, animals, and locomotion-related elevation of

the exploratory activity in the treated vs. control rats.

Supplementary information. The online version

contains supplementary material available at https://

doi.org/10.1134/S0006297924100043.

Contributions. A.V.G. planned and performed

animal experiments; A.V.A. assayed dehydrogenases

of 2-oxo acids; O.N.S. assayed TKT and ThDP; A.L.K.

quantified amino acids; A.V.G., A.V.A., and V.I.B. ana-

lyzed and visualized results. V.I.B. developed the study

concept and design, supervised the study, and wrote

the manuscript.

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1707

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

Funding. This work was supported by the State

Program AAAA-A19-119042590056-2.

Ethics declarations. This work does not contain

any studies involving human subjects. All animal ex-

periments were approved by the Bioethics Committee

of the Lomonosov Moscow State University (protocol

137-d 11.11.2022). The authors of this work declare

that they have no conflicts of interest.

REFERENCES

1. White, P. J., McGarrah, R. W., Herman, M. A., Bain,

J. R., Shah, S. H., and Newgard, C. B. (2021) Insulin

action, type 2 diabetes, and branched-chain amino ac-

ids: a two-way street, Mol. Metab., 52, 101261, https://

doi.org/10.1016/j.molmet.2021.101261.

2. Tobias, D. K., Mora, S., Verma, S., and Lawler, P. R.

(2018) Altered branched chain amino acid metabo-

lism: toward a unifying cardiometabolic hypothesis,

Curr. Opin. Cardiol., 33, 558-564, https://doi.org/10.1097/

HCO.0000000000000552.

3. Desine,S., Gabriel, C.L., Smith, H.M., Antonetti, O.R.,

Wang, C., Calcutt, M. W., Doran, A. C., Silver, H. J.,

Nair, S., Terry, J. G., Carr, J. J., Linton, M. F., Brown,

J. D., Koethe, J.R., and Ferguson, J.F. (2023) Associ-

ation of alpha-aminoadipic acid with cardiometabol-

ic risk factors in healthy and high-risk individuals,

Front. Endocrinol. (Lausanne), 14, 1122391, https://

doi.org/10.3389/fendo.2023.1122391.

4. Zhao,X., Zhang,X., Pei,J., Liu,Y., Niu,W., and Sun,H.

(2023) Targeting BCAA metabolism to potentiate met-

formin’s therapeutic efficacy in the treatment of di-

abetes in mice, Diabetologia, 66, 2139-2153, https://

doi.org/10.1007/s00125-023-05985-6.

5. Rivera, C.N., Watne, R.M., Brown, Z.A., Mitchell, S.A.,

Wommack, A. J., and Vaughan, R.A. (2023) Effect of

AMPK activation and glucose availability on myo-

tube LAT1 expression and BCAA utilization, Amino

Acids, 55, 275-286, https://doi.org/10.1007/s00726-022-

03224-7.

6. Navarro, D., Zwingmann, C., and Butterworth, R. F.

(2008) Impaired oxidation of branched-chain ami-

no acids in the medial thalamus of thiamine-de-

ficient rats, Metab. Brain Dis., 23, 445-455, https://

doi.org/10.1007/s11011-008-9105-6.

7. Navarro, D., Zwingmann, C., and Butterworth, R. F.

(2008) Region-selective alterations of glucose oxidation

and amino acid synthesis in the thiamine-deficient rat

brain: a re-evaluation using 1H/13C nuclear magnetic

resonance spectroscopy, J.Neurochem., 106, 603-612,

https://doi.org/10.1111/j.1471-4159.2008.05410.x.

8. Tsepkova, P.M., Artiukhov, A.V., Boyko, A.I., Aleshin,

V.A., Mkrtchyan, G.V., Zvyagintseva, M. A., Ryabov,

S. I., Ksenofontov, A. L., Baratova, L.A., Graf, A. V.,

and Bunik, V. I. (2017) Thiamine induces long-term

changes in amino acid profiles and activities of 2-ox-

oglutarate and 2-oxoadipate dehydrogenases in rat

brain, Biochemistry (Moscow), 82, 723-736, https://

doi.org/10.1134/S0006297917060098.

