Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/yexcr

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 6

http://dx.doi.org/10.10014-4827/& 2014 El

Abbreviations: NHmatrix; Runx2, Run

nCorresponding au47000, Sungai Buloh

E-mail addressessharaniza_abrahim@smarina@salam.uitm.e

Research Article

Short-term moderate hypothermia stimulatesalkaline phosphatase activity and osteocalcinexpression in osteoblasts by upregulating Runx2 andosterix in vitro

M.D. Aishaa, M.N.K. Nor-Ashikina,c, A.B. Sharanizac, H.M. Nawawib,c,M.Y. Kapitonovaa,c, G.R.A. Froemminga,c,n

aInstitute of Medical Molecular Biotechnology & Faculty of Medicine, Universiti Teknologi MARA, 47000 Sungai Buloh,Selangor, MalaysiabCenter for Pathology Diagnostic and Research Laboratories, Clinical Training Center, Universiti Teknologi MARA,47000 Sungai Buloh, Selangor, MalaysiacDDH, Universiti of Teknologi MARA, Shah Alam, Selangor, Malaysia

a r t i c l e i n f o r m a t i o n

Article Chronology:

Received 11 January 2014Received in revised form18 May 2014Accepted 4 June 2014Available online 11 June 2014

Keywords:

HypothermiaOsteoblastsCell cytoskeletonTranscription factorChaperone proteinFracture healing

016/j.yexcr.2014.06.003sevier Inc. All rights reserv

Ost, Normal human osteot-related transcription fathor at: Institute of Medi, Selangor, Malaysia. Fax: þ: aisha.mohddin@gmail.coalam.uitm.edu.my (A.B. Shdu.my (M.Y. Kapitonova),

a b s t r a c t

Exposure of Normal Human Osteoblast cells (NHOst) to a period of hypothermia may interrupttheir cellular functions, lead to changes in bone matrix and disrupt the balance between boneformation and resorption, resulting in bone loss or delayed fracture healing. To investigate thispossibility, we exposed NHOst cells to moderate (35 1C) and severe (27 1C) hypothermia for 1, 12,24 and 72 h. The effects of hypothermia with respect to cell cytoskeleton organization, metabolicactivity and the expression of cold shock chaperone proteins, osteoblast transcription factors andfunctional markers, were examined. Our findings showed that prolonged moderate hypothermia

retained the polymerization of the cytoskeletal components. NHOst cell metabolism was affecteddifferently according to hypothermia severity. The osteoblast transcription factors Runx2 andosterix were necessary for the transcription and translation of bone matrix proteins, wherealkaline phosphatase (Alp) activity and osteocalcin (OCN) bone protein were over expressedunder hypothermic conditions. Consequently, bone mineralization was stimulated after exposureto moderate hypothermia for 1 week, indicating bone function was not impaired. The cold shockchaperone protein Rbm3 was significantly upregulated (po0.001) during the cellular stressadaption under hypothermic conditions. We suggest that Rbm3 has a dual function: one as achaperone protein that stabilizes mRNA transcripts and a second one in enhancing thetranscription of Alp and Ocn genes. Our studies demonstrated that hypothermia permitted thein vitro maturation of NHOst cells probably through an osterix-dependent pathway. For that

ed.

blasts; Rbm3, RNA binding motif protein 3; Hsp70, heat shock protein 70; ECM, extracellularctor 2; ALP, alkaline phosphatase; OCN, osteocalcincal Molecular Biotechnology & Faculty of Medicine, Universiti Teknologi MARA, Jalan Hospital,6 03 61265082.m (M.D. Aisha), noras011@salam.uitm.edu.my (M.N.K. Nor-Ashikin),araniza), hapizah@salam.uitm.edu.my (H.M. Nawawi),gabriele@salam.uitm.edu.my (G.R.A. Froemming).

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 6 47

reason, we suggest that moderate hypothermia can be clinically applied to counteract heatproduction at the fracture site that delays fracture healing.

& 2014 Elsevier Inc. All rights reserved.

Introduction

Mammalian cells are active and proliferate within a narrow rangeof temperatures. Any deviation from the 36–38 1C range leads toan arrest of cell division. Hypothermia has been documented ascausing cell cycle arrest and a downregulation of cell metabolicactivities; as such, hypothermia is clinically used as a protectionagainst hypoxia during cardiac surgery or an ischemic infarct[1,2]. However, the downregulation of cellular metabolic activitiesstrongly depends on the temperature, duration of exposure andcell type. Chinese Hamster Ovary (CHO) cells, for example, remainmetabolically active and produce more antibodies when cooled to35-36 1C [3]. So far, most hypothermia studies have been per-formed with either myocytes or nerve cells [2]. Very little isknown about bone cells and how they react to hypothermia.Osteoblasts are responsible for extracellular matrix and boneformation. Therefore, disruptions in osteoblast function canchange the balance between bone formation and resorption,ultimately leading to a loss of bone mass or delayed fracturehealing. A recent study hypothesized that the observed drop inthe core body temperature of elderly adults (aged 465) maycontribute to the loss of bone mass in this age group [4]. Bone lossin humans is a complicated process, as it is influenced by manyfactors such as nutritional status and systemic acting hormones,among others. The delayed reunion of fractures or the lack ofintegration or rejection of implants is associated with increasedheat production. Thus, cooling the fracture or operation areamight be beneficial. Currently, little is known about the cellularand molecular responses of osteoblasts to hypothermia.

Therefore, understanding the cellular and molecular mechan-isms involved in the response of osteoblasts to hypothermia isnecessary. Various cell culture studies have shown that hypother-mia affects the integrity of the cell cytoskeleton, metabolism,cellular proliferation, mRNA transcription and protein translation[5,6]. Proper assembly of the cell cytoskeleton is necessary for theorganization of the protein synthesis machinery to performaccurate protein translation. During hypothermia, cell membranepermeability is altered, leading to a decrease in intracellular pH.Moreover, hypothermia modifies the interaction of actin binding-1α elongation factor (EF1α) with the actin cytoskeleton throughthe change in pH, which in turn affects protein translation [7,8].Depolymerization of the actin fibers has been associated with areduction of protein synthesis [7,9]. An in vitro study has shownthat mild hypothermia increases the polymerization of actinfibers while accelerating protein production [10]. However, thereverse is seen with regards to microtubules. Tubulin fibers aredepolymerized under severe hypothermia, leading to mitoticspindle deactivation [7].The link between the actin and tubulinfibers is important for the maintenance of an intact cytoskeleton,which in turn is crucial for protein synthesis, especially duringhypothermic conditions.

