PAPER www.rsc.org/analyst | Analyst

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online / Journal Homepage / Table of Contents for this issue

Evaluation and discrimination of simvastatin-induced structural alterations inproteins of different rat tissues by FTIR spectroscopy and neural networkanalysis†

Sebnem Garip,a Engin Yapici,b Nihal Simsek Ozek,b Mete Severcanc and Feride Severcan*b

Received 19th July 2010, Accepted 15th October 2010

DOI: 10.1039/c0an00540a

Statins are commonly used to control hypercholesterolemia and to prevent cardiovascular diseases.

Among the statins, Simvastatin is one of the most frequently prescribed statins because of its efficacy in

reducing LDL lipoprotein cholesterol levels, its tolerability, and its reduction of cardiovascular risk and

mortality. Conflicting results have been reported with regard to benefits (pleiotropic effects) as well as

risks (adverse effects) of simvastatin on different soft and hard tissues. In the current study, Attenuated

Total Reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy was used to obtain detailed

information about protein conformational changes due to simvastatin therapy of soft tissues namely

liver, testis, sciatic nerve and hard tissues such as femur and tibia. Protein secondary structural changes

were predicted by intensity calculations from second derivative spectra and neural network (NN)

analysis, using the amide I band (1700�1600 cm�1) of FTIR spectra. Moreover, based on protein

secondary structural differences, hierarchial cluster analysis was carried out in the 1700�1600 cm�1

region. The results of our study in liver, testis and sciatic nerve tissues revealed that simvastatin

treatment significantly decreased alpha helix structure and beta sheet structure at 1638 cm�1, while

increased the anti-parallel and aggregated beta sheet and random coil structures implying

a simvastatin-induced protein denaturation in treated groups. Different to soft tissues, the results of

hard tissue studies on femur and tibia bones revealed increased alpha helix structure and decreased anti-

parallel beta sheet, aggregated beta sheet and random coil structures implying more strengthened bone

tissues in simvastatin-treated groups. Finally, the simvastatin-treated and control groups for all soft

and bone tissues were successfully differentiated using cluster analysis. According to the heterogeneity

values in the cluster analysis of these tissues, the sciatic nerve tissue was found to be the most affected

tissue from simvastatin treatment among the studied soft tissues. In addition, the high heterogeneity

value implied high secondary structural difference between control and simvastatin-treated groups in

tibia bone tissues. These findings reveal that FTIR spectroscopy with bioinformatic analyses such as

neural network and hierarchical clustering, allowed us to determine the simvastatin-induced protein

conformational changes as adverse and pleitropic effects of the drug on different soft and hard tissues.

Introduction

Statins are potent and specific inhibitors of 3-hydroxy-3-meth-

ylglutaryl-CoA(H MG-CoA) reductase, a key enzyme of

cholesterol synthesis.1 Statins lower plasma low-density lipo-

protein (LDL) cholesterol by causing intracellular cholesterol

depletion and upregulating the expression of LDL receptors.2

Apart from cholesterol, mevalonate is also the substrate for the

synthesis of nonsteroid isoprenoids including farnesylpyr-

ophosphate, geranylgeranylpyrophosphate (both attached to

aDepartment of Biochemistry, Middle East Technical University, Ankara,06531 Turkey. E-mail: feride@metu.edu.tr; Fax: +90 312 210 76 79;Tel: +90 312 210 51 66bDepartment of Biological Sciences, Middle East Technical University,Ankara, 06531, TurkeycDepartment of Electrical and Electronics Engineering, Middle EastTechnical University, Ankara, 06531, Turkey

† This article is part of a themed issue on Optical Diagnosis. This issueincludes work presented at SPEC 2010 Shedding Light on Disease:Optical Diagnosis for the New Millennium, which was held inManchester, UK June 26th–July 1st 2010.

This journal is ª The Royal Society of Chemistry 2010

small GTP-binding proteins by protein prenyltransferases),

coenzyme Q, dolichol, isopentenyladenosine, etc. which have

pivotal roles in cell biology and human physiology and potential

relevance to benefits (pleiotropic effects) as well as risks (adverse

effects) of statins.3–6

The adverse effects of statins on soft tissues (such as muscle,

peripheral nerves, liver, testis, etc.) may result from impaired

protein prenylation, deficiency of coenzyme Q involved in

mitochondrial electron transport and antioxidant protection,

abnormal protein glycosylation due to dolichol shortage, or

deficiency of selenoproteins.2 Myopathy is the most frequent side

effect of statins and in some cases may have a form of severe

rhabdomyolysis.7 Less common adverse effects include hepato-

toxicity, peripheral neuropathy, problems with sex steroids

impaired myocardial contractility and autoimmune diseases.8 In

recent years, among the pleiotropic effects, the effect of statins on

bone tissue and osteoporosis is the most challenging issue.9 Like

anti-resorptive agents such as bisphosphonates, simvastatin

inhibits osteoclast activation by inhibiting the prenylation of

certain GTP binding proteins (Rho, Rac, Rab etc.).10

Analyst, 2010, 135, 3233–3241 | 3233

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

Simvastatin is one of the most frequently prescribed statins

because of its efficacy in reducing LDL lipoprotein cholesterol

levels, its tolerability, and its reduction of cardiovascular risk and

mortality.11 Conflicting results have been reported with regard to

adverse effects of simvastatin on liver, testis and sciatic nerve

tissues and pleiotropic effects on bone tissues.

Prescribing information for simvastatin (and other statins) has

typically recommended routine liver function monitoring.12

Although statin therapy can increase hepatic transaminases,13, 14

there is still controversial results about the hepatotoxicity of the

drug. According to some studies in the literature,15–17 there is

little evidence to support any increased risk for hepatotoxicity

while some studies supported the increased lipid peroxidation in

liver of rats18 and humans19–21 with administration of statins.

Since, in the recent literature more serious hepatotoxicity cases

related to statins have been reported,19–21 the vital importance of

early detection of these serious adverse drug reactions should be

emphasised.

