Ni

Ca

b

a

ARA

KCNMB

1

cfPpLniFswaLDpih

0h

Experimental and Toxicologic Pathology 65 (2013) 941– 948

Contents lists available at SciVerse ScienceDirect

Experimental and Toxicologic Pathology

jo ur nal homepa ge: www.elsev ier .de /e tp

ew insight in colistin induced neurotoxicity with the mitochondrial dysfunctionn mice central nervous tissues

hongshan Daia, Jichang Lia,∗, Jian Lib

College of Veterinary Medicine, Northeast Agricultural University, 59 Mucai Street, Xiangfang District, Harbin 150030, PR ChinaFacility for Anti-infective Drug Development and Innovation, Monash Institute of Pharmaceutical Sciences, Monash University, 381 Royal Parade, Parkville, Victoria 3052, Australia

r t i c l e i n f o

rticle history:eceived 24 October 2012ccepted 8 January 2013

eywords:olistin sulfateeurotoxicityitochondrial dysfunction

ehavioral test

a b s t r a c t

In the present study, the mechanism of colistin-induced neurotoxicity was investigated with a focus onbehavioral characters, mitochondrial ultrastructures and functions of the central nerve tissues in micefollowed by administrating intravenously 15 (divided into two dose and 12 h apart), 7.5 and 5 mg/kg bwcolistin sulfate for 1, 3 or 7 days successively. To assess the recoverability of colistin-induced neurotoxic-ity, the neurotoxicity was also examined on day 15 (8 post colistin sulfate administration for 7 days). Theresults showed that, the spontaneous activities of mice were significantly decreased on days 3 and 7 in the15 mg/kg group compared with the correspondingly control group. The abnormal ultrastructure changesof mitochondria were presented in their nervous tissues and changed in a dose- and time-dependentmanner, e.g. severe vacuolation and fission on days 3 and 7 in the 15 mg/kg group and more slight on day7 in the 7.5 mg/kg group. In addition, mitochondrial permeability transition (MPT), membrane potential

(� m) and activities of mitochondrial succinate dehydrogenase changed, showing that colistin affectedthe mitochondrial functions. The recoverability of colistin-induced neurotoxicity was showed and onlyslight injury occurred in the nerve tissues of mice on day 15 in the 15 mg/kg group and it had no abnor-mal changes in the behavioral and neuropathology characters in mice on day 15 in the 7.5 and 5 mg/kggroups. The results suggested that mitochondrial dysfunction might partly account for the mechanismof neurotoxicity induced by colistin sulfate.

. Introduction

In recent ten years, colistin (also known as polymyxin E) islinically being used increasingly as a last-line therapeutic optionor infections of multidrug resistant gram-negative bacteria, e.g.seudomonas aeruginosa, Acinetobacter baumannii, and Klebsiellaneumoniae (Hermsen et al., 2003; Falagas and Kasiakou, 2006;andman et al., 2008; Lim et al., 2010); It is possible of inducingeurotoxicity in human beings and sensitive animals, affecting

ts optimal therapeutic efficacy in clinic (Landman et al., 2000;alagas and Kasiakou, 2006; Lim et al., 2010). The neurotoxicigns and symptoms include apathy, ataxia, myopathy, musculareakness, polymyoneuropathy, facial and peripheral paresthesia

nd even neuromuscular blockade (Landman et al., 2000, 2008;in et al., 2005; Falagas and Kasiakou, 2006; Wabhy et al., 2010;ai et al., 2012a). Though several literatures reported that the

roposed mechanism of colistin neurotoxicity was a noncompet-

tive myoneuronal presynaptic blockade of acetylcholine release,ypocalcemia-induced prolongation of depolarization, or the

∗ Corresponding author. Tel.: +86 451 5519 0674; fax: +86 451 5519 0363.E-mail address: lijichang828@yahoo.cn (J. Li).

940-2993/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.etp.2013.01.008

© 2013 Elsevier GmbH. All rights reserved.

interaction of colistin with neurons owning high lipid contentinduced neurotoxicity (Duncan, 1973; Falagas and Kasiakou, 2006),the proper mechanism is unknown. In our pervious study (Dai et al.,2012a,b), the ultrastructural studies of sciatic nerves showed thataxonal degeneration and demyelization, and the overtly abnormalelectrophysiological characterizations of sciatic nerves occurredwhen the female mice were treated with colistin sulfate (15 mg/kgper day) intravenously for 3 and 7 days successively. Mitochondriaare important regulators of neuronal function in addition totheir role in energy production, and their dysfunction is a keydeterminant of neurodegeneration (Mancuso et al., 2006; Masoudet al., 2009), such as some neurodegenerative diseases includingHuntington’s disease, Parkinson’s disease, Alzheimer’s disease, orneurotoxicity induced by some chemotherapy agents including thepaclitaxel and cisplatin (Lin and Beal, 2006; Mancuso et al., 2006;Melli et al., 2008; Jewel et al., 2011). Up to now, little is known aboutthe correlations between the mitochondrial dysfunction and neu-rotoxicity induced by colistin. So we built the animal models of thecolistin-induced neurotoxicity by the behavioral tests in mice, fol-

lowed by examination of the ultrastructural and functional changesof the mitochondria in central nervous tissue (CNS) in a dose- andtime-dependent manner, to illustrate whether mitochondrial dys-function account for the development of the colistin neurotoxicity.