Similar to the heat shock response, cells exposed to hypother-mia immediately elicit an adaptation process by expressing cold-

shock proteins (CSPs) [6,11], particularly the RNA binding motifprotein 3 (Rbm3) [12]. By binding to the newly synthesized mRNAstrands, Rbm3 stabilizes the mRNA and facilitates the translationof the newly formed mRNAs into proteins at low temperatures[12]. Rbm3 binds to the 50 end of the mRNA molecules and speedsup the recruitment and assembly of the 40S and 60S ribosomalsubunits needed for protein synthesis [13]. Moreover, posttransla-tional modification processes required for the production offunctional proteins have been shown to be enhanced underhypothermic conditions [6,11]. Therefore, cells are able to effi-ciently tolerate hypothermia by reducing the amount of mRNAdegradation while increasing mRNA translation.Eukaryotic transcription is highly dependent on the action of

transcription factors for the initiation of transcription and thestabilization of the transcription bubble. During hypothermia, it isimportant that the mRNA molecules are stable, and the transcrip-tion factors involved in osteoblast differentiation and function arenot affected. During bone development, sequential expression ofseveral transcription factors is required to activate the osteoblastdifferentiation pathway. Runt-related transcription factor 2(Runx2) mRNA was shown to be expressed at several early stagesof osteoblast differentiation [14]. Runx2 directs pre-osteoblastcells into immature osteoblasts, which then leads to the promoteractivation and expression of the bone phenotypic protein genes,such as type-I collagen (Col1a1), alkaline phosphatase (Alp),osteocalcin (Ocn), osteopontin (Opn), and bone sialoprotein(Bsp) [15,16]. Nevertheless, the expression levels of the bonematrix protein genes vary according to the stages of osteoblastmaturation. Immature osteoblasts express the COL1A1 and OPNproteins, while early and late mature osteoblasts strongly expressOCN protein[17]. Although immature osteoblasts express bonematrix proteins, the cells are unable to induce bone mineraliza-tion. In the later stages of osteoblast differentiation, osterix isrequired for the expression of the mature osteoblast phenotypemarkers, which are necessary for fully functioning osteoblasts.Osterix activates the expression of OCN, which is responsible forbone matrix mineralization [18]. The importance of osterix wasdemonstrated by the fact that osterix-null mice exhibited no bonemineralization. Similarly, osterix expression was completely lostin Runx2-null mice, suggesting that osterix functions downstreamof Runx2 during osteoblast differentiation [19]. Given that Runx2and osterix are transcriptional activators for the bone matrixprotein genes, severe bone abnormalities have been observed inRunx2 and osterix-deficient mouse models [20,21].Therefore, we investigated the different responses of osteo-

blasts to acute, short-term, and long-term exposures to moderateand severe hypothermia, as cellular stress responses and copingmechanisms strongly depend on the time of exposure, as well asthe severity of the stressor. Of special interest were the effects ofhypothermia on the organization of the cell cytoskeleton, themetabolic activity of osteoblasts, and cell viability were examined.We also measured the expression of the NHOst transcriptionfactors, cold and heat shock proteins and bone function markers

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 648

after exposure to moderate (35 1C) and severe hypothermia(27 1C) for 1, 12, 24, and 72 h.

Material and methods

Reagents

NHOst cells, Osteoblast Basal Medium™ (OBM), fetal bovineserum (FBS), ascorbic acid, gentamicin and amphotericin-B werepurchased from Lonza (USA). Dimethylsulfoxide (DMSO), Tritons

X-100, phosphate buffered saline (PBS), α-anti tubulin-FITC anti-body, paraformaldehyde, bovine serum albumin (BSA), Accutaseand Trypan blue dye were purchased from Sigma-Aldrichs (USA).Additionally, 4,6-diamidino-2-phenylindole (DAPI), Prolongs

Gold, and Alexa Fluors 635 phalloidin were purchased fromInvitrogen (USA). CellTiter 96s Aqueous One Solution was pur-chased from Promega (USA),and the RNeasys Mini kit andSensiscripts RT kit were purchased from Qiagen (Germany).The iQ™ SYBRs Green Supermix was purchased from Bio-Rad (USA).

Cell culture

NHOst cells were cultured in T-25 Flasks (Orange Scientific,Belgium) containing OBM supplemented with 10% FBS, 0.1%ascorbic acid, 0.1% gentamicin and 0.1% amphotericin-B. The cellswere grown at 37 1C with 5% CO2. Confluent NHOst monolayerswere passaged every 5–6 days, using Accutase for cell detach-ment. Cells at passage 8 were used for all experimental assays.

Hypothermia treatment

A total of 3.75�105 cells/cm2 were seeded per T-75 flask andpreincubated at 37 1C for 24 h. The cells were then exposed todifferent temperatures, including moderate (35 1C) and severe(27 1C) hypothermia for 1, 12, 24, and 72 h in a water-jacketedincubator (Binder, Germany) with 5% CO2; physiological tempera-ture was used as a control (37 1C). Each temperature experimentwas carried out in triplicate (n¼3).

Cytoskeletal changes

After each temperature treatment, NHOst cells that were culturedon Lab-Tek II chamber slides (Thermo Fisher Scientific Inc., USA)were fixed with 4% paraformaldehyde in PBS for 30 min at roomtemperature. The fixed cells were incubated with DAPI (300 nM inPBS) for 20 min to stain the nucleus and then permeabilized with0.1% Triton X-100 in PBS for 10 min. Alexa Fluors 635 phalloidin(6.6 mM in PBS) and the anti-α-tubulin-FITC conjugated antibodydiluted 1:50 in blocking solution (10% BSA and 1% Triton X-100 inPBS) were added to the permeabilized cells and incubated for 1 hto stain the actin and tubulin fibers respectively. After 1 h, thecells were washed twice with PBS and counterstained with DAPI(300 nM in PBS) for 30 min. The chambers were removed fromthe slides, and cover slips were mounted with the Prolongs GoldAntifade reagent to minimize photobleaching. The slides werestored at 4 1C in the dark until they were scanned using a confocallaser scanning microscope (CLSM) (Leica TCS SP5). A total of 30cells (n¼30) were analyzed using Leica QWin software. The

fluorescence intensity of the actin (excitation: 633 nm, emission:647 nm) and tubulin (excitation: 494 nm, emission: 518 nm)fibers were measured according to their wavelengths.