Simvastatin could have potential side effects on the adrenal

gland, ovary, and testis, as these three glands use cholesterol for

their hormonal biosynthesis.22 However, in the literature there

are conflicting results about the testicular adverse effects of sta-

tins. Some data indicate that statins reduce serum testosterone

concentrations and induce erectile dysfunction (ED),23 but other

data indicate that statins have no effect on sex hormones or

spermatogenesis.22 Urologic adverse effects of statins rarely

occur but should not be discarded.24

Simvastatin is also presumed to cause peripheral neuropathy.25

However, the adverse effects of the drug on peripheral nerves

including sciatic nerve and neural regeneration is still unclear.

There are published case reports of peripheral neuropathy in

patients taking statins.26,27 Despite that, some recent studies28,29

supported the neuroprotective properties of simvastatin through

modifying intracellular or extracellular environments, making it

favorable for regeneration.30

The adverse effects of statins appear to be small relative to the

significant cardioprotective benefits.31 However, due to

increasing number of patients taking statins, monitoring for any

side effects and intense research to recognize their mechanisms

are mandatory to further improve the safety of these drugs.

The pleiotropic effect of simvastatin on bone quality is still

controversial in the literature. Some studies32–34 have reported

positive effects of simvastatin on bone tissue by increasing bone

formation through induction of BMP-2 and by the accumulation

of bone matrix proteins like type 1 collagen. On the other hand,

there are also other studies that support the ineffectiveness of

simvastatin therapy on bone tissues.35–37

In our previous studies, the compositional and structural

changes of macromolecules induced by simvastatin therapy on

healthy rat extensor digitorum longus (EDL) muscle38 and tibia

bone tissue9 were investigated using FTIR spectroscopy with

KBR pellet method. The results of muscle study revealed that 20

mg kg�1 day�1 simvastatin treatment induced a significant

decrease in lipid, nucleic acid, protein and glycogen content.

Detailed secondary structure analysis of the amide I band by

intensity calculations from second derivative spectra revealed

a significant decrease in b sheet structure and a significant

increase in antiparallel and aggregated b sheet and random coil

structure which indicate protein denaturation. This was the first

3234 | Analyst, 2010, 135, 3233–3241

study which reported the simvastatin induced-protein secondary

structure changes on muscle tissues. The effects of the drug on

protein secondary structure of other soft tissues such as liver,

testis and sciatic nerve, are not known yet.

In our previous bone study,9 it was determined that 50 mg kg�1

day�1 simvastatin treatment strengthen tibia tissues of rats even

in the absence of any disease state. However, it was reported that

simvastatin therapy at this dose may induce lipid peroxidation

and cause changes in the physical properties of the lipid envi-

ronment, such as disordering and increasing the fluidity of the

membrane. Moreover, detailed secondary structure analysis by

intensity calculations from second derivative spectra implied

a significant increase in a-helical structure due to the increase of

collagen I production and decrease in random coil structure

implying more strengthened bone tissue. This study was also the

first in the literature which reported the protein secondary

structure changes induced by simvastatin therapy on tibia

tissues.9

In the current study, Attenuated Total Reflectance-Fourier

transform infrared (ATR-FTIR) spectroscopy was used to

obtain detailed information of the protein secondary structure of

soft tissues namely liver, testis, sciatic nerve and hard tissues

namely femur and tibia after simvastatin therapy. Recently,

a number of secondary structure prediction methods based on

reference sets of FTIR spectra from proteins with known struc-

ture from X-ray crystallography have been suggested.39–41 In the

present study, protein secondary structural changes induced by

simvastatin were firstly predicted by intensity calculations from

second derivative spectra and neural network (NN) analysis,

using the amide I band (1700�1600 cm�1) of FTIR spectra.

Moreover, based on protein secondary structural differences,

hierarchial cluster analysis was carried out in the

1700�1600 cm�1 region.

As an emergent area of research, a significant number of FTIR

studies have been undertaken on soft38,42 and hard9,43 tissues to

detect the early alterations induced by the administration of

drugs or the development of pathologies, which are not easily

detectable by morphological methods.44 Attenuated Total

Reflection Fourier Transform Infrared spectroscopy (ATR-

FTIR) is a well-established analytical tool applicable to full-

spectrum characterisation of chemical and biological species

without particular sample preparation.45 Another advantage of

ATR-IR over transmission-IR, is the limited path length into the

sample. This avoids the problem of strong attenuation of the IR

signal in highly absorbing media.46

Experimental

Materials

Simvastatin (Zocor) was purchased from the company name

Merck, Sharp and Dohme (West Point, PA, USA) as tablets

containing 40 mg in each.

Animal studies

Adult male Wistar rats (12–14 weeks) weighing 250–300 g (HıfzıSıhha Animal Center, Ankara) were fed with a standard diet with

water ad libitum, and kept in a conventional room with

controlled light (12 : 12, dark : light), temperature (22 � 1 �C),

This journal is ª The Royal Society of Chemistry 2010

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

relative humidity (40–50%) and ventilation (15 air changes per

hour). They were allowed to adapt to their environment for one

week prior to the experiments.

The rats were separated into two groups as control (n ¼ 10)

and simvastatin-treated (n ¼ 10). The control group received

serum physiologic solution, while the simvastatin treatment

group received simvastatin (50 mg kg�1 day�1) in serum physio-

logic daily by gavage for 30 days.9,38 In humans 80 mg day�1 of

simvastatin is the highest recommended dose for treatment of

hypercholesterolaemia.47,48 For an individual who weighs 70 kg,

this dose corresponds to 1.1 mg kg�1.49 Since rats metabolize

statins more rapidly than humans, a human dose of 1.1 mg kg�1

day�1 is comparable to 50 mg kg�1 day�1 used in our rat model.

The dose used in this study was based on earlier animal studies

on simvastatin treated rat models.9,47,50–54 At the end of the

treatment period rats were decapitated and liver, testis, sciatic

nerve, femur and tibia samples were taken and stored at �80 �C

for ATR-FTIR studies. Approval for the study was obtained

from the Animal Ethical Committee of Hacettepe University.