9 xicolo

2

2

0r(sBoTCSw

2

RwfavAmaoio

2

ctmsatacdw5w

iwwsTsa

TT

42 C. Dai et al. / Experimental and To

. Materials and methods

.1. Chemicals, solvents and reagents

Colistin sulfate was purchased from Sigma–Aldrich (Lot:95K1048, St. Louis, MO, USA; 20195 U/mg). Rhodamine 123,otenone, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acidHEPES) were purchased from Sigma Chemical Co. (St. Louis, Mis-ouri, USA). BCATM protein assay kit was purchased from Wuhanoster Bio-engineering Limited Co. (Wuhan, Hubei, China). Allther chemicals were of highest quality commercially available.he ZZ-6 mice spontaneous activity apparatus was produced byhengdu Taimeng Science and Technology, Co., Ltd. (Chengdu,ichuan, China). JSM25610LV transmission electron microscopeas produced by JEOL Ltd. (JEOL Ltd., Japan).

.2. Animals

Female Kunming mice (supplied by the Harbin Veterinaryesearch Institute, China) at 8 week of age and weighing 18–20 gere used in these experiments, were kept in separate cages, had

ree access to feed and water and were acclimated to housingnd environmental conditions (temperature, humidity, light, andentilation) for at least 1 week prior to beginning of treatments.nimals were housed in a limited access animal facility where ani-al room temperature and relative humidity were set at 22 ± 2 ◦C

nd 55 ± 10%, respectively. Artificial lighting provided a 24 h cyclef 12 h light/12 h dark (light 7 a.m.–7 p.m.). All procedures involv-ng animals were approved by the Animal Care and Use Committeef Heilongjiang Province, PR China.

.3. Experimental design

Three experiments were set up to investigate the effect ofolistin sulfate on the behavior changes of mice (experiment 1),he structures (experiment 2) and functions (experiment 3) of

itochondria in central nerve tissues including the brains andpinal cords. Doses of 15, 7.5 and 5 mg/kg colistin sulfate weredministered intravenously as 10 ml/kg solutions for a successivereatment period of 1, 3 and 7 days, respectively in each group,nd the dose regimens were showed in Table 1. For the 15 mg/kgolistin sulfate group, the dose of 15 mg/kg was divided into twoosages and 12 h apart, i.e. 7.5 mg/kg/12 h, which was consistentith the pervious study (Dai et al., 2012a). And for the 7.5 and

mg/kg groups, the dose was once a day. Those in the control groupere given 0.9% normal saline of the same volume.

In the experiment 1, the effects of colistin sulfate on the behav-oral changes including the spontaneous activity and gait scores

ere studied. Forty mice were randomly divided into four groupsith 10 mice in each group. Before the mice administration, the

pontaneous activity and gait scores were tested as the base line.hen the mice were administrated for 7 days continuously. Thepontaneous activity and gait scores were tested on days 1, 3 and 7fter the first administration. To observe the reversibility of the

able 1he dose regimens in all the experiments.

Doses (mg/kg) Times of administration Times of investigati

15a Successive 1 to 7 days inexperiment 1 and 1, 3 and 7days, respectively inexperiments 2 and 3

On days 1, 3, 7, and

after the firstadministrationb

7.550

a 15 mg/kg per day was divided into two doses, 12 h apart (Dai et al., 2012a,b).b Day 15, i.e. 8 days after the mice were administrated colistin sulfate for 7 d.c Days 1, 3, 7, and 15 were based on the days of the first dose as day 0.d It is a convenient intravenous procedure in mice.

gic Pathology 65 (2013) 941– 948

neurotoxicity induced by colistin sulfate, on day 15, i.e. on the8th day after 7 days of successive administration, these behavioralindexes were also been tested.

In the experiment 2, fifty two mice were randomly divided intofour groups, which were randomly divided into 15 mg/kg (n = 16),7.5 mg/kg (n = 16) and 5 mg/kg (n = 16) colistin sufate-given groupsand the control group (n = 4). For the colistin sulfate groups, themice were administrated colistin sulfate successively for 1, 3 and7 days including 4, 4 and 8 mice, respectively, and on days 1, 3, 7and 15 after the first administration, they (four at each time point)were respectively anesthetized by intraperitoneal injection of anoverdose of sodium pentobarbital (80 mg/kg bw), perfused trans-cardially with 100 mL 0.9% saline and 1000 mL 0.1 M phosphatebuffer (PB; pH 7.4) containing 4% (w/v) paraformaldehyde. Thecerebral cortex and spinal cord were obtained and used to observethe changes of mitochondrial ultrastructures.

In the experiment 3, one hundred and sixty mice were randomlydivided into four groups, i.e. 15, 7.5, 5 mg/kg colistin sufate-givengroups and the control group (n = 40 in each group). Ten mice ineach group were randomly selected and sacrificed on the corre-sponding time points of days 1, 3, 7 and 15, respectively (n = 10 ineach time point). The cerebral cortex and spinal cord were quicklydissected and prepared for mitochondrial functional assay.