Cell viability assay

NHOst cell viability was determined using an MTS-based assay byadding CellTiter 96s Aqueous One solution to the culturemedium. Briefly, 2.5�104 cells per well were cultured in 96-well plates (Techno Plastic Products, Switzerland). CellTiter 96s

Aqueous One Solution (20 ml) was added to the medium in eachwell (1:5), and the plates were incubated at 37 1C for 2 h in ahumidified 5% CO2 incubator. The formation of formazan bymitochondrial reduction was determined by measuring theabsorbance at 490 nm using a microplate reader (Biomeks FX,Beckman Coulter). Blank measurements were obtained from wellswithout cells and were used as a reference reading.

Gene expression

Total RNA was extracted from treated NHOst cells at each timepoint using the RNeasys Mini kit. The cells were first lysed andhomogenized using a QIAshredder column (Qiagen, Germany).Total RNA was isolated and on-column genomic DNA contamina-tion was eliminated using the RNase-free DNase kit (Qiagen,Germany) according to the manufacturer's protocol. The concen-tration and purity of the eluted total RNA was determined using aNanoDropspectrophotometer (NanoDrop ND-1000). Complemen-tary DNA (cDNA) synthesis was performed using 40 ng of totalRNA as a template in a reaction volume of 20 ml; Sensiscript RT(Qiagen, Germany) and oligo-dT primers were used for real-time PCR.

Real-time quantitative PCR

The expression levels of Hsp70 and Rbm3 were analyzed usingreal-time quantitative PCR (RT-qPCR). A cocktail of 1� iQ™ SYBRGreen Supermix, 25 mM forward and reverse primers (Table 1),and 2 ml of cDNA template in a 25 ml volume was amplified usingthe iQ™ 5 Real-Time PCR detection system (Bio-Rad, USA). Theexpression level of each gene was normalized to the expressionlevels of GAPDH and β-actin as reference genes using iQ™ 5optical system software, version 2.0 (Bio-Rad, USA). Significantdifferences were identified using Student's t-test between thebiological replicates (n¼3) for the control and treatment groups.Our RT-qPCR experiment was conducted in accordance with andadhered to the MIQE guidelines [22].

RT2 Profiler PCR Array system

The expression of the osteoblast transcription markers Runx2(Catalog Num: PPH01897) and osterix (Catalog Num: PPH00705)were analyzed using the RT2 Profiler PCR Array system. The RT2

First Strand Kit (Qiagen, Germany) was used to eliminate thegenomic DNA present in the total RNA samples and to transcribethe RNA into cDNA. Genomic DNA Elimination mixture wasprepared by adding 1.5 mg of total RNA into 2.0 ml of 5X gDNAElimination Buffer; RNase-free water was added to a final volumeof 10 ml. The mixture was incubated at 42 1C for 5 min on athermal block. The reaction was immediately terminated by

Table 1 – Real-time PCR primer sequences.

Gene Primer sequence Amplicon size GenBank

Hsp70 F: 50–TTTTACCACTGAGCAAGTGACTG-30 106 NM_002154R: 50- ACAAGGAACCGAAACAACACA-30

Rbm3 F: 50- GAGGGCTCAACTTTAACACCG-30 169 NM_006743R: 50- ATGGCTCTCATGGCAACTGAA-30

GAPDH F: 50 - TGCACCACCAACTGCTTAGC- 30 197 NM_002046R: 50 - GGCATGGACTGTGGTCATGAG-30

β-actin F: 50- CCTTCCTCCCTGGGCATGGA-30 250 NM_001101R: 50 - CACACGGAGTACTTGCGCTCA-30

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 6 49

storing the mixture tubes on ice for at least 1 min. Reversetranscription (RT) cocktail (10 ml) was prepared by adding 5C RTbuffer 3, primer and external control mix, RT enzyme mix 3 andRNase-free water together. The RT cocktail was added to 10 ml ofthe genomic DNA elimination mixture. The mixture was thenincubated on a thermal block for precisely 15 min at 42 1C,followed by an immediate reaction termination by heating at95 1C for 5 min. The first strand cDNA synthesis mixture wasdiluted by adding 91 ml of RNase-free water. A volume of 25 ml/reaction of the RT2 qPCR master mix was prepared by adding 2XSABiosciences RT2 qPCR master mixes, diluted first strand cDNAsynthesis reaction and RNase-free water. A total of 25 ml of mastermix was loaded onto the Custom RT² PCR Array plate andamplified using the iQ™ 5 Real-Time PCR detection system (Bio-Rad, USA). Gene expression was normalized to the expression ofGAPDH (Catalog Num: PPH00150) and Hprt1 (Catalog Num:PPH01018) as reference genes using the RT2 Profiler PCR ArrayData Analysis Version 3.5 software. Significant differences wereidentified using Student's t-test from the biological replicate ΔΔCTvalues for each gene in the control and treatment groups (n¼3).

Alkaline phosphatase activity

Alkaline phosphatase activity was measured using the alkalinephosphatase colorimetric assay kit (BioVision, USA) in which thetreated sample supernatant was directly used to measure thealkaline phosphatase activity (U/mL). A volume of 60 ml of treatedsample supernatant was added to each well, and 20 ml of assaybuffer was then added to the wells. pNPP solution (50 ml of 5 mMsolution) was added to the sample and blank wells. The reactionwas incubated for 60 min at room temperature and was thenhalted by the addition of 20 ml of stop solution. The O.D. wasmeasured at 405 nm using a microplate reader. The amount ofpNP produced by ALP from the control and treated samples wascompared to the standard curve and then applied to the followingequation to generate the ALP activity: ALP activity (U/mL)¼A/V/T,where A is the amount of pNP produced by a sample (mmol), V isthe volume of the sample/well (mL), and T is the reactiontime (min).