Sample preparation for ATR-FTIR studies

The soft tissues were placed on the Diamond/ZnSe (Di/ZnSe)

crystal plate of the Universal ATR unit of the FTIR spectrom-

eter. The cortices of bone samples first were ground in liquid

nitrogen in a liquid nitrogen-cooled colloid mill (Retsch MM200)

to obtain tissue powder and then bone powder put on the crystal

like soft tissues.

Data acquisition and spectroscopic analysis

FTIR spectra of tissues were recorded with an ATR unit (Perkin

Elmer) combined with a Perkin-Elmer Spectrum 100 FTIR

spectrometer (Perkin-Elmer Inc., Norwalk, CT, USA). The

spectra were recorded in the 4000–400 cm�1 region at room

temperature. A total of 100 scans were taken for each inter-

ferrogram at 4 cm�1 resolution.

Collections of spectra and data manipulations were carried out

using Spectrum 100 software (Perkin-Elmer). Each sample was

scanned as three different replicates under the same conditions,

all of which gave identical spectra. The average spectra of these

three replicates were then used in detailed data and statistical

analysis.

For the determination of simvastatin-induced protein

secondary structure variations, intensity calculations from

second derivative spectra using the amide I band (1700 � 1600

cm�1) was performed using OPUSNT data collection software

package (Bruker Optics, Reinstetten, Germany). Firstly the

second derivatives spectra were obtained by applying a Savitzky–

Golay algorithm with nine smoothing points and these deriva-

tives vector normalized at 1700 � 1600 cm�1 and then the peak

intensities were calculated. The peak minima of the second

derivative signals were used, since they correspond to the peak

positions of the original absorption spectra.38,55

Neural Network (NN) analysis

Protein secondary structure predictions were also carried out

using neural networks. Neural networks were first trained using

FTIR spectra of 18 water soluble proteins recorded in water

This journal is ª The Royal Society of Chemistry 2010

whose secondary structures were known from X-ray Crystal-

lography.56 Before applying to the neural networks, the Amide I

band (1700 � 1600 cm�1) was preprocessed which involves

normalization and discrete cosine transformation (DCT) of the

amide I band of the FTIR spectra. To improve the training of the

neural networks, the size of the training data set was increased by

interpolating the available FTIR spectra. For each structure

parameter, a separate NN was trained using Bayesian regulari-

zation whose number of inputs, i.e., the number of DCT coeffi-

cients, and the number of hidden neurons were optimized. The

trained NNs have standard error of prediction values of 4.19%

for alpha helix, 3.49% for beta sheet and 3.15% for turns. The

secondary structure parameters of the new proteins were pre-

dicted by applying the inputs of the trained NNs the pre-

processed FTIR data as reported in detailed in Severcan et al.

(2004).39

Cluster analysis

For comparison of control and treated soft and hard tissues with

cluster analysis, second derivative spectra using a nine smoothing

point Savitzky-Golay algorithm were independently vector-

normalised in 1700 � 1600 cm�1 spectral region for the analysis

of signals from protein secondary structures. Cluster analysis is

a procedure that groups the spectra based on similarities of their

spectral characteristics. When graphically displayed, the result of

the analysis forms a dendrogram. The change in variances

between the spectra of samples is represented by heterogeneity

values. Higher heterogeneity in between the clusters demon-

strates higher differences among analyzed groups. Heterogeneity

values were calculated automatically by the OPUS 5.5 software

and for the calculation Euclidean distances and Ward’s algo-

rithm method were used. The details of the calculation and

algorithm can be found in Severcan et al. (2010).57

Statistics

The results were expressed as mean � standard deviation (SD).

Data were analyzed statistically by using Mann–Whitney U test

and the p values less than or equal to 0.05 were considered as

statistically significant (*p # 0.05; **p # 0.01; ***p # 0.001).

Results and discussion

In the present study, protein secondary structural changes of soft

tissues namely liver, testis, sciatic nerve and hard tissues namely

femur and tibia, induced by 50 mg kg�1 day�1 simvastatin

therapy were predicted by intensity calculations from second

derivative spectra and neural network (NN) analysis. Moreover,

based on protein secondary structural differences, hierarchial

cluster analysis was carried out in the 1700–1600 cm�1 region.

Clinically relevant simvastatin doses vary depending on the

studied animal model. Since rat and mouse models metabolize

statins more rapidly than humans, higher doses are required for

these animals to access similar effective doses in humans.58,59 In the

previous studies, it was reported that high therapeutic simvastatin

doses in humans correspond to 50–100 mg kg�1 day�1 dose in

rats50–54 and 50–202 mg kg�1 day�1 dose in mouse.58,60, 61Therefore

in the current study, the simvastatin dose used as 50 mg kg�1 day�1

Analyst, 2010, 135, 3233–3241 | 3235

Table 1 The assignments of secondary structure sub-bands underAmide I band in 1700–1600 cm�1 region

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

for rats is in the lower limit of high dose and corresponds to

a clinically relevant dose of simvastatin.

Peak Number Mean Frequencies/cm�1 Assignment

1 1694 Antiparallel beta sheet1676 Turns

2 1652 Alpha helix3 1642 Random coil4 1638 Beta sheet5 1625 Aggregated beta sheet

Adverse effects on soft tissues

Fig. 1 shows a typical absorbance (A) and second derivative (B)

FTIR spectrum of rat liver, testis and sciatic nerve tissues in the

1700–1600 cm�1 region. The assignments of secondary structure

sub-bands under Amide I band in this region are given in Table 1.

The changes in the protein structure were determined from the

intensities of the sub-bands in the second derivative of the amide

I band (Fig. 1B). The intensity values of these bands in liver,

testis and sciatic nerve tissues of control and simvastatin-treated

groups are listed in Table 2. The peak located at 1690 cm�1 is due

to antiparallel b sheet structure, the peak at 1682 cm�1 arises

from turns and bends, the peak located at 1638 cm�1 is due to

b sheet structures, the peak at 1659 cm�1 corresponds to a a-helix

structure, the peak at 1648 cm�1 is assigned to random coil

structure and the peak around 1629 cm�1 is attributed to

aggregated b sheets structure.9,38,62

The results revealed that simvastatin treatment significantly

decreased alpha helix structure in liver (p<0.01) and sciatic nerve

(p<0.001), while a slight decrease was observed in testis tissues.