2.4. Determination of neurological behavior

2.4.1. Gait scoresIn the experiment 1, on days 1, 3, 7 and 15 after the first intra-

venous administrations of colistin sulfate, the gait abnormalitieswere measured as the previous study (Dai et al., 2012a,b). The testedmice were placed in a clear plexiglas box and were observed for3 min each time, the assigned scores were as follows: 1 = normalgait; 2 = slightly abnormal gait (slight ataxia, hopping gait and footsplay); 3 = moderately abnormal gait (obvious ataxia and foot splaywith limb abduction during ambulation); 4 = severely abnormalgait (inability to support body weight and foot splay).

2.4.2. Spontaneous activitiesThe tested mice were placed into the mice spontaneous activity

apparatus, 1 mouse in each cabin, the times of spontaneous activ-ities and the times of arising within 10 min were recorded afteradaptation for 5 min (Li, 1992).

2.5. Observation of mitochondrial ultrastructures

In the experiment 2, on days 1, 3, 7 and 15, the cerebral cortexand spinal cord were obtained from 4 mice for the ultrastructuralstudies in each group; the method of ultrastructure observationwas as our previous method (Dai et al., 2012a,b). Small pieces of tis-sue were dissected from the cerebral cortex and spinal cord and cut

into approximately 1 mm cubes, then were fixed by immersion with2.5% glutaraldehyde in 0.12 mM phosphate buffer (pH 7.2), andkept overnight at 4 ◦C in the fixative. Post-fixation was performedwith 2% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.4) for

onc The volume of injection Route of administration

15 10 mL/kg Administeredintravenously in thetail veind

xicologic Pathology 65 (2013) 941– 948 943

2t(lm

2

2

s(w1wfc0pmE3p1fict

2(

a1a5isi

2

iaodKuCb

*

*

Fig. 1. The gait scores changes in mice treated by colistin sulfate. The gait scoreswere determined in mice on days 1, 3, 7 and 15 after intravenous treatment ofcolistin sulfate. The results were presented as mean ± SD (n = 10, on days 0, 1, 3, 7,

Fdwc

C. Dai et al. / Experimental and To

h at 4 ◦C. The specimens were then dehydrated in graded ace-one solutions and embedded in araldite-epon. Ultrathin sections50 nm) were prepared and counterstained with uranyl acetate andead citrate, examined with the JSM25610LV transmission electron

icroscope.

.6. Measurement of mitochondrial function

.6.1. Isolation and purification of the mitochondriaIn the experiment 3, mitochondria of the mice central nerve tis-

ues were isolated by the method described by Clark and Nicklas1970). Briefly, the dissected cerebrum and spinal cord samplesere homogenized (10% w/v) in ice-cold buffer (0.25 M sucrose,

0 mM Tris–HCl, 0.5 mM K+-EDTA, pH 7.4). Then the homogenateas centrifuged at 2000 × g for 4 min at 4 ◦C, the supernatant was

urther centrifuged at 12,500 × g for 8 min at 4 ◦C. The crude mito-hondrial pellet was resuspended in a 3% Ficoll medium (3% Ficoll,.12 M mannitol, 0.03 M sucrose, 25 mM K+-EDTA, 5 mM Tris–HCl,H 7.4), this suspension was carefully layered onto a 6% Ficolledium (6% Ficoll, 0.24 M mannitol, 0.06 M sucrose, 50 mM K+-

DTA, 10 mM Tris-HCl, pH7.4) and centrifuged at 11,500 × g for0 min at 4 ◦C, the supernatant was decanted and the mitochondrialellet was resuspended in isolation medium and recentrifuged at2,500 × g for 10 min at 4 ◦C to get the purified mitochondria. Thenal pellet was reconstituted in the isolation medium. The proteinoncentration of the mitochondrial suspension was quantified byhe BCATM protein assay kit.

.6.2. Measurement of mitochondrial permeability transitionMPT)

MPT is associated with mitochondrial swelling, which wasssayed by the absorbance at 540 nm (A540 nm) (Uyemura et al.,997). The isolated mitochondria were modulated to 0.5 mg/mLnd incubated in the assay buffer (125 mM sucrose, 65 mM KCl,

mM succinate, 5 mM rotenone, 10 mM Tris-HCl, pH 7.4). MPT wasnitiated by adding 50 �M calcium chloride, and monitored by mea-uring the A540 nm at 37 ◦C using an ultraviolet spectrophotometern 5 min.

.6.3. Detection of mitochondrial membrane potential (� m)Changes in mitochondrial membrane potential were monitored

n the presence of the fluorescent dye Rhodamine 123 (Rh123)ccording to Emaus et al. (1986a,b) with modifications. Briefly, flu-rescence with excitation at 503 nm and emission at 527 nm wasetected in a reaction buffer (250 mM sucrose, 2 mM HEPES, 0.5 mM

H2PO4, 4.2 mM sodium succinate, pH 7.4, and 0.3 mM Rh123)sing F-4500FL Spectrophotometer (Hitachi High-Technologieso., Japan). The mitochondria were diluted to 0.5 mg/mL in theuffer and incubated for 3 min. The fluorescence was recorded

ig. 2. Influence of colistin sulfate on the spontaneous activities of mice. (A) Times of spoays 1, 3, 7 and 15 after intravenous treatment of colistin sulfate. The results were presenith ANOVA, followed by LSDs post hoc tests, and significant statistical difference was in

ontrol.