9. Bunik, V.I. (2023) Editorial: experts’ opinion in medi-

cine 2022, Front. Med. (Lausanne), 10, 1296196, https://

doi.org/10.3389/fmed.2023.1296196.

10. Muley, A., Fernandez, R., Green, H., and Muley, P.

(2022) Effect of thiamine supplementation on glycae-

mic outcomes in adults with type 2 diabetes: a system-

atic review and meta-analysis, BMJ Open, 12, e059834,

https://doi.org/10.1136/bmjopen-2021-059834.

11. Tamaki, H., Tsushima, H., Kachi, N., and Jimura, F.

(2022) Cardiac dysfunction due to thiamine deficiency

after hemodialysis for biguanide-related lactic acido-

sis, Intern. Med., 61, 2905-2909, https://doi.org/10.2169/

internalmedicine.8697-21.

12. Alston, T. A. (2003) Does metformin interfere with

thiamine? Arch. Intern. Med., 163, 983, https://doi.org/

10.1001/archinte.163.8.983.

13. Greenwood, J., and Pratt, O. E. (1985) Comparison

of the effects of some thiamine analogues upon thi-

amine transport across the blood-brain barrier of

the rat, J. Physiol., 369, 79-91, https://doi.org/10.1113/

jphysiol.1985.sp015889.

14. Dudeja, P.K., Tyagi,S., Kavilaveettil, R.J., Gill,R., and

Said, H. M. (2001) Mechanism of thiamine uptake

by human jejunal brush-border membrane vesicles,

Am. J. Physiol. Cell Physiol., 281, C786-C792, https://

doi.org/10.1152/ajpcell.2001.281.3.C786.

15. Bizon-Zygmanska, D., Jankowska-Kulawy, A.,

Bielarczyk, H., Pawelczyk, T., Ronowska, A.,

Marszall, M., and Szutowicz, A. (2011) Acetyl-CoA

metabolism in amprolium-evoked thiamine pyro-

phosphate deficits in cholinergic SN56 neuroblasto-

ma cells, Neurochem. Int., 59, 208-216, https://doi.org/

10.1016/j.neuint.2011.04.018.

16. Bunik, V. I., Tylicki, A., and Lukashev, N. V. (2013)

Thiamin diphosphate-dependent enzymes: from

enzymology to metabolic regulation, drug design

and disease models, FEBS J., 280, 6412-6442, https://

doi.org/10.1111/febs.12512.

17. Wernery,U., Haydn-Evans,J., and Kinne,J. (1998) Am-

prolium-induced cerebrocortical necrosis (CCN) in

dromedary racing camels, Zentralbl. Veterinarmed. B,

45, 335-343, https://doi.org/10.1111/j.1439-0450.1998.

tb00802.x.

18. Kasahara, T., Ichijo, S., Osame,S., and Sarashina, T.

(1989) Clinical and biochemical findings in bovine

cerebrocortical necrosis produced by oral adminis-

tration of amprolium, Nihon Juigaku Zasshi, 51, 79-85,

https://doi.org/10.1292/jvms1939.51.79.

19. Frye, T.M., Williams, S.N., and Graham, T.W. (1991)

Vitamin deficiencies in cattle, Vet. Clin. North Am.

Food Anim. Pract., 7, 217-275, https://doi.org/10.1016/

s0749-0720(15)30817-3.

GRAF et al.1708

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

20. Tanwar, R. K., Malik, K. S., and Gahlot, A. K. (1994)

Polioencephalomalacia induced with amprolium

in buffalo calves – clinicopathologic findings, Zen-

tralbl. Veterinarmed. A, 41, 396-404, https://doi.org/

10.1111/j.1439-0442.1994.tb00106.x.

21. Nikov, S., Ivanov, I. T., Simeonov, S. P., Stoikov, D.,

and Iordanova,V. (1983) Cerebrocortical necrosis in

calves, sheep and goats, Vet. Med. Nauki, 20, 58-67.

22. Thornber, E. J., Elliott, L. E., Kerr, D., Marriott,

J. M., and Massera, F. C. (1983) Thiamin inadequa-

cy in infants: lack of evidence of amprolium in egg

yolk, Aust. N. Z. J. Med., 13, 51-52, https://doi.org/

10.1111/j.1445-5994.1983.tb04549.x.