Beta sheet structure at 1638 cm�1 was also significantly decreased

in all tissues of treated groups. However, anti-parallel and

aggregated beta sheet structures were significantly increased in

simvastatin-treated liver (p<0.05, p<0.001, respectively), sciatic

nerve (p<0.001, p<0.05, respectively) and testis (p<0.01) tissues

compared to the control tissues. In addition, the random coil

structure was increased in liver (p<0.01) and sciatic nerve

Fig. 1 The average (A) absorbance and (B) second derivative spectra of

control liver, testis and sciatic nerve tissue in 1700�1600 cm�1 region.

3236 | Analyst, 2010, 135, 3233–3241

(p<0.001) tissues. The increase in random coil structure in testis

tissue of treated group was not statistically significant. The

decrease in beta sheet and the increase in random coil and

aggregated b sheet structures indicate simvastatin-induced

protein denaturation.38,63

These results from secondary structure were further supported

by neural networks results. The results are presented in Table 3.

It is clearly seen from the table that simvastatin treatment caused

significant changes in the protein secondary structures of liver

and sciatic nerve tissues by decreasing the content of alpha helix

and by increasing the content of beta sheet and random coil

structures, while there was no obvious changes in the secondary

structure of testis tissue. According to the intensities of the sub-

bands in the second derivative of the amide I band, the beta sheet

at 1638 cm�1 was decreased while, in neural network results there

is an increase in beta sheet structures. The neural network is

applied to the whole 1600–1700 cm�1 region without considering

the location of substructures, therefore the results of the analysis

give us the total amount of each secondary structural parameters,

i.e., the content of the beta sheet includes the sum of native beta

sheet at 1638 cm�1, antiparallel beta sheet at 1690 cm�1 and

aggregated beta sheet at 1629 cm�1. Since the results of neural

network analysis give the total content of beta sheet structures,

this increase in beta sheet content was due to the increase in

antiparallel and aggregated beta sheets, which was also revealed

by the second derivative spectra. On the other hand the same type

of behaviour such as decrease in alpha helix and increase in beta

sheet is almost a general trend at the early stage of several

diseases such as diabetes64 and neurodegenerative diseases.65 Our

results clearly show that protein secondary structure is signifi-

cantly effected from simvastatin treatment.

HMG-CoA reductase inhibitors depresses the synthesis not

only of cholesterol, but also of the isoprenoid lateral chain of

ubiquinone Q10 (coenzyme Q: a natural protector against free

radical oxidation).66 Thus, simvastatin induced inhibition of

ubiquinone Q10 biosynthesis in tissues has adverse consequences

such as lipid peroxidation.9,66 On the other hand, Lankin et al.66

reported in their study a sharp decrease in the enzymatic systems

activity which utilizes reactive oxygen species and lipid peroxides

(such as superoxide dismutase and glutathione peroxidase) in the

hepatocytes, blood cells and other soft tissues. It was assumed

that the decreasing of antioxidative enzymes activity may be

a direct cause of intensification of free radical lipoperoxidation in

the blood and other tissues during statin treatment.66,67

There are several studies9,38,68 that reported the effects of

oxidation on the structure and stability of proteins. Jayaraman

et al. (2007)69 investigated the effect of the oxidation on

secondary structure of apolipoprotein B-100 moiety of low

This journal is ª The Royal Society of Chemistry 2010

Table 2 The results of the changes in the intensities of main proteinsecondary structures for control and simvastatin treated soft tissues;liver, testis and sciatic nerve. [The values are the mean� standard error ofmean for each group. *p # 0.05; **p # 0.01; ***p # 0.001]

Functional Groups Control (n ¼ 10) Treated (n ¼ 10)

Livera-helical structure

(at 1658 cm�1)�0.116 � 0.005 �0.108 � 0.004**

b-sheet structure(at 1638 cm�1)

�0.159 � 0.008 �0.154 � 0.009

Antiparallel b-sheetstructure(at 1690 cm�1)

�0.018 � 0.004 �0.27 � 0.008*

Aggregated b-sheetstructure(at 1628 cm�1)

�0.178 � 0.005 �0.190 � 0.007***

Random coilstructure(at 1645 cm�1)

�0.073 � 0.004 �0.080 � 0.007*

Testisa-helical structure �0.106 � 0.018 �0.093 � 0.009b-sheet structure �0.190 � 0.013 �0.134 � 0.016***Antiparallel b-sheet

structure�0.016 � 0.011 �0.033 � 0.007**

Aggregated b-sheetstructure

�0.175 � 0.016 �0.186 � 0.009

Random coil structure �0.098 � 0.011 �0.102 � 0.007Sciatic Nervea-helical structure �0.150 � 0.026 �0.114 � 0.011***b-sheet structure �0.214 � 0.019 �0.205 � 0.011*Antiparallel b-sheet

structure�0.035 � 0.003 �0.044 � 0.006***

Aggregated b-sheetstructure

�0.239 � 0.010 �0.260 � 0.010***

Random coilstructure

�0.068 � 0.008 �0.085 � 0.012***

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

density lipoprotein (LDL) particles. This protein assumes a pen-

tapartite structure with alternating alpha helices and beta-plea-

ted sheets (a1–b1–a2–b2–a3).70 Beta-sheets are structurally rigid

and engaged in electrostatic interactions with the phospho-

lipids.70 The oxidative modifications render the LDL particle

electronegatively charged (LDL-) as compared to native LDL

(nLDL).69 Oxidatively-modified LDL also contains elevated

Table 3 The results of the neural network predictions based on FTIRdata in 1600–1700 cm�1 spectral region (Amide I band) for control andsimvastatin treated soft tissues; liver, testis and sciatic nerve. [The valuesare the mean � standard error of mean for each group. *p # 0.05; **p #0.01; ***p # 0.001]