15). Statistical analysis was performed with ANOVA, followed by LSDs post hoc tests,and significant statistical difference was indicated by *p < 0.05 with respect to thecorresponding time-matched control.

again, and the alteration of the � m was detected by the decreaseof the fluorescence.

2.6.4. Activities of the mitochondrial respiratory chainMTT reduction was used to assess the activities of the mitochon-

drial respiratory chain by the method of Liu et al. (1997). The dyewas converted to water-insoluble purple formazan on the reduc-tive cleavage of its tetrazolium ring by the succinate dehydrogenasesystem of the active mitochondria. The reaction mixture containingmitochondrial preparation (80–100 mg protein) and 0.02 mL MTT(0.1 mg/mL) was incubated at 37 ◦C for 30 min and then centrifugedat 1000 × g for 5 min at room temperature to obtain the formazanpellet. The pellet was dissolved in 1 mL acidic isopropanol and themixture was recentrifuged at 1000 × g for 5 min at room temper-ature. Then the absorbance of the supernatant was measured at595 nm. Results were expressed as A595 nm/mg protein.

2.7. Statistical analyses

Results are represented as mean ± standard deviations (SD). Thedata were analyzed with one-way analysis of variance (ANOVA),followed by LSD’s post hoc tests to compare the control andtreatment groups. The differences were considered significantly atp < 0.05 level.

3. Results

3.1. Gait scores

On days 0, 1, 3, 7 and 15, the gait scores were measured and thedata were shown in Fig. 1. Some mice showed slightly abnormal gaiton day 3 in the 15 mg/kg group (mean clinical score = 1.23 ± 0.44).

ntaneous activities; (B) times of arising within 10 min were determined in mice onted as mean ± SD (n = 10, on days 0, 1, 3, 7, 15). Statistical analysis was performed

dicated by *p < 0.05 and **p < 0.01 with respect to the corresponding time-matched

944 C. Dai et al. / Experimental and Toxicologic Pathology 65 (2013) 941– 948

Fig. 3. Electron micrographs of cerebrum in mice showing morphological changes of mitochondria in neuron in experimental groups on days 1, 3, 7 and 15 after exposure tocolistin sulfate and mice in control group (n = 4). (A) control; (B–F) respectively represented the mitochondria morphological changes on days 1(B), 3(C), 7(D and E) and 15(F)in the 15 mg/kg group; (G and H) respectively represented the mitochondria morphological changes on day 7 in the 7.5 and 5 mg/kg groups. The black arrows representedm es. Frt unkenm

Cp(gtcaoc

3

wt

itochondria. In A, B and G, the structures of mitochondria had no abnormal changhe mitochondrial cristae ruptured, decreased and even disappeared. In E, dark shr

itochondria cristae was found.

linical signs progressed to disturbances with slight ataxia, hop-ing gait and foot splay, and signs were severe on day 7 thereaftermean clinical score = 2.2 ± 0.63) in the 15 mg/kg group, and theait scores of two mice were 3. In the 7.5 mg/kg group, mice begano take on slightly abnormal gait on day 7 thereafter (mean clini-al score = 1.3 ± 0.48). On day 15, forty percent of the mice showedbnormal gait in the 15 mg/kg group. In contrast, little clinical signsf the neurotoxicity were observed in mice in the 5 mg/kg andontrol groups.

.2. Changes of spontaneous activities

On days 0, 1, 3, 7 and 15, the changes of spontaneous activitiesere shown in Fig. 2. In the 15 mg/kg group, markedly change of the

imes of spontaneous activities and times of arising within 10 min

om C to D, the progressive morphological changes of mitochondria were observed, neuron with swollen mitochondria was observed. In F and H, slightly decrease of

were showed on days 3 and 7 (p < 0.05 and p < 0.01, respectively),and the changes recovered on day 15 compared with the control.However, there were no significant changes in 7.5 and 5 mg/kggroups during the experiment period.

3.3. Ultrastructure changes of mitochondria

3.3.1. Ultrastructure changes of mitochondria in the cerebralcortex

In the cerebral cortex, progressive morphological changes ofmitochondria were observed and showed in Fig. 3. In the 15 mg/kg

group, decreases of mitochondrial cristae were shown on day 3(Fig. 3C), and more severe pathology changes including the swollenand vacuolated mitochondria, and ruptured, decreased, even disap-peared mitochondrial cristae were shown on day 7 (Fig. 3D); Some

C. Dai et al. / Experimental and Toxicologic Pathology 65 (2013) 941– 948 945

Fig. 4. Electron micrographs of spinal cord in mice showing morphological changes of mitochondria in neuron in experimental group on days 1, 3, 7 and 15 after exposure tocolistin sulfate and mice in control group (n = 4). (A) control; (B–F) respectively represented the mitochondria morphological changes on days 1(B), 3(C), 7(D and E) and 15(F)in the 15 mg/kg group; (G and H) respectively represented the mitochondria morphological changes on day 7 in the 7.5 and 5 mg/kg groups. The black arrows representedm Fromm nosisF

soccoec(

3

mgcu

itochondria. In A and G, the structures of mitochondria had no abnormal changes.itochondrial cristae ruptured, decreased and even disappeared. In E, the karyopyk

and H, slightly decrease of mitochondria cristae was found.