Human osteocalcin

Osteocalcin (OCN) was quantitatively measured using the humanOsteocalcin Instant ELISA kit (Bender MedSystems™, Austria). Ineach control and sample well, 25 ml of the supernatant was addedto 100 ml of distilled water. The plate was then incubated for 2 hat room temperature on a plate shaker at 100 rpm. After

incubation, the wells were emptied and rinsed three times with400 ml of Wash Buffer. Then, 100 ml of TMB Substrate Solution wasadded to the wells and incubated at room temperature for 15 min.A total of 100 ml of stop solution was then added to all wellsto stop the enzymatic reaction. The absorbance of each wellwas read at 450 nm (620 nm as reference wavelength) using amicroplate reader. Human OCN concentration was read from thestandard curve.

Bone mineralization

NHOst cells were culture in a 6-well plate with a total of 1.0�105

cells per well with OBM. After 1 week, cells were cultured withOBM supplemented with hydrocortisone-21-hemisuccinate(200 nM) and beta-glycerophosphate (10 nM). By using a water-jacketed incubator (Binder, Germany), the cells were exposed tomoderate hypothermia at 35 1C and 37 1C served as a control.After stimulation for 1 week, culture medium was discarded andmultilayer cells were washed with PBS. Next, NHOst cells werestained with silver nitrate (Merck, Millipore) and plates wereplaced under ultraviolet light for 2 h. Finally cells were washedwith PBS and presence of mineral deposition in a form of noduleswas visualized under the light microscope.

Statistical analysis

Statistical analyses were performed using the statistical packageSPSS (Statistical Package for the Social Sciences, version 17.0)software employing the independent-sample t-test to determinethe level of significance between the control and treatmentgroups. Statistically significant levels were set at nPo0.05, nnPo0.01, and nnnPo 0.001. All results are presented as the mean7standard deviation (SD).

Results

Hypothermia reorganized NHOst cell cytoskeleton

Cell shape and movement are highly dependent on the composi-tion and arrangement of the cell cytoskeleton. Our CLSM resultsshowed disturbances in the NHOst cell cytoskeleton in the form ofcell shape changes, fiber redistribution and intensity. Compared tocells grown under physiological temperature (Fig. 1a), exposure tohypothermia resulted in the perturbation of the NHOst cytoske-leton (Fig. 1b and c). NHOst cells exposed to hypothermia were

Hypothermia Control (37°C)

1h

12h

24h

72h

Moderate (35°C) Severe (27°C)

Fig. 1 – Fixed Normal Human Osteoblast cells (NHOst) were stained with Anti-α-Tubulin-FITC; tubulin (green), Alexa Fluors635phalloidin; actin (red), and DAPI; nucleus (blue). Change in cell cytoskeleton during control (a), moderate hypothermia (b), andsevere hypothermia (c) was viewed under Confocal Laser Scanning microscope at 40x magnification (scale bar: 75 lm) after 1, 12,24, and 72 h of treatment.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 650

larger and possessed a spindle-like shape, and these cells becameincreasingly irregular in shape after 12, 24 and 72 h exposure.Tubulin fibers increased after 1 h of moderate hypothermiacompared to the control. In severe hypothermia, tubulin fiberswere localized around the NHOst nucleus after 12 and 24 h.Despite the cytoskeletal changes, NHOst cells showed no mor-phological signs of apoptosis or necrosis under moderate andsevere hypothermia at any time point of exposure.Both moderate and severe hypothermia significantly increased

the fluorescence of actin fibers (po0.001), except for exposure to27 1C after 12 h (Fig. 2a). The fluorescence intensity of actinpeaked at 24 h for the 35 1C treatment and at 1 h for the 27 1Ctreatment, which both had the same intensity value (186%). Undermoderate hypothermia, the tubulin intensity significantly

increased (po0.001) from 1 to 24 h before gradually decreasingto the level of the control at 72 h (Fig. 2b). Exposure to severehypothermia for 12 h significantly reduced the intensity of thetubulin fibers. No significant differences were observed at theother time points (Fig. 2a).

The severity of hypothermia affects NHOst mitochondrialactivity

Metabolically active cells are able to reduce the MTS tetrazoliumcompound [3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfophenyl)-2 H-tetrazolium] to formazan, which isbrown in color. This reduction reaction was used to measure themetabolic activities of the cells. Acute (1 h) and prolonged (72 h)

0

20

40

60

80

100

120

140

160

180

200

1 h 12 h 24 h 72 h

% o

f act

in in

tens

ity

Duration (h)

Control35°C 27°C

35°C 27°C

0

20

40

60

80

100

120

140

160

180

200

1 h 12 h 24 h 72 h

% o

f tub

ulin

inte

nsity

Duration (h)

Control

***

***

***

***

****** *** ***

***

***

***

***

Fig. 2 – Exposure to hypothermia conditions significantly increased actin fluorescence intensity between 1 and 72 h.Simultaneously, acute and short exposure to moderate hypothermia between 1 and 24 h significantly increased tubulinfluorescence intensity. Cells stained with Alexa Fluors 635 Phalloidin and Anti-α-tubulin-FITC after 1, 12, 24 and 72 h for actin (a)and tubulin (b) fibers respectively was analyzed using Leica QWin software. Data obtained from 50 cells (n¼50) is shown as themean value7SD of percentage change relative to control (100%) using Independent-sample t-test. nnn: P o 0.001.

0

20

40

60

80

100

120

140

1 h 12 h 24 h 72 h

% o

f mito

chon

dria

l act

ivity

Duration (h)

Control35°C27°C

***

******

***

***

***

Fig. 3 – Hypothermia modified NHOst mitochondrial activityaccording to temperature severity. Production of formazanwas increased under moderate hypothermia and reducedunder severe hypothermia between 1 and 72 h. Formazanabsorbance was measured at 490 nm. Results represent themeans7SD of six replicate samples (n¼6). nnn: P o 0.001.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 6 51

exposure of NHOst cells to moderate hypothermia (35 1C) sig-nificantly increased the amount of formazan formed; therefore,these cells exhibited increased mitochondrial activity. Other timepoints (12 and 24 h) showed no significant differences (Fig. 3).The metabolic activity under moderate hypothermia conditionsincreased by 17.2% (po0.001) and 17% (po0.001) at 1 h and 72 h,respectively, when compared to the control. Severe hypothermiaresulted in a significant inhibition of mitochondrial activity(po0.001) compared to the control. Metabolic activity wasreduced by 35, 42, 51, and 15% at 1, 12, 24, and 72 h, respectively.