Functional groups Control (n ¼ 10) Treated (n ¼ 10)

Livera-helical structure 12.45 � 0.17 11.80 � 0.11**b-sheet structure 44.88 � 0.16 45.71 � 0.63Turns 24.30 � 0.6 22.96 � 0.8*Random coil structure 18.37 � 0.45 19.52 � 0.33**Testisa-helical structure 10.89 � 0.64 10.21 � 0.38b-sheet structure 44.36 � 0.33 45.64 � 0.97*Turns 23.65 � 0.5 22.20 � 0.3*Random coil structure 21.09 � 1.00 21.95 � 0.36Sciatic Nervea-helical structure 13.81 � 0.65 12.00 � 0.75**b-sheet structure 46.63 � 0.95 49.30 � 0.89*Turns 22.83 � 0.9 19.38 � 1.00*Random coil structure 16.74 � 0.99 19.33 � 1.20***

This journal is ª The Royal Society of Chemistry 2010

levels of lipid peroxides and aldehydes that are implicated in

protein unfolding. In the study of Jayaraman et al. (2007),69

oxidatively-modified LDL-induced specific protein modifica-

tions and conformational changes were studied by liquid chro-

matography/tandem mass spectrometry (LC/MS/MS) analyses

and circular dichroism (CD). According to the results, oxidation

assisted the loss of alpha-helical structure and an increase in anti-

parallel beta sheets, and random coil structures. These reported

results are in agreement with our spectral and neural network

results. Since low density lipoprotein (LDL) particles transport

cholesterol, cholesterol esters, lipids and phospholipids to

peripheral tissues,69 statin-induced oxidation of this particle may

affect the composition and secondary structure of proteins in

these tissues.

Hierarchical cluster analysis was also performed for compar-

ison of normal and simvastatin-treated groups for liver, testis

and sciatic nerve tissues. The results of cluster analysis using

1700–1600 cm�1 spectral region of control and treated groups for

investigated soft tissues were shown in Fig. 2. Hierarchy of

clusters from individual elements is represented as dendrograms,

which are tree-like diagrams showing the arrangements of the

clusters. As seen from the figure, two distinct clusters were

produced corresponding to control and simvastatin treated

groups for all tissues in the spectral regions subjected to cluster

analysis. There was only one misclassification (simvastatin

treated group as control) in liver tissues which is shown with ‘‘†’’

in Fig. 2A; and two misclassifications (control as simvastatin

treated group) in testis and sciatic nerve tissues which are shown

with ‘‘†’’ in Fig. 2B and C, respectively. The heterogeneity value

which represents the spectral distance, was found to be about

0.62 for liver and 0.72 for sciatic nerve tissues. It can be noted

that the heterogeneity value for testis tissue was very low (0.20)

implying a low variance between control and simvastatin-treated

groups. Moreover, according to the heterogeneity values in the

cluster analysis of these tissues, the sciatic nerve tissue proteins

were found to be the most affected tissue proteins from simvas-

tatin treatment among the studied soft tissues. It was expected

that the most affected tissue would be the liver, since simvastatin

is metabolized by the important 3A4 isoenzyme of the cyto-

chrome P450 system of the liver.71 It was reported that simvas-

tatin can induce asymptomatic mild elevation of serum

transaminases with an incidence quoted between 1% and 1.5%.19

The elevation of serum transaminases is often self-limiting and

thought to relate to alteration of the hepatocyte cellular

membrane with enzyme leakage rather than direct severe liver

injury.19 In our study, this can be the reason of the slight effect of

the drug on liver when compared to the other investigated soft

tissues. In addition to that, nervous toxicity of simvastatin may

be explained by the inhibition of cholesterol synthesis that may

alter myelin and nerve membrane function and prevent mito-

chondrial respiratory chain enzyme synthesis which may disturb

neuron energy use.72

Pleitropic effects on hard tissues

Fig. 3 shows a typical absorbance (A) and second derivative (B)

FTIR spectrum of a rat femur and tibia tissues in the 1700–1600

cm�1 region. The assignments of secondary structure sub-bands

Analyst, 2010, 135, 3233–3241 | 3237

Fig. 2 The dendrogram of a hierarchical cluster analysis of control (K)

and 50 mg simvastatin treated (TR) groups in (A) liver, (B) testis and (C)

sciatic nerve tissues. Cluster analysis was performed on the second

derivative and vector normalized spectra of tissue samples and resulting

from Ward’s algorithm. The study was conducted in the 1700–1600 cm�1

spectral region.

Fig. 3 The average (A) absorbance and (B) second derivative spectra of

control femur and tibia bone tissue in 1700–1600 cm�1 region.

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

under Amide I band in this region were mentioned before and

given in Table 1.

The intensity values of the sub-bands in the second derivative

of the amide I band of control and treated groups were given in

Table 4. As seen from the table, antiparallel b sheet structure at

1690 cm�1, b sheets structures at 1638 cm�1 and aggregated

b sheets structure at 1629 cm�1 were significantly decreased in

both femur (p<0.05, p<0.01, p<0.01, respectively) and tibia

3238 | Analyst, 2010, 135, 3233–3241

(p<0.05, p<0.001, p<0.001, respectively) tissues of simvastatin-

treated groups. Random coil structure was also significantly

decreased (p<0.01) in femur tissue, while a slight decrease was

observed for this structure in tibia tissues of treated group

compared to the control group. Moreover, in both bone tissues

of simvastatin-treated group, alpha helix was significantly

increased (p<0.05).

The protein region, corresponding to absorption values

between 1600 and 1700 cm�1 was also analysed using neural

networks based on FTIR data to estimate the simvastatin

induced alterations on protein secondary structure of bone

tissues. The results are presented in Table 5. As seen from the

table, simvastatin treatment caused a significant increase in alpha

helix structure in both femur (p<0.01) and tibia (p<0.001)

tissues. Moreover, supporting the results mentioned above

significant decrease in beta sheet and random coil structures were

observed in both bone tissues of treated group.