hrunken neuron cells with swollen mitochondria were observedn day 7 in the 15 mg/kg group (Fig. 3E), up to day 15, the pathologyhanges showed a recovery trend, only with slight mitochondrialristae decrease. In the 7.5 mg/kg group, slightly abnormal changesf mitochondria were showed only on day 7 (Fig. 3H). During thexperiment period, there was no abnormal change in the neuronells and their mitochondria in 5 mg/kg group and control groupsFig. 3A and G).

.3.2. Ultrastructural changes of mitochondria in the spinal cordIn the spinal cord, the progressive morphological changes of

itochondria were observed and shown in Fig. 4. In the 15 mg/kgroup, a little mitochondria began to show slightly abnormalhanges (Fig. 4C) on day 3, and almost all in neurons were vac-olated on day 7 (Fig. 4D), which leaded to necrosis of neuron

B to D, the progressive morphological changes of mitochondria were observed; the caused to increase its perinuclear space, which certified the necrotic in neuron. In

including karyopyknosis and increase of its perinuclear space(Fig. 4E); up to day 15, the mitochondrial structures in spinalcord showed slight abnormalities, such as mild swelling and slightdecrease of mitochondrial cristae, while the swollen endoplasmicreticulum was much more obvious than others (Fig. 4F). In the 7.5and 5 mg/kg groups, the slightly abnormal changes in the neuroncells and mitochondria occurred only on day 7 (Fig. 4H).

3.4. Effects of colistin sulfate on mitochondrial functions

3.4.1. Effects of colistin slfate on Ca2+-induced MPT

Compared with the control group, with time went on, on day

3, the A540 nm was significantly decreased by 14.7% and 22.2%(p < 0.05 and p < 0.01, respectively) in cerebrum and spinal cordin the 15 mg/kg group, which were decreased by 30.2% and 48.7%

946 C. Dai et al. / Experimental and Toxicologic Pathology 65 (2013) 941– 948

Fig. 5. Effects of colistin sulfate on Ca2+-induced mitochondrial permeability transition (MPT). (A) cerebrum; (B) spinal cord. The results were represented as the mean ± SD(n = 10). Statistical significance was determined by using one-way analysis of variance (ANOVA). *p < 0.05 and **p < 0.01 compared with the control group. Ca2+-induced MPTwere shown as the progressive drop of A540 nm.

Fig. 6. Effects of colistin sulfate on � m. (A) cerebrum; (B) spinal cord. The results were represented as a mean ± SD (n = 10). Statistical significance was determined by usingone-way analysis of variance (ANOVA). *p < 0.05, **p < 0.01 compared with the control group. Significant differences of � m were only presented in the spinal cord of thehigh dose group treated mice on days 7 and 15 (p < 0.01 and p < 0.05, respectively).

F mices nce ww ere ex

(shAcc

3

oiti

3

sdg1gd

ig. 7. The activity of mitochondrial succinate dehydrogenase by mitochondria frompinal cord. The results were represented as a mean ± SD (n = 10). Statistical significaith the control group. The activities of mitochondrial succinate dehydrogenase w

both p < 0.01) on day 7. In the 7.5 mg/kg group, the A540 nm wasignificantly decreased by 18.8% (p < 0.05) in spinal cord, while itad no significant changes (p > 0.05) in the cerebrum. On day 15, the540 nm was still decreased by 19.3% and 19.9% (both p < 0.05) inerebrum and spinal cord in the 15 mg/kg group, and no significanthanges were shown in the other groups (Fig. 5).

.4.2. Effect of colistin sulfate on the � m

� m was examined by monitoring the fluorescence quenchingf Rh123 dynamically (Fig. 6). There was no significant differencen the cerebrum, but significant differences of � m were shown inhe spinal cord on days 7 and 15 (p < 0.01 and p < 0.05, respectively)n the 15 mg/kg group.

.4.3. Activities of the mitochondrial respiratory chainThe results showed in Fig. 7. In the cerebrum, there was no

ignificant decrease in the activity of mitochondrial succinateehydrogenase on days 1 and 3; however, on day 7 in the 15 mg/kg

roup, the enzyme activities were significantly decreased by4.8% (p < 0.01) compared with the correspondingly controlroup. In the spinal cord, the activities of mitochondrial succinateehydrogenase were decreased by 9.6% (p < 0.01) and 15.7%

central nerve tissues following administration of colistin sulfate. (A) cerebrum; (B)as determined by using one-way analysis of variance (ANOVA). **p < 0.01 comparedpressed as A595 nm/mg protein.

(p < 0.01), respectively on days 3 and 7 in the 15 mg/kg group. Onday 15, it still decreased by 9.84% (p < 0.01) and 8.82% (p < 0.01)in the cerebrum and spinal cord in the 15 mg/kg group comparedwith the control group.