Expression of Rbm3 and Hsp70 increased underhypothermia conditions

The expression of the cold shock RNA chaperone protein Rbm3 wasconstitutively enhanced under moderate and severe hypothermiaduring all but the 1 h treatments (Fig. 4a). Rbm3 mRNA was highlyupregulated at 12 h by 2.5270.43-fold and 2.6270.44-fold for35 1C and 27 1C, respectively (po0.001). However, our studiesshowed that acute exposure to hypothermia for 1 h significantlydecreased the expression of Hsp70 (Fig. 4b) by 0.3670.16-fold and0.4270.17-fold for 35 and 27 1C, respectively (po0.001). Shortexposure to hypothermia for 12 h and 24 h increased the expres-sion of Hsp70 mRNA. Although the expression was upregulated, agradual decrease in expression was observed from 12 h onwards.At 72 h, Hsp70 expression was downregulated in both the moder-ate and severe hypothermia samples.

Hypothermia affects the expression of osteogenictranscription factors

The key osteoblast transcription regulators, Runx2 and osterix,that are involved in early osteogenesis and late mineralization,respectively, were measured using the RT2 Profiler PCR ArraySystem. Acute hypothermia (1 h) significantly downregulatedRunx2 expression by 0.2070.15-fold (po0.01) and 0.3670.27-fold

(po0.01) for 35 1C and 27 1C, correspondingly (Fig. 5a). Runx2expression showed a V-shaped curve increase in expression undermoderate hypothermia conditions between 12 h (po0.05) and 72 h(po0.01), and Runx2 expression significantly peaked under 27 1Cat 12 h by 12.970.18-fold (po0.001). However, the expressiongradually returned to basal levels by 24 h and was finally reducedat 72 h (po0.05) relative to control under severe hypothermiaconditions. Our studies showed that osterix was significantlyupregulated under severe hypothermia (Fig. 5b) with the highestfold change at 24 h (6.970.29-fold). Expression of osterix after 1 hunder moderate hypothermia was significantly reduced by0.2370.13-fold (po0.01). Prolonged exposure to moderatehypothermia resulted in the high expression of osterix between12 h to 72 h.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

1 h 12 h 24 h 72 h

Hsp

70 F

old

Cha

nge

Duration (h)

Control35°C27°C

35°C27°C

0

0.5

1

1.5

2

2.5

3

3.5

1 h 12 h 24 h 72 h

Rbm

3 Fo

ld C

hang

e

Duration (h)

Control

******

******

******

*****

**

********

Fig. 4 – NHOst cells responded to hypothermia conditioning by upregulating cold shock chaperone protein, Rbm3 (a). Rbm3 mRNAwas consecutively expressed after 12 h and up to 72 h. Short exposure to moderate and severe hypothermia after 12 h significantlyincreased heat shock chaperone protein, Hsp70 (b) between 12 and 24 h. Analysis was performed by Real-time PCR using specificprimer sequence (Table 1). Expression levels for Rbm3 and Hsp70 mRNA were normalized with β-actin and GAPDH. Data areexpressed as mean7SD (n¼6). Statistical analysis by Independent-sample t-test showed significant differences. nn: P o 0.01;nnn: P o 0.001.

0

1

2

3

4

5

6

7

8

1 h 12 h 24 h 72 h

Ost

erix

Fol

d C

hang

e

Duration (h)

Control

0

2

4

6

8

10

12

14

1 h 12 h 24 h 72 h

Run

x2 F

old

Cha

nge

Duration (h)

Control35°C27°C

** **

**

**

*

*

*

******

**

**

*

35°C27°C

Fig. 5 – Osteoblast transcription factor Runx2 (a) and osterix (b) mRNAwas induced under hypothermic conditions between 12 and72 h. Expression of Runx2 peaked after 12 h of severe hypothermia while osterix after 24 h. Analysis was performed by RT2 ProfilerPCR Array System. Data are expressed as means7SD (n¼3). Statistical analysis by Independent-sample t-test showed significantdifferences. n: P o 0.05, nn: P o 0.01, nnn: P o 0.001.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 652

Hypothermia retained the osteoblastic properties

Osteoblast function was assessed by measuring ALP activity andosteocalcin protein expression. As shown in Fig. 6a, NHOst cellsshowed similar patterns of activity under moderate and severehypothermia. After 1 h at 35 1C and 27 1C, ALP activity was slightlyreduced. However, after 12 h, ALP activity was significantlyincreased (po0.001) compared to the 37 1C temperature controls.Prolonged exposure to moderate and severe hypothermia sug-gests that osteoblastic functional marker with regards to tissuemineralization was enhanced. ALP activity increased andremained significantly elevated for up to 72 h compared to the

control. However, OCN expression was reduced after 1 h undermoderate (5.070.012 ng/mL) hypothermia compared to the con-trol (Fig. 6b). The production of OCN after 12 h (5.670.013 ng/mL)and 24 h (6.070.025 ng/mL) remained increased under moderatehypothermia relative to the control concentration. The OCNexpression level after 72 h of exposure to moderate hypothermiawas on par with the OCN protein expression level in the controlcells. Meanwhile under severe hypothermia, acute 1 h exposureincreased OCN protein by 9.070.080 ng/mL and remained increaseafter 12 h (8.070.063 ng/mL) and 72 h (6.870.042 ng/mL). Level ofOCN protein returned to basal level after 72 h of severe hypother-mia exposure.