The strength of bone tissues depend both on the mineral and

the matrix (primarily type I collagen fibrils) constituents.73

Simvastatin is thought to up-regulate the expression of a series of

growth factors in osteoblasts, ultimately enhancing matrix and

collagen production.9 Collagen is essential for maintaining the

tensile strength in bone.74 The distinct features of bone (type I)

collagen are its cross-linking chemistry and its molecular packing

structure.75 Type I collagen can be differentiated from other

types of collagen by the higher amounts of a-helix and triple

This journal is ª The Royal Society of Chemistry 2010

Table 4 The results of the changes in the intensities of main proteinsecondary structures for control and simvastatin treated hard tissues;femur and tibia. [The values are the mean � standard error of the meanfor each group. *p # 0.05; **p # 0.01; ***p # 0.001]

Functional Groups Control (n ¼ 10) Treated (n ¼ 10)

Femura-helical

structure(at 1658 cm�1)

�0.134 � 0.012 �0.149 � 0.013*

b-sheetstructure(at 1638 cm�1)

�0.123 � 0.008 �0.115 � 0.009

Antiparallel b-sheetstructure(at 1690 cm�1)

�0.038 � 0.006 �0.028 � 0.006*

Aggregatedb-sheet structure(at 1628 cm�1)

�0.190 � 0.009 �0.168 � 0.009**

Random coilstructure(at 1645 cm�1)

�0.293 � 0.017 �0.256 � 0.022**

Tibiaa-helical

structure�0.131 � 0.004 �0.158 � 0.005*

b-sheet structure �0.183 � 0.012 �0.109 � 0.011***Antiparallel b-sheet

structure�0.038 � 0.004 �0.032 � 0.005*

Aggregated b-sheetstructure

�0.186 � 0.005 �0.093 � 0.009***

Random coilstructure

�0.288 � 0.009 �0.265 � 0.011

Table 5 The results of the neural network predictions based on FTIRdata in 1600–1700 cm�1 spectral region (Amide I band) for control andsimvastatin treated hard tissues; femur and tibia. [The values are themean � standard error of mean for each group. *p # 0.05; **p # 0.01;***p # 0.001]

Functional Groups Control (n ¼ 10) Treated (n ¼ 10)

Femura-helical structure 16.40 � 0.63 17.49 � 1.03b-sheet structure 51.87 � 0.98 49.16 � 0.68*Turns 19.49 � 0.8 22.52 � 0.4*Random coil structure 12.34 � 0.74 10.83 � 1.10**Tibiaa-helical structure 15.77 � 0.65 18.74 � 1.00***b-sheet structure 52.97 � 1.21 50.98 � 0.44*Turns 18.71 � 1.05 21.24 � 1.0**Random coil structure 12.55 � 1.05 9.05 � 1.60***

Fig. 4 The dendrogram of a hierarchical cluster analysis of control (K)

and 50 mg simvastatin treated (T) groups in (A) femur and (B) tibia bone

tissues. Cluster analysis was performed on the second derivative and

vector normalized spectra of tissue samples and resulting from Ward’s

algorithm. The study was conducted in the 1700–1600 cm�1 spectral

region.

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

helix.9 The increase in the content of a-helical structure could be

due to the increase of collagen I production which was induced

by simvastatin therapy.9 This increase in collagen type I with the

decrease of aggregated beta sheet and random coil structure

imply a more strengthened bone tissue in simvastatin-treated

groups.9

Hierarchical cluster analysis was also performed for compar-

ison of normal and simvastatin-treated groups for femur and

tibia tissues in 1700–1600 cm�1 spectral region. The results of

cluster analysis were shown in Fig. 4. As seen from the figure,

there was only one misclassification (control as simvastatin

treated group) in femur tissue which is shown with ‘‘‡’’ in

This journal is ª The Royal Society of Chemistry 2010

Fig. 4A. There was no misclassification between control and

treated groups in tibia tissues (Fig. 4B). The heterogeneity value

was about 0.22 for femur and 1.42 for tibia tissues. These values

imply high secondary structural difference between control and

simvastatin-treated groups in tibia bone tissues.

As discussed above the simvastatin-treated and control groups

for all soft and bone tissues were successfully differentiated

between using cluster analysis on the basis of spectral patterns of

main protein amide I band (Fig. 2 and 4). These findings reveal

that simvastatin treatment causes some important changes in the

FTIR spectra in protein secondary structure parameters, which

can be successfully determined by using cluster analysis.

Conclusion

There are some controversial results reported in the literature

about the adverse effects of simvastatin on soft tissues; liver,

testis and sciatic nerve and pleitropic effects on hard tissues;

femur and tibia. In this study, it was determined that simvastatin

treatment caused significant alterations on protein conforma-

tions of soft and hard tissues. The sciatic nerve tissue was found

to be the most negatively affected tissue from simvastatin

Analyst, 2010, 135, 3233–3241 | 3239

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

treatment among the studied soft tissues. However, different

from soft tissues, simvastatin-treatment strengthened the bone

tissues.

In the current study, FTIR spectroscopy together with bio-

informatics methods such as neural network and hierarchical

clustering, allowed us to determine the simvastatin induced

protein conformational changes as adverse and pleitropic effects

of the drug on different soft and hard tissues.