23. Moraes, J. O., Rodrigues, S.D.C., Pereira, L.M., Me-

deiros, R.C.N., de Cordova, C.A.S., and de Cordova,

F. M. (2018) Amprolium exposure alters mice behav-

ior and metabolism in vivo, Animal Model Exp. Med.,

1, 272-281, https://doi.org/10.1002/ame2.12040.

24. Lemos, C., Faria, A., Meireles, M., Martel, F., Mon-

teiro, R., and Calhau, C. (2012) Thiamine is a sub-

strate of organic cation transporters in Caco-2 cells,

Eur.J. Pharmacol., 682, 37-42, https://doi.org/10.1016/

j.ejphar.2012.02.028.

25. Aleshin, V.A., Mkrtchyan, G.V., and Bunik, V.I. (2019)

Mechanisms of non-coenzyme action of thiamine:

protein targets and medical significance, Biochem-

istry (Moscow), 84, 829-850, https://doi.org/10.1134/

S0006297919080017.

26. Liang, X., Chien, H.C., Yee, S. W., Giacomini, M. M.,

Chen, E.C., Piao, M., Hao,J., Twelves,J., Lepist, E.I.,

Ray, A.S., and Giacomini, K.M. (2015) Metformin is a

substrate and inhibitor of the human thiamine trans-

porter, THTR-2 (SLC19A3), Mol. Pharm., 12, 4301-4310,

https://doi.org/10.1021/acs.molpharmaceut.5b00501.

27. Oliveira, W. H., Nunes, A. K., Franca, M. E., Santos,

L.A., Los, D.B., Rocha, S.W., Barbosa, K.P., Rodrigues,

G.B., and Peixoto, C.A. (2016) Effects of metformin on

inflammation and short-term memory in streptozoto-

cin-induced diabetic mice, Brain Res., 1644, 149-160,

https://doi.org/10.1016/j.brainres.2016.05.013.

28. Gould, T. D., Dao, D. T., and Kovacsics, C. E. (2009)

The Open Field Test, in Mood and Anxiety Related

Phenotypes in Mice (Gould, T. D., ed.) Humana Press/

Springer Nature, pp.1-20, https://doi.org/10.1007/978-

1-60761-303-9_1.

29. Aleshin, V.A., Graf, A.V., Artiukhov, A.V., Boyko, A.I.,

Ksenofontov, A.L., Maslova, M.V., Nogues,I., di Salvo,

M. L., and Bunik, V. I. (2021) Physiological and bio-

chemical markers of the sex-specific sensitivity to ep-

ileptogenic factors, delayed consequences of seizures

and their response to vitamins B1 and B6 in a rat

model, Pharmaceuticals (Basel), 14, 737, https://doi.org/

10.3390/ph14080737.

30. Graf, A. V., Maslova, M. V., Artiukhov, A. V., Kseno-

fontov, A. L., Aleshin, V. A., and Bunik, V. I. (2022)

Acute prenatal hypoxia in rats affects physiology and

brain metabolism in the offspring, dependent on sex

and gestational age, Int. J. Mol. Sci., 23, 2579, https://

doi.org/10.3390/ijms23052579.

31. De La Haba, G., Leder, I. G., and Racker, E. (1955)

Crystalline transketolase from bakers’ yeast: isolation

and properties, J. Biol. Chem., 214, 409-426, https://

doi.org/10.1016/S0021-9258(18)70978-2.

32. Aleshin, V.A., Kaehne,T., Maslova, M.V., Graf, A.V.,

and Bunik, V.I. (2024) Posttranslational acylations of

the rat brain transketolase discriminate the enzyme

responses to inhibitors of ThDP-dependent enzymes

or thiamine transport, Int.J. Mol. Sci., 25, 917, https://

doi.org/10.3390/ijms25020917.

33. Hinman, L.M., and Blass, J.P. (1981) An NADH-linked

spectrophotometric assay for pyruvate dehydroge-

nase complex in crude tissue homogenates, J. Biol.

Chem., 256, 6583-6586, https://doi.org/10.1016/S0021-

9258(19)69029-0.