0

1

2

3

4

5

6

7

8

9

10

1 h 12 h 24 h 72 h

Ost

eoca

lcin

(ng/

mL

)

Duration (h)

0

0.001

0.002

0.003

0.004

0.005

0.006

1 h 12 h 24 h 72 h

AL

P ac

tivity

(U/m

L)

Duration (h)

Control35°C 27°C

*

**

** **

**

** *

Control35°C 27°C

Fig. 6 – Hypothermia retained osteoblastic properties by enhancing the production of alkaline phosphatase activity (a) andosteocalcin protein (b) between 12 and 72 h. Data are expressed as mean7SD (n¼6) with significant differences. Amount of pNPproduced by ALP from control and treated samples was applied to [ALP activity (U/mL)¼pNP produced (lmol)/sample volume (mL)/ reaction time (min)] equation to generate ALP activity. Statistical analysis by Independent-sample t-test showed significantdifferences. n: P o 0.05, nn: P o 0.01.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 6 53

Moderate hypothermia stimulated bonemineralization

NHOst cells were stained with von Kossa to observe for anymineral deposition (Fig. 7). After stimulating osteoblast cells withhydrocortisone-21-hemisuccinate and beta-glycerophosphate for1 week, pronounced staining pattern was observed comprising ofnodule-like areas. Larger nodule-like area was observed undercontrol condition while smaller nodule-like area was observedunder moderate hypothermia after 1 week. Although stained areaunder moderate hypothermia was slightly smaller than control,the amount of nodule-like area was more.

Discussion

In contrast to cells grown at physiological temperature (37 1C),the cells under hypothermic conditions show reduced metabolicactivities, resulting in reduced ATP production, cytoskeletalchanges, and the suppression of mRNA and protein synthesis[6]. Numerous studies have shown that the cell cytoskeleton playsa role in managing the protein synthesis machinery in mamma-lian cells [5,6,23,24]. Our cell culture experiments showed thatNHOst cells treated with hypothermia had an altered cytoskele-ton, whereas the severity of the changes depended on the severityand duration of the hypothermia treatment. These results are inaccordance with studies performed in CHO cells, where thecytoskeletal architecture was minimally disturbed after exposureto moderate hypothermia [23]. Consistent with our findings, CHOcells under severe hypothermic conditions (28 1C) exhibitedtubulin fibre disassembly, which corresponded to a reduction inprotein synthesis [23].

A well-organized and functional cell cytoskeleton is essentialfor cell adhesion, motility, proliferation, cell-to-cell and cell-to-ECM signaling interactions[25]. We showed that the actin inten-sity increased during both moderate (35 1C) and severe (27 1C)

hypothermia. These results showed that the actin fibers were ableto polymerize under moderate and severe hypothermia, whichcontradicts results from other studies [7,9]. However, studies onblood platelets showed that hypothermia did not affect the actinfibers in the platelets [10]. It has been proposed that cells exposedto hypothermia respond by disassembling their cytoskeleton,which leads to an overall suppression of protein synthesis. Anincreasing number of studies have proven that protein translationcomponents, such as ribosomes, eukaryotic initiation factors(eIFs), aminoacyl-tRNAs, and mRNAs, are co-localized with thecytoskeleton components, particularly the actin and tubulin fibers[8]. Mammalian cells subjected to mild hypothermia did notundergo cytoskeleton disassembly. Similarly, our findings showedthat the actin and tubulin fibers remained polymerized underprolonged exposure to moderate hypothermic conditions (35 1C).However, other studies suggested that protein synthesis underhypothermia conditions might be limited due to the de/phos-phorylation of the initiation and elongation factors. For instance,elongation factor 1α (EF1α) serves as a linker between the proteintranslation components (mRNAs) and the cytoskeleton. EF1α non-selectively binds the newly formed mRNAs to the cytoskeletonthrough the F-actin polymer [23]. At the same time, EF1α binds tothe adapter molecular of the amino acid-carrying tRNA and bringsit to the ribosome to facilitate the translation of the nascentmRNAinto protein polymers [26].Maintaining tubulin organization is important and required for

the formation of microtubules, which are involved in cell division[7] and intracellular trafficking [24]. Our results showed anincrease in the tubulin intensities from 1 to 24 h, confirming thatmore tubulin fibers were formed under moderate hypothermia.Bosc and Job (2003) showed that the formation of stable-tubule-only polypeptides (STOP) was induced under moderate hypother-mia conditions [27]. This microtubule-associated protein isinvolved in stabilizing the microtubule network. In addition,protein kinase CK2 has been reported to stabilize the polymerizedmicrotubules from hypothermia induced depolymerization [28].

Control

Moderate hypothermia

35°C

6-well plate micrograph

Fig. 7 – Moderate hypothermia induced bone mineralization after 1 week. NHOst cells stimulated with hydrocortisone-21-hemisuccinate and beta-glycerophosphate was stained with silver nitrate and exposed to ultraviolet light. Cells were then viewedunder the microscope to observe for mineralized nodules.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 654

CHO cells exposed to severe hypothermic conditions (28 1C)initiated the disassembly of the tubulin fibers. Our findingsconfirmed that the tubulin fibers were affected by severehypothermia as early as 1 h after exposure. As moderatehypothermia is considered to be within the normal pathophysio-logical body temperature range [29], the higher intensity value at35 1C suggests that more actin and tubulin fibers were formedcompared to the control at 37 1C.Recent studies have suggested that severe hypothermia results

in a reduction in mitochondrial metabolism, and therefore,a reduction in available ATP, which is required for many cellfunctions, particularly cell division [30,31]. As observed in rathepatoma cells, severe hypothermia (25 1C) halted cell prolifera-tion by 95% after 48 h exposure [32]. Hypothermia slows downthe enzymatic reactions by reducing ATP consumption in themitochondria, contributing to low cellular metabolic activity [33].The depletion of mitochondrial energy directly decreases themitochondrial activity. We were able to show that under moder-ate hypothermia, NHOst cells remained metabolically active after1, 12 and 72 h. The NHOst cells exposed to moderate hypothermiawere able to maintain cytoskeleton integrity by increasing theactin and tubulin fibers, but also showed an increase in themitochondrial activity compared to the NHOst cells grown atphysiological conditions. However, less formazan formation wasobserved after exposure to severe hypothermia, and the NHOstmitochondrial activity was suppressed after 1 h exposure. NHOstmitochondrial activity remained reduced up to 72 h, suggestingthat the cells could be arrested in either quiescence or senescencestage, or were in fact dead. We hypothesized that the decrease in

mitochondrial activity was most likely due to the decrease inenergy consumption, as protein synthesis is usually inhibitedduring severe hypothermia [34].