References

1 T. Koga, K. Fukuda, Y. Shimada, M. Fukami, H. Koike andY. Tsujita, Eur. J. Biochem., 1992, 209, 315–319.

2 J. Be1towski, G. W�ojcicka and A. Jamroz-Wi�sniewska, Curr. DrugSaf., 2009, 4(3), 209–28.

3 B. A. Golomb and M. A. Evans, Am. J. Cardiovasc. Drugs, 2008, 8(6),373–418.

4 J. L. Goldstein and M. S. Brown, Nature, 1990, 343, 425–30.5 F. Mach, Circulation, 2004, 109, II-15–7.6 J. Keyhani and E. Keyhani, FEBS Lett., 1978, 93, 271–4.7 M. Evans and A. Rees, Curr. Opin. Lipidol., 2002, 13, 415–420.8 W. V. Brown, Curr. Opin. Lipidol., 2008, 19(6), 558–62.9 S. Garip and F. Severcan, J. Pharm. Biomed. Anal., 2010, 52(4), 580–

8.10 S. B. Jadhav and G. K. Jain, JPP., 2006, 58, 3–18.11 J. Szendroedi, C. Anderwald, M. Krssak, M. Bayerle-Eder,

H. Esterbauer, G. Pfeıler, A. Brehm, P. Nowotny, A. Hofer,W. W. Ausl and M. Roden, Diabetes Care, 2008, 32, 209–214.

12 MRC/BHF Heart Protection Study Collaborative Group, BMCPharmacol., 2009, 9, 6, DOI: 10.1186/1472-6904-9-6.

13 K. G. Tolman, Am. J. Cardiol., 2002, 89, 1374–80.14 H. Bays, Am. J. Cardiol., 2006, 97, 6C–26C.15 H. Bays, Am. J. Cardiol., 2008, 101(8A), 3B–8B.16 M. Law and A. R. Rudnicka, Am. J. Cardiol., 2006, 97(suppl), 52C–

60C.17 P. H. Chong, A. Boskovich, N. Stevkovic and R. E. Bartt,

Pharmacotherapy, 2004, 24(9), 1194–203.18 V. Z. Lankin, M. V. Ivanova, G. G. Konovalova, A. K. Tikhaze,

A. I. Kaminnyi and V. V. Kukharchuk, Bull. Exp. Biol. Med., 2007,143(4), 414–7.

19 A. T. Clarke and P. R. Mills, Dig. Liver Dis., 2006, 38, 772–777.20 L. Perger, M. Kohler, K. Fattinger, R. Flury, P. Meier and C. J. Pauli-

Magnus, J. Hepatol., 2003, 39, 1096–7.21 M. Geoghegan, V. Smith and J. R. B. Green, Gut, 2004, 53(Suppl.

III), A123.22 C. Azzarito, L. Boiardi, M. Zini, A. Agosti, C. Dotti, R. Biagi and

I. Portioli, Metab., Clin. Exp., 1992, 41(2), 148–53.23 G. Corona, V. Boddi, G. Balercia, G. Rastrelli, G. De Vita, A. Sforza,

G. Forti, E. Mannucci and M. Maggi, J. Sex. Med., 2010, 7(4 Pt 1),1547–56.

24 J. M. Backes and P. A. Howard, Ann Pharmacother., 2007, 41(1), 138–42.

25 E. Daglioglu, M. Berker, M. Demirci, M. Tuncel, E. Karabulut andA. Tuncel, Folia Neuropathol., 2010, 48(1), 49–56.

26 P. H. Chong, A. Boskovich, N. Stevkovic and R. E. Bartt,Pharmacotherapy, 2004, 24, 1194–1203.

27 J. L. K. Anderson, J. B. Muhlestein, T. L. Bair, S. Morris,A. N. Weaver, D. L. Lapp�e, D. G. Renlund, R. R. Pearson,K. R. Jensen and B. D. Horne, Am. J. Cardiol., 2005, 95, 1097–1099.

28 L. H. Weimer and N. Sachdev, Curr. Neurol. Neurosci. Rep., 2009,9(1), 69–75.

29 M. R. Gholami, F. Abolhassani, P. Pasbakhsh, M. Akbari,A. Sobhani, M. R. Eshraghian, N. Kamalian, F. A. Amoli,A. R. Dehpour and D. Sohrabi, Eur. J. Pharmacol., 2008, 590(1–3),111–4.

30 H. C. Pan, D. Y. Yang, Y. C. Ou, S. P. Ho, F. C. Cheng andC. J. Chen, Neurosurgery, 2010, Jun 9, In Press.

31 S. A. Linnebur and W. H. Hiatt, Ann Pharmacother., 2003, 37(2),274–8.

32 R. Garrett and G. R. Mundy, Arthritis Res., 2002, 4(4), 237–240.

3240 | Analyst, 2010, 135, 3233–3241

33 D. C. Bauer, G. R. Mundy, S. A. Jamal, D. M. Black, J. A. Cauley,K. E. Ensrud, M. van der Klift and H. A. Pols, Arch. Intern. Med.,2004, 164, 146–152.

34 Y. Lee, M. J. Schmid, D. B. Marx, M. W. Beatty, D. M. Cullen,M. E. Collins and R. A. Reinhardt, Biomaterials, 2008, 29, 1940–1949.

35 A. L. Anbinder, J. C. Junqueira, M. N. G. Mancini, I. Balducci,R. F. da Rocha and Y. R. Carvalho, Brazilian Dental Journal,2006, 17(4), 267–273.

36 T. Van staa, S. Wegman, F. de Vries, B. Leufkens and C. Cooper,JAMA, J. Am. Med. Assoc., 2001, 285(14), 1850–1856.

37 J. Hsia, M. Morse and V. Levin, BMC Musculoskeletal Disord., 2002,3, 7.

38 N. S. Ozek, Y. Sara, R. Onur and F. Severcan, Biosci. Rep., 2010,30(1), 41–50.

39 M. Severcan, P. I. Haris and F. Severcan, Anal. Biochem., 2004,332(2), 238–44.

40 J. A. Hering, P. R. Innocent and P. I. Haris, Proteomics, 2004, 4(8),2310–9.

41 E. Goormaghtigh, J. M. Ruysschaert and V. Raussens, Biophys. J.,2006, 90(8), 2946–2957.

42 E. Bogomolny, M. Haleihel, A. Salman, A. Zwielly, R. Moreh andS. Mordechai, Analyst, 2010, 135, 1934–1940.

43 H. Boyar, F. Zorlu, M. Mut and F. Severcan, Anal. Bioanal. Chem.,2004, 379, 433–438.

44 P. Carmona, A. Rodr�ıguez-Casado, I. Alvarez, E. de Miguel andA. Toledano, Biopolymers, 2008, 89(6), 548–554.

45 J. Greener, B. Abbasi and E. Kumacheva, Lab Chip, 2010, 10, 1561–1566.

46 ‘‘T-IR Spectroscopy—AttenuatedTotal Reflectance (ATR)’’. PerkinElmer Life and Analytical Sciences, 2005, Retrieved 2007-01-26.