34. Schwab, M. A., Kolker,S., van den Heuvel, L.P., Sau-

er,S., Wolf, N.I., Rating,D., Hoffmann, G.F., Smeitink,

J.A., and Okun, J. G. (2005) Optimized spectrophoto-

metric assay for the completely activated pyruvate de-

hydrogenase complex in fibroblasts, Clin. Chem., 51,

151-160, https://doi.org/10.1373/clinchem.2004.033852.

35. Artiukhov, A. V., Graf, A.V., Kazantsev, A. V., Boyko,

A.I., Aleshin, V.A., Ksenofontov, A.L., and Bunik, V.I.

(2022) Increasing inhibition of the rat brain 2-oxog-

lutarate dehydrogenase decreases glutathione redox

state, elevating anxiety and perturbing stress adapta-

tion, Pharmaceuticals (Basel), 15, 182, https://doi.org/

10.3390/ph15020182.

36. Ksenofontov, A. L., Boyko, A. I., Mkrtchyan, G. V.,

Tashlitsky, V.N., Timofeeva, A.V., Graf, A.V., Bunik,

V.I., and Baratova, L.A. (2017) Analysis of free ami-

no acids in mammalian brain extracts, Biochemis-

try (Moscow), 82, 1183-1192, https://doi.org/10.1134/

S000629791710011X.

37. Hagberg,H., Lehmann,A., Sandberg,M., Nystrom,B.,

Jacobson,I., and Hamberger, A.V. (1985) Ischemia-in-

duced shift of inhibitory and excitatory amino acids

from intra- to extracellular compartments, J. Cereb.

Blood Flow Metab., 5, 413-419, https://doi.org/10.1038/

jcbfm.1985.56.

38. Ellison, D. W., Beal, M. F., and Martin, J. B. (1987)

Phosphoethanolamine and ethanolamine are de-

creased in Alzheimer’s disease and Huntington’s dis-

ease, Brain Res., 417, 389-392, https://doi.org/10.1016/

0006-8993(87)90471-9.

39. Suarez, L.M., Munoz, M. D., Martin Del Rio,R., and

Solis, J.M. (2016) Taurine content in different brain

structures during ageing: effect on hippocampal syn-

aptic plasticity, Amino Acids, 48, 1199-1208, https://

doi.org/10.1007/s00726-015-2155-2.

40. Turner, O., Phoenix, J., and Wray, S. (1994) Devel-

opmental and gestational changes of phosphoeth-

anolamine and taurine in rat brain, striated and

METFORMIN, AMPROLIUM, AND AMINO ACIDS IN THE RAT BRAIN 1709

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

smooth muscle, Exp. Physiol., 79, 681-689, https://

doi.org/10.1113/expphysiol.1994.sp003800.

41. Artiukhov, A.V., Aleshin, V.A., Karlina, I. S., Kazant-

sev, A.V., Sibiryakina, D.A., Ksenofontov, A.L., Luka-

shev, N.V., Graf, A.V., and Bunik, V.I. (2022) Phospho-

nate inhibitors of pyruvate dehydrogenase perturb

homeostasis of amino acids and protein succinyla-

tion in the brain, Int. J. Mol. Sci., 23, 13186, https://

doi.org/10.3390/ijms232113186.

42. Artiukhov, A.V., Pometun, A.A., Zubanova, S.A., Tish-

kov, V.I., and Bunik, V.I. (2020) Advantages of formate

dehydrogenase reaction for efficient NAD(+) quanti-

fication in biological samples, Anal. Biochem., 603,

113797, https://doi.org/10.1016/j.ab.2020.113797.

43. Kochetov, G. A. (1982) Transketolase from yeast, rat

liver, and pig liver, Methods Enzymol., 90, 209-223,

https://doi.org/10.1016/s0076-6879(82)90128-8.

44. Artiukhov, A. V., Solovjeva, O. N., Balashova, N. V.,

Sidorova, O. P., Graf, A. V., and Bunik, V. I. (2024)

Pharmacological doses of thiamine benefit patients

with the Charcot–Marie–Tooth neuropathy by chang-

ing thiamine diphosphate levels and affecting regu-

lation of thiamine-dependent enzymes, Biochemis-

try (Moscow), 89, 1161-1182, https://doi.org/10.1134/

S0006297924070010.