Our results showed that NHOst cell metabolism was affectedaccording to the degree of severity of the hypothermia and thatthis was associated with an increase in the regulation of the coldshock protein Rbm3. Temperature downshift is known to causesecondary structure formation in mRNA molecules, resulting inthe masking of the 50- region (UTR) and the suppression of mRNAtranslation [11]. It has been proposed that under hypothermicconditions, Rbm3 binds to mRNA to stabilize the molecule and toallow the mRNA to bind to the 40 S ribosome unit; this binding isthen followed by the assembly of the total ribosome and thetranslation of mRNA under hypothermic conditions [35,13,36,37].Rbm3 may have additional functions in cancer cells. Highexpression of Rbm3 was reported in cervical carcinoma, renalcarcinoma, bladder carcinoma, hepatoma, and leukemia cells linesexposed to hypothermia [29].

The response to hypothermia-induced heat shock proteins (Hsp)is rarely examined. The molecular chaperone Hsp70 is expressed atbasal levels in non-stressed cells and is induced during cellularstress to aid in the response. Although the main function of Hsp70is to ensure proper protein folding, some additional functions havebeen associated with Hsp70, especially in cancer cells [38,39].The few studies that investigated the effects of hypothermia on Hspproteins were performed in brown adipose tissue (BAT) from mice.Hypothermia-induced Hsp70 in BAT was associated with theelevated binding of the heat shock transcription factors (HSF1) tothe DNA, resulting in Hsp70 gene transcription [40]. Our studies

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 6 55

showed that NHOst cells significantly upregulated Hsp70 undermild hypothermic conditions (35 1C) after 12 and 24 h. In nature,hypothermia imposes cellular stress by denaturing the proteinstructure and disrupting its functions[38,41]. Our results suggestedthat the induction of Hsp70 under moderate and severe hypother-mia after 12 and 24 h is part of the cold shock response to assist thefolding of the new proteins, as well as to repair the misfoldedproteins[38,42]. Similarly, studies have shown that the expressionof the molecular chaperone proteins Rbm3 and Hsp70 weredownregulated after 1 h under moderate and severe hypothermia.We suggest that the acute response to hypothermic conditionsreduced the overall gene expression to preserve energy and toreduce the number of misfolded proteins, thereby preventingfurther hypothermia-induced damage.

Regulation of the Runx2 and osterix transcription factors isnecessary for the transcription of bone matrix protein genes atdifferent stages of osteoblast maturation. Studies have shown thatRUNX2 protein is strongly expressed in immature osteoblastspositive for OPN protein and in early matured osteoblasts positivefor OCN protein. However, in late mature osteoblasts positive forOCN, hypothermia resulted in the deregulation of the RUNX2protein[17].Studies have suggested that low levels of Runx2mRNA are required to maintain COL1A1 and OCN protein expres-sion in mature osteoblasts. Our studies showed that Runx2 wasconstitutively expressed along with an increase in ALP activityand OCN protein, suggesting that NHOst cells are in an immaturestage under hypothermic conditions. Conversely, a recent studyby Patel et al.[43] showed that prolonged hypothermia (34–35.5 1C) caused a malfunction/decrease of Runx2 in rat calvarialosteoblast cells; this reduction in turn resulted in the loss of ALPand OCN, and therefore, a loss of osteoblast function. The reasonfor the contradicting results could be due to the differences in theduration and severity of the hypothermia exposure.

Studies from Nakashima et al. showed that osterix-deficientmice are able to express Runx2, while Runx2-deficient mice werenot able to produce osterix, leading to bone deformation [19]. Thisindicates that osterix plays a role downstream of Runx2 duringosteoblast differentiation[15]. Runx2 and osterix may reside in thesame pathway or may control osteoblast differentiation indepen-dently. Our studies showed that the upregulation of Runx2 mRNAcorrelated with an upregulation of osterix mRNA; this finding hasbeen confirmed by other studies [19]. Our studies also showedthat exposure to hypothermia regulated osterix mRNA, suggestingthat NHOst cells were able to differentiate into functionalosteoblasts, as determined by the accumulation of OCN protein.Because Runx2 and osterix are expressed at the early and lateosteoblast maturation stages, respectively, we predict that ourNHOst culture consisted of both immature and mature NHOstcells. Although Runx2 and osterix are required for the transcrip-tion of the bone matrix protein genes, the maintenance of theproteins by Runx2 and osterix is not essential.

We also showed that Rbm3 not only functioned as an mRNAstabilizer, but also enhanced the transcriptional process. Studieshave confirmed that Rbm3 is able to bind to a specific region inthe promoter of target genes, leading to an increase in transcrip-tion [44]. This was further confirmed by our results that showedthat over expression of Runx2 and osterix resulted in increasedtranscription and translation of the OCN protein. Nevertheless, wesuggest that hypothermia permitted the maturation of NHOstcells through an osterix-dependent pathway in vitro. On the other

hand, we were able to demonstrate mineralized bone areas after 1week of moderate hypothermia (Fig. 7). On the contrary, in studiesperformed by Patel et al., bone mineralization was stronglyinhibited when exposed to 34 1C over a period of 14–16 days[43].The authors elaborated that the degree of mineralization inhibitionwas dependent on the hypothermic severity and duration. Patelet al., demonstrated that longer exposure to hypothermia decreasesthe rate of mineralization between 50–95% [43].Our in vitro studies suggested that the normal osteoblast

cellular function is dependent on the severity and longevity ofhypothermia exposure. NHOst cells under moderate hypothermiawere able to retain intact cytoskeleton components that arecrucial for cell motility, intra- and extracellular signaling, prolif-eration, and physiological functions, particularly after 12 h expo-sure. We showed that short-term exposure to moderatehypothermia increased enzymatic reactions, which was reflectedby increased NHOst mitochondrial activity. In our study, theinduction of the temperature chaperone proteins resulted in thestabilization of mRNA and in increased transcription of bonefunctional markers, such as ALP activity and OCN.Because our finding showed that normal bone function was

maintained under moderate hypothermia, we hypothesize thathypothermia at 35 1C can be used therapeutically to help cooldown fracture site and may prevent the rejection of boneimplants. Nevertheless, the application of moderate hypothermiafor bone fracture repair in vivo may result in different responsesthan those observed in our in vitro study.