4. Discussion

Neuro- and nephrotoxicities are potential adverse effects thatare of greatest concern to physicians when colistin as a valu-able therapeutic option used against Gram-negative pathogens.Fewer rates of neurotoxicity (about 0–7%) occurred in clinical whenpatients received the currently recommended dosage regimensthan the nephrotoxicity (Falagas and Kasiakou, 2006; Lim et al.,2010; Yousef et al., 2011). However, most of these patients were oldand some being treated with sedatives and myorelaxants, assess-ment of the central and especially peripheral toxicity is difficult,the neurotoxicity rates may be more than the present (Nasnaset al., 2009; Wabhy et al., 2010). In our previous study, we had

confirmed that the peripheral neurotoxicity occurred when micewere intravenously administrated colistin sulfate at 15 mg/kg perday (divided into two dosages) for 3 days (Dai et al., 2012a). Andwe also found that the mitochondrial dysfunction may play an

xicolo

i2

awoaaeeiwnsi(p1p(

ct2toii2siamtw

bggeov2dcmrctcrgffofrl

obiitLo

C. Dai et al. / Experimental and To

mportant role in colistin-induced neurotoxicity in vivo (Dai et al.,012b).

In this paper, these signs including slight ataxia, hopping gaitnd foot splay were shown in the mice administrated intravenouslyith 15 mg/kg colistin sulfate, which were similar with the previ-

us study (Landman et al., 2000; Dai et al., 2012a); The spontaneousctivities in mice, and the functions and structure of mitochondrialso changed in the dose- and time-dependent manners (Landmant al., 2000; Falagas and Kasiakou, 2006; Landman et al., 2008; Limt al., 2010), which may be relative with the concentration of col-stin in the nerve tissues or serum level. When the chick and mice

ere treated subcutaneously with 40 mg/kg colistin sulfate, sig-ificant neurotoxicity was only found in chick at the maximumerum concentration of 55.6 �g/mL but no any signs of neurotox-cty in mice at the maximum serum concentration of 36.1 �g/mLLandman et al., 2000; Jin et al., 2009). When colistin is used inatients, the drug accumulation will occur, and a serum level of5 �g/mL was reported following dosing of 2.5 mg/kg (total dosageer day) for 7 days, but if the neurotoxicity occurred was not knownLandman et al., 2008).

In the previous studies, the cell apoptosis or necrosis had beenertified in colistin-induced nephrotoxicity by Wallace et al. (2008),he same results were also been certified by Yousef et al. (2011,012). Some authors suggested that colistin should induce neuro-oxicity mediated by the d-amino acid and fatty acid componentsf colisin, especially the lipid A of its model structure, which isnteracted with the high lipid neuron, as same as the colistinnduced nephrotoxicity (Liu et al., 1997; Falagas and Kasiakou,006; Nasnas et al., 2009; Ozyilmaz et al., 2011). In the presenttudy, the mitochondria ultrastructure changed evidently, signif-cant increase of Ca2+-induced MPT and obvious decrease in thectivity of mitochondrial enzymes in the central nerve tissues inice treated with 15 mg/kg colsitn sulfate for 7 days. It suggested

hat the mitochondrial dysfunction played an important role, whichas same as the precious study in vivo (Dai et al., 2012b).

The mitochondria are vital cellular machineries for maintainingasic cellular functions such as cellular energy metabolism, whichenerate ATP accompanied with the production of reactive oxy-en species (ROS) (Lin and Beal, 2006; Masoud et al., 2009; Smitht al., 1999). Consequently, oxidative stress-induced cell apoptoticr necrosis in neurons has been shown to be related with the ele-ation of ROS levels-induced mitochondrial damage (Schulz et al.,000; He et al., 2007; Tastekin et al., 2006). In this paper, theegeneration or necrosis in neurons of cerebral cortex and spinalord were presented in the 15 mg/kg group (Figs. 3 and 4). Colistinay produce amount of ROS in vivo on primary chick cortex neu-

ons because of the oxidative stress (Zhang, 2011). Excessive ROSould impair nervous system by triggering the cascade of peroxida-ive events, leading to damage of cell membranes and intracellularytoplasmic organelles, such as increased MPT, decreased � m oreduced MTT of mitochondria in this study (Figs. 5–7), which sug-ested that mitochondrial dysfunction might be partly responsibleor the development of the neurotoxicity induced by colistin sul-ate. The obvious deprementia, decrease of spontaneous activitiesr myasthenia of limbs occurred in mice treated with colistin sul-ate intravenously, which were same as Lin et al. (2005), it may beelated with the metabolism of energy resulting in reduction of ATPevel in nervous system.