47 K. M. Thelen, R. Laaksonen, H. P€aiv€a, T. Lehtim€aki andD. L€utjohann, J. Clin. Pharmacol., 2006a, 46(7), 812–6.

48 T. N. Patel, M. H. Shishehbor and D. L. Bhatt, Eur. Heart J., 2007,28(6), 664–72.

49 F. Fassbender, M. Simons, C. Bergmann, M. Stroick, D. L€utjohann,P. Keller, H. Runz, S. K€uhl, T. Bertsch, K. Von Bergmann,M. Hennerici, K. Beyreuther and T. Hartmann, Neurobiology, 2001,98(10), 5856–5861.

50 N. Cartier, S. Guidoux, F. Rocchiccioli and P. Aubourg, FEBS Lett.,2000, 478, 205–208.

51 L. N. Johnson-Anuna, G. P. Eckert, J. H. Keller, U. Igbavboa,C. Franke, T. Fechner, M. Schubert-Zsilavecz, M. Karas,W. E. M€uller and W. G. Wood, J. Pharmacol. Exp. Ther., 2005,312, 786–793.

52 K. M. Thelen, K. M. Rentsch, U. Gutteck, M. Heverin, M. Olin,V. Andersson, A. von Eckardstein, I. Bj€orkhem and D. L€utjohann,J. Pharmacol. Exp. Ther., 2006, 316(3), 1146–1152.

53 L. Marcoff and P. D. Thompson, J. Am. Coll. Cardiol., 2007, 49(23),2231–2237.

54 M. Jasi�nska, J. owczarek and D. Orszulak-Michalak, Acta PoloniaePharmaceutica, 2006, 63(5), 386–90.

55 N. Toyran, P. Lasch, D. Naumann, B. Turan and F. Severcan,Biochem. J., 2006, 397, 427–436.

56 M. Severcan, F. Severcan and P. I. Haris, J. Mol. Struct., 2001, 565–566, 383–387.

57 F. Severcan, O. Bozkurt, R. Gurbanov and G. Gorgulu,J. Biophotonics, 2010, 8–9, 621–631, DOI: 10.1002/jbio.201000016.

58 B. von Tresckow, E. P. von Strandmann, S. Sasse,S. Tawadros, A. Engert and H. P. Hansen, Haematologica,2007, 92, 682–685.

59 S. Youssef, O. St€uve, J. C. Patarroyo, P. J. Ruiz, J. L. Radosevich,E. M. Hur, M. Bravo, D. J. Mitchell, R. A. Sobel, L. Steinman andS. S. Zamvil, Nature, 2002, 420, 78–84.

60 T. B. Newman and S. B. Hulley, JAMA, J. Am. Med. Assoc., 1996,275(1), 55–60.

61 J. Golledge, B. Cullen, C. Moran and C. Rush, Cardiovasc DrugsTher., 2010, In Press.

62 A. Ausili, A. Torrecillas, M. M. Mart�ınez-Senac, S. Corbal�an-Garc�ıa and J. C. G�omez-Fern�andez, J. Struct. Biol., 2008,164(1), 146–52.

63 S. W. Ellepola, S. M. Choi and C. Y. Ma, Int. J. Biol. Macromol.,2005, 37(1–2), 12–20.

64 J. C. Lin and H. L. Liu, Curr. Drug Discovery Technol., 2006, 3, 145–153.

This journal is ª The Royal Society of Chemistry 2010

Publ

ishe

d on

29

Oct

ober

201

0. D

ownl

oade

d by

Mon

ash

Uni

vers

ity o

n 25

/10/

2014

20:

40:5

9.

View Article Online

65 L. Kupfer, W. Hinrichs and M. H. Groschup, Curr. Mol. Med., 2009,9, 826–835.

66 V. Z. Lankin, A. K. Tikhaze, V. V. Kukharchuk, G. G. Konovalova,O. I. Pisarenko, A. I. Kaminnyi, K. B. Shumaev and Y. N. Belenkov,Mol. Cell. Biochem., 2003, 249, 129–140.

67 V. Z. Lankin, Free Radical Biol. Med., 1994, 16, 8.68 R. T. Hamilton, L. Asatryan, J. T. Nilsen, J. M. Isas, T. K. Gallaher,

T. Sawamura and T. K. Hsiai, Arch. Biochem. Biophys., 2008, 479(1), 1–14.69 S. Jayaraman, D. L. Gantz and O. Gursky, Biochemistry, 2007, 46,

5790–5797.70 P. Holvoet, J. Vanhaecke, J. Janssens, F. Van de Werf and D. Collen,

Circulation, 1998, 98, 1487–1494.

This journal is ª The Royal Society of Chemistry 2010

71 R. L. Talbert, J. Am. Pharm. Assoc., 2003, 46(4), 479–88; quiz488–90.

72 M. Formaglio and C. Vial, Rev Neurol (Paris)., 2006, 162(12),1286–9.

73 E. Durchschlag, E. P. Paschalis, R. Zoehrer, P. Roschger, P. Fratzl,R. Recker, R. Phipps and K. Klaushofer, Journal of Bone andMineral Research, 2006, 21(10), 1581–1590.

74 G. T. Charras, B. A. Williams, S. M. Sims and M. A. Horton,Biophys. J., 2004, 87, 2870–2884.

75 F. Parhami, Y. Tintut, W. G. Beamer, N. Gharavi, W. Goodman andL. L. Demer, Journal of Bone and Mineral Research, 2001, 16, 182–188.

Analyst, 2010, 135, 3233–3241 | 3241