Conflict of interest

All authors have no conflicts of interest.

Acknowledgments

Our special thanks go to Professor Dr. Methil Kannan Kutty fromthe Faculty of Medicine, Universiti Teknologi MARA for his helpfuldiscussions and assistance in the preparation of this manuscript.We would also like to thank the staff of the Institute of MedicalMolecular Biotechnology (IMMB),particularly Ms. Salina Othmanfor her excellent operating skills of the Leica Confocal LaserScanning Microscope and Mrs. Norita Salim for her skillfulmolecular technical assistance. This study was funded by theFundamental Research Grant Scheme (FRGS) under the MalaysianMinistry of Higher Education (MOHE). Project code: 600 RMI/ST/FRGS5/3/Fst (183/2010).

r e f e r e n c e s

[1] N M Nikolov, A J Cunningham, Mild therapeutic hypothermia toimprove the neurologic outcome after cardiac arrest, Surv.Anesthesiol. 47 (2003) 219–220.

[2] J Sahuquillo, M Mena, A Vilalta, M Poca, Moderate hypothermiain the management of severe traumatic brain injury: a good ideaproved ineffective?, Curr. Pharmaceu. Des. 10 (2004) 2193–2204.

[3] N Kumar, P Gammell, M Clynes, Proliferation control strategies toimprove productivity and survival during CHO based productionculture, Cytotechnology 53 (2007) 33–46.

[4] R Fox, P M Woodward, A Exton-Smith, M Green, D Donnison,M Wicks, Body temperatures in the elderly: a national study of

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 6 ( 2 0 1 4 ) 4 6 – 5 656

physiological, social, and environmental conditions, Br. Med. J. 1(1973) 200–206.

[5] J Fujita, Cold shock response in mammalian cells, J. Mol. Micro-biol. Biotechnol. 1 (1999) 243–255.

[6] M B Al-Fageeh, C M Smales, Control and regulation of the cellularresponses to cold shock: the responses in yeast and mammaliansystems, Biochem. J. 397 (2006) 247–259.

[7] R Stapulionis, S Kolli, M P Deutscher, Efficient mammalianprotein synthesis requires an intact F-actin system, J. Biol. Chem.272 (1997) 24980–24986.

[8] A Hudder, L Nathanson, M P Deutscher, Organization of mam-malian cytoplasm, Mol. Cell. Biol. 23 (2003) 9318–9326.

[9] L L Gisselsson, A Matus, T Wieloch, Actin redistribution underliesthe sparing effect of mild hypothermia on dendritic spinemorphology after in vitro ischemia, J. Cereb. Blood Flow Metab.25 (2005) 1346–1355.

[10] K M Hoffmeister, H Falet, A Toker, K L Barkalow, T P Stossel,J H Hartwig, Mechanisms of cold-induced platelet actin assem-bly, J. Biol. Chem. 276 (2001) 24751–24759.

[11] M Lleonart, A new generation of proto-oncogenes: cold-inducibleRNA binding proteins, Biochim. Biophys. Acta 1805 (2010) 43–52.

[12] S A Chappell, G C Owens, V P Mauro, A 5 Leader of Rbm3, a coldstress-induced mRNA, mediates internal initiation of translationwith increased efficiency under conditions of mild hypothermia,J. Biol. Chem. 276 (2001) 36917–36922.

[13] J Dresios, A Aschrafi, G C Owens, P W Vanderklish, G M Edelman,V P Mauro, Cold stress-induced protein Rbm3 binds 60 Sribosomal subunits, alters microRNA levels, and enhances globalprotein synthesis, Proc. Natl. Acad. Sci. United States of Am. 102(2005) 1865–1870.

[14] R T Franceschi, C Ge, G Xiao, H Roca, D Jiang, Transcriptionalregulation of osteoblasts, Ann. New York Acad. Sci. 1116 (2007)196–207.

[15] T Matsubara, K Kida, A Yamaguchi, K Hata, F Ichida, H Meguro,H Aburatani, R Nishimura, T Yoneda, BMP2 regulates Osterixthrough Msx2 and Runx2 during osteoblast differentiation, J.Biol. Chem. 283 (2008) 29119–29125.

[16] C-Y Wang, S-F Yang, Z Wang, J-M Tan, S-M Xing, D-C Chen,S-M Xu, W Yuan, PCAF acetylates Runx2 and promotes osteoblastdifferentiation, J. Bone Miner. Metab. 31 (2013) 1–9.

[17] Z Maruyama, C A Yoshida, T Furuichi, N Amizuka, M Ito,R Fukuyama, T Miyazaki, H Kitaura, K Nakamura, T Fujita, Runx2determines bone maturity and turnover rate in postnatal bonedevelopment and is involved in bone loss in estrogen deficiency,Dev. Dyn. 236 (2007) 1876–1890.

[18] K Nakashima, B de Crombrugghe, Transcriptional mechanisms inosteoblast differentiation and bone formation, TRENDS in Genet.19 (2003) 458–466.

[19] K Nakashima, X Zhou, G Kunkel, Z Zhang, J M Deng, R RBehringer, B de Crombrugghe, The novel zinc finger-containingtranscription factor osterix is required for osteoblast differentia-tion and bone formation, Cell 108 (2002) 17–29.

[20] P J Marie, Transcription factors controlling osteoblastogenesis,Arch. Biochem. Biophys. 473 (2008) 98–105.

[21] C Zhang, W Tang, Y Li, F Yang, D R Dowd, P N MacDonald,Osteoblast-Specific Transcription Factor Osterix Increases Vita-min D Receptor Gene Expression in Osteoblasts, PloS one 6(2011) 26504–26512.

[22] S A Bustin, V Benes, J A Garson, J Hellemans, J Huggett, M Kubista,R Mueller, T Nolan, M W Pfaffl, G L Shipley, The MIQE guidelines:minimum information for publication of quantitative real-timePCR experiments, Clin. Chem. 55 (2009) 611–622.

[23] S Becerra, J Berrios, N Osses, C Altamirano