45. Huang, H. M., Chen, H. L., and Gibson, G. E. (2010)

Thiamine and oxidants interact to modify cellular

calcium stores, Neurochem. Res., 35, 2107-2116, https://

doi.org/10.1007/s11064-010-0242-z.

46. Karuppagounder, S. S., Xu, H., Shi, Q., Chen, L. H.,

Pedrini, S., Pechman, D., Baker, H., Beal, M. F., Gan-

dy, S. E., and Gibson, G. E. (2009) Thiamine defi-

ciency induces oxidative stress and exacerbates the

plaque pathology in Alzheimer’s mouse model, Neu-

robiol. Aging, 30, 1587-1600, https://doi.org/10.1016/

j.neurobiolaging.2007.12.013.

47. Nicoletti, V.G., Santoro, A.M., Grasso,G., Vagliasindi,

L.I., Giuffrida, M.L., Cuppari,C., Purrello, V.S., Stella,

A. M., and Rizzarelli, E. (2007) Carnosine interac-

tion with nitric oxide and astroglial cell protection,

J.Neurosci. Res., 85, 2239-2245, https://doi.org/10.1002/

jnr.21365.

48. Ke, C. J., He, Y. H., He, H. W., Yang, X., Li, R., and

Yuan, J. (2014) A new spectrophotometric assay for

measuring pyruvate dehydrogenase complex activi-

ty: a comparative evaluation, Anal. Methods, 6, 6381-

6388, https://doi.org/10.1039/c4ay00804a.

49. Roelofs, K. (2017) Freeze for action: neurobiological

mechanisms in animal and human freezing, Philos.

Trans. R. Soc. Lond. B Biol. Sci., 372, 20160206, https://

doi.org/10.1098/rstb.2016.0206.

50. Page, S. W. (2008) Antiparasitic drugs, in Small Animal

Clinical Pharmacology (Maddison, J. E., Page, S. W.,

and Church, D.B., eds) Saunders Ltd., Philadelphia,

pp. 198-260, https://doi.org/10.1016/B978-070202858-

8.50012-9.

51. Bunik, V.I., Artiukhov, A.V., Kazantsev, A.V., Aleshin,

V.A., Boyko, A.I., Ksenofontov, A.L., Lukashev, N.V.,

and Graf, A.V. (2022) Administration of phosphonate

inhibitors of dehydrogenases of 2-oxoglutarate and

2-oxoadipate to rats elicits target-specific metabolic

and physiological responses, Front. Chem., 10, 892284,

https://doi.org/10.3389/fchem.2022.892284.

52. Sambon, M., Pavlova, O., Alhama-Riba, J., Wins, P.,

Brans,A., and Bettendorff, L. (2022) Product inhibi-

tion of mammalian thiamine pyrophosphokinase is

an important mechanism for maintaining thiamine

diphosphate homeostasis, Biochim. Biophys. Acta

Gen. Subj., 1866, 130071, https://doi.org/10.1016/j.

bbagen.2021.130071.

53. Bunik, V. I., Raddatz, G., and Strumilo, S. A. (2013)

Translating enzymology into metabolic regulation:

the case of the 2-oxoglutarate dehydrogenase mul-

tienzyme complex, Curr. Chem. Biol., 7, 74-93, https://

doi.org/10.2174/2212796811307010008.

54. Graf,A., Trofimova,L., Ksenofontov,A., Baratova,L.,

and Bunik, V. (2020) Hypoxic adaptation of mito-

chondrial metabolism in rat cerebellum decreases

in pregnancy, Cells, 9, 139, https://doi.org/10.3390/

cells9010139.

55. Kazyken, D., Dame, S.G., Wang, C., Wadley, M., and

Fingar, D. C. (2024) Unexpected roles for AMPK in

the suppression of autophagy and the reactivation

of MTORC1 signaling during prolonged amino acid

deprivation, Autophagy, 20, 2017-2040, https://doi.org/

10.1080/15548627.2024.2355074.