It is known that mitochondrial Ca2+ homeostasis is at the centerf wide interest in the scientific community today because of its roleoth in the modulation of numerous physiological responses and

ts involvement in cell death (Giacomello et al., 2007). Alterations

n Ca2+-induced MPT and � m are now thought to be a cen-ral regulatory mechanism for cell death induction (Lopachin andehning, 1997). Once mitochondria membrane permeabilizationccurs, cells would die either by apoptosis or necrosis. When the

gic Pathology 65 (2013) 941– 948 947

excessive mitochondrial Ca2+ loads occurred, it might be incurredthe Ca2+-induced respiratory impairment by inhibition of electrontransport chain and oxidative phosphorylation, which may acti-vate the key enzymes responsible for increased ROS generation(Giacomello et al., 2007; Kaur et al., 2007). The structure of fivepositively charged amino groups in colistin sulfate, which wereconsidered as the major structure inducing toxicity in the mam-malian cell and the mechanism of its killing the bacteria, wasthought to change the membrane permeability by effecting the con-centration of Na+, K+ and Ca2+ (Landman et al., 2008). Ozyilmaz et al.(2011) reported that endothelial NOS (e-NOS), Ca2+ and calmod-ulin (CaM)-dependent activation, was deceased in kidney and lungtissue after rats were administered 10 mg/kg intraperitoneal (i.p.)colistin for 6 days. In our previous study (Dai et al., 2012b), thechanges were certified in vivo on the primary chick cortex neurons.

Especially, on day 15, the capability of self-recovery was shownfrom the behavioral tests and mitochondrial functions, which wereconsistent with some literatures, i.e., the colistin-induced neu-rotoxicity had recoverability (Hermsen et al., 2003; Falagas andKasiakou, 2006; Wabhy et al., 2010; Dai et al., 2012a,b). But upto now, the recoverability mechanism was unknown. It may berelative with the self-regulation of Bac-2 and nerve growth fac-tor (NGF), which had been confirmed in regularizing neurotoxicitycaused by diabetic or cisplatin (Hellweg and Hartung, 1990; Gilland Windebank, 1998; Schmidt et al., 2000; Sayre et al., 2008).

In this paper, the result showed obvious pathological changes,increased Ca2+-induced MPT, decrease of � m and mitochon-drial succinate dehydrogenase occurred in the mice injected withcolistin sulfate, indicating that mitochondrial dysfunction mightbe partly responsible for the development of the neurotoxicityinduced by colistin sulfate.

Conflict of interest

The authors declare that there are no conflicts of interest.

Acknowledgements

This work was supported by the National Natural Science Foun-dation of China (31272613) and the Scientific Research Fund ofHeilongjiang Provincial Education Department (12521043).

References

Clark JB, Nicklas WJ. The metabolism of rat brain mitochondria. Preparation andcharacterization. The Journal of Biological Chemistry 1970;245:4724–31.

Dai C, Li J, Lin W, Li G, Sun M, Wang F, et al. Electrophysiology and ultrastruc-tural changes in mouse sciatic nerve associated with colistin sulfate exposure.Toxicology Mechanisms and Methods 2012a;22:592–6.

Dai C, Zhang D, Li J, Li J. Effect of colistin exposure on calcium homeostasis andmitochondria functions in chick cortex neurons. Toxicology Mechanisms andMethods 2012b.

Duncan DA. Colistin toxicity. Neuromuscular and renal manifestations. Two casestreated by hemodialysis. Minnesota Medicine 1973;56:31–5.

Emaus RK, Grunwald R, Lemasters JJ. Rhodamine 123 as a probe of transmembranepotential in isolated rat-liver mitochondria: spectral and metabolic properties.Biochimica et Biophysica Acta 1986a;85:436–48.

Emaus RK, Grunwald R, Lemasters JJ. Rhodamine 123 as a probe of transmembranepotential in isolated rat-liver mitochondria: spectral and metabolic properties.Biochimica et Biophysica Acta 1986b;850:436–48.

Falagas ME, Kasiakou SK. Toxicity of polymyxins: a systematic review of the evidencefrom old and recent studies. Critical Care 2006;10:R27.

Giacomello M, Drago I, Pizzo P, Pozzan T. Mitochondrial Ca2+ as a key regulator ofcell life and death. Cell Death and Differentiation 2007;14:1267–74.

Gill JS, Windebank AJ. Cisplatin-induced apoptosis in rat dorsal root ganglion neu-rons is associated with attempted entry into the cell cycle. The Journal of ClinicalInvestigation 1998;101:2842–50.

He X, Feng L, Xiao H, Li Z, Liu N, Zhao Y, et al. Unambiguous effects of lanthanum?Toxicology Letters 2007;170:94–6.

Hellweg R, Hartung HD. Endogenous levels of nerve growth factor (NGF) are alteredin experimental diabetes mellitus: a possible role for NGF in the pathogenesisof diabetic neuropathy. Journal of Neuroscience Research 1990;26:258–67.

9 xicolo

H

J

J

K

L

L

L

L

L

L

L

L

M

M

48 C. Dai et al. / Experimental and To

ermsen ED, Sullivan CJ, Rotschafer JC, Polymyxins:. pharmacology, pharmacokine-tics, pharmacodynamics, and clinical applications. Infectious Disease Clinics ofNorth America 2003;17:545–62.

ewel LPodratz, Andrew MKnight, Lauren ETa, Nathan PStaff, Jennifer MGass, Kon-stantin Genelin, et al. Cisplatin induced mitochondrial DNA damage in dorsalroot ganglion neurons. Neurobiology of Disease 2011;41:661–8.

in L, Li J, Nation RL, Nicolazzo JA. Brain penetration of colistin in mice assessedby a novel high-performance liquid chromatographic technique. AntimicrobialAgents and Chemotherapy 2009;53:4247–51.