56. Barnaba,C., Broadbent, D.G., Kaminsky, E.G., Perez,

G.I., and Schmidt, J.C. (2024) AMPK regulates phago-

phore-to-autophagosome maturation, J.Cell Biol., 223,

e202309145, https://doi.org/10.1083/jcb.202309145.

57. Seliger, C., Rauer, L., Wuster, A. L., Moeckel,S., Leid-

gens, V., Jachnik, B., Ammer, L. M., Heckscher, S.,

Dettmer,K., Riemenschneider, M.J., Oefner, P.J., Proe-

scholdt,M., Vollmann-Zwerenz,A., and Hau,P. (2023)

Heterogeneity of amino acid profiles of proneural and

mesenchymal brain-tumor initiating cells, Int. J. Mol.

Sci., 24, 3199, https://doi.org/10.3390/ijms24043199.

58. Welch,N., Singh, S.S., Kumar,A., Dhruba, S.R., Mish-

ra,S., Sekar, J., Bellar, A., Attaway, A. H., Chelluboy-

ina, A., Willard, B. B., Li, L., Huo, Z., Karnik, S. S.,

Esser,K., Longworth, M.S., Shah, Y. M., Davuluri, G.,

Pal,R., and Dasarathy,S. (2021) Integrated multiom-

ics analysis identifies molecular landscape pertur-

bations during hyperammonemia in skeletal muscle

and myotubes, J. Biol. Chem., 297, 101023, https://

doi.org/10.1016/j.jbc.2021.101023.

59. Fitzpatrick, S. M., Hetherington, H. P., Behar, K. L.,

and Shulman, R.G. (1989) Effects of acute hyperam-

monemia on cerebral amino acid metabolism and

pHi in vivo, measured by 1H and 31P nuclear mag-

netic resonance, J. Neurochem., 52, 741-749, https://

doi.org/10.1111/j.1471-4159.1989.tb02517.x.

GRAF et al.1710

BIOCHEMISTRY (Moscow) Vol. 89 No. 10 2024

60. Haberle,J. (2013) Clinical and biochemical aspects of

primary and secondary hyperammonemic disorders,

Arch. Biochem. Biophys., 536, 101-108, https://doi.org/

10.1016/j.abb.2013.04.009.

61. Boyko, A. I., Artiukhov, A. V., Kaehne, T., di Salvo,

M. L., Bonaccorsi di Patti, M. C., Contestabile, R.,

Tramonti,A., and Bunik, V.I. (2020) Isoforms of the

DHTKD1-encoded 2-oxoadipate dehydrogenase, iden-

tified in animal tissues, are not observed upon the

human DHTKD1 expression in bacterial or yeast

systems, Biochemistry (Moscow), 85, 920-929, https://

doi.org/10.1134/S0006297920080076.

62. Artiukhov, A. V., Grabarska, A., Gumbarewicz, E.,

Aleshin, V. A., Kahne, T., Obata, T., Kazantsev, A.V.,

Lukashev, N.V., Stepulak,A., Fernie, A.R., and Bunik,

V.I. (2020) Synthetic analogues of 2-oxo acids discrim-

inate metabolic contribution of the 2-oxoglutarate and

2-oxoadipate dehydrogenases in mammalian cells

and tissues, Sci. Rep., 10, 1886, https://doi.org/10.1038/

s41598-020-58701-4.

63. Fan,J., Li,D., Chen, H.S., Huang, J.G., Xu, J. F., Zhu,

W. W., Chen, J. G., and Wang, F. (2019) Metformin

produces anxiolytic-like effects in rats by facilitat-

ing GABA(A) receptor trafficking to membrane, Br. J.

Pharmacol., 176, 297-316, https://doi.org/10.1111/

bph.14519.

64. Li, G.F., Zhao,M., Zhao,T., Cheng,X., Fan,M., and Zhu,

L.L. (2019) Effects of metformin on depressive behav-

ior in chronic stress rats, Zhongguo Ying Yong Sheng

Li Xue Za Zhi, 35, 245-249, https://doi.org/10.12047/

j.cjap.5775.2019.052.

Publisher’s Note. Pleiades Publishing remains

neutral with regard to jurisdictional claims in published

maps and institutional affiliations. AI tools may have

been used in the translation or editing of this article.