aur P, Radotra B, Minz RW, Gill KD. Impaired mitochondrial energy metabolismand neuronal apoptotic cell death after chronic dichlorvos (OP) exposure in ratbrain. Neurotoxicology 2007;28:1208–19.

andman D, Georgescu C, Martin DA, Quale J. Polymyxins revisited. Clinical Micro-biology Reviews 2008;21:449–65.

andman WJ, Dwars RM, Keukens HJ, Berendsen BJ. Polymyxin E-1 (colistin sulphate)(neuro-) intoxication in young ostriches (Struthio camelus spp). Avian Pathology2000;29:593–601.

i YK. Pharmaco-empirical technology of traditional Chinese medicine. Shanghai,China: Shanghai Science and Technology Publishing House; 1992. p. 330.

im LM, Ly N, Anderson D, Yang JC, Macander L, Jarkowski A, et al. Resurgence ofcolistin: a review of resistance, toxicity, pharmacodynamics, and dosing. Phar-macotherapy 2010;30:1279–91.

in B, Zhang C, Xiao X. Toxicity, bioavailability and pharmacokinetics of a newlyformulated colistin sulfate solution. Journal of Veterinary Pharmacology andTherapeutics 2005;28:349–54.

in MT, Beal MF. Mitochondrial dysfunction and oxidative stress in neurodegenera-tive disease. Nature 2006;443:787–95.

iu Y, Peterson DA, Kimura H, Schubert D. Mechanism of cellular 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction.Journal of Neurochemistry 1997;69:581–93.

opachin RM, Lehning EJ. Mechanism of calcium entry during axon injury and degen-eration. Toxicology and Applied Pharmacology 1997;143:233–44.

ancuso M, Coppede F, Migliore L, Siciliano G, Murri L. Mitochondrial dysfunc-

tion, oxidative stress and neurodegeneration. Journal of Alzheimer’s Disease2006;10:59–73.

asoud A, Kiran R, Sandhir R. Impaired mitochondrial functions in organophos-phate induced delayed neuropathy in rats. Cellular and Molecular Neurobiology2009;29:1245–55.

gic Pathology 65 (2013) 941– 948

Melli G, Taiana M, Camozzi F, Triolo D, Podini P, Quattrini A, et al. Alpha-lipoic acidprevents mitochondrial damage and neurotoxicity in experimental chemother-apy neuropathy. Experimental Neurology 2008;214:276–84.

Nasnas R, Saliba G, Hallak P. The revival of colistin: an old antibiotic for the 21stcentury. Pathologie-Biologie (Paris) 2009;57:229–35.

Ozyilmaz E, Ebinc FA, Derici U, Gulbahar O, Goktas G, Elmas C, et al. Could nephrotox-icity due to colistin be ameliorated with the use of N-acetylcysteine? IntensiveCare Medicine 2011;37:141–6.

Sayre LM, Perry G, Smith MA. Oxidative stress and neurotoxicity. Chemical Researchin Toxicology 2008;21:172–88.

Schmidt RE, Dorsey DA, Roth KA, Parvin CA, Hounsom L, Tomlinson DR. Effectof streptozotocin-induced diabetes on NGF, P75(NTR) and TrkA contentof prevertebral and paravertebral rat sympathetic ganglia. Brain Research2000;867:149–56.

Schulz JB, Lindenau J, Seyfried J, Dichgans J. Glutathione, oxidative stress and neu-rodegeneration. European Journal of Biochemistry 2000;267:4904–11.

Smith RAJ, Porteous CM, Murphy CM. Selective targeting of an antioxidant tomitochondria. European Journal of Biochemistry 1999;263:709–16.

Tastekin A, Gepdiremen A, Ors R, Buyukokuroglu ME, Halici Z. Protective effect of l-carnitine against bilirubin-induced neuronal cell death. Brain and Development2006;28:436–9.

Uyemura SA, Santos AC, Mingatto FE, Jordani MC, Curti C. Diclofenac sodiumand mefenamic acid: potent inducers of the membrane permeability transi-tion in renal cortex mitochondria. Archives of Biochemistry and Biophysics1997;342:231–5.

Wabhy K, Chopra T, Chandrasekar P. Intravenous and inhalational colistin-inducedrespiratory failure. Clinical Infectious Diseases 2010;50:e38–40.

Wallace SJ, Li J, Nation RL, Rayner CR, Taylor D, Middleton D, et al. Subacute toxicityof colistin methanesulfonate in rats: comparison of various intravenous dosageregimens. Antimicrobial Agents and Chemotherapy 2008;52:1159–61.

Yousef JM, Chen G, Hill PA, Nation RL, Li J. Ascorbic acid protects against thenephrotoxicity and apoptosis caused by colistin and affects its pharmacokine-tics. Journal of Antimicrobial Chemotherapy 2012;67:452–9.

Yousef JM, Chen G, Hill PA, Nation RL, Li J. Melatonin attenuates colistin-induced nephrotoxicity in rats. Antimicrobial Agents and Chemotherapy2011;55:4044–9.

Zhang DX. Study on the neurotoxicity and its mechanism induced by colistin in chick.Harbin, China: Northeast Agricultural University; 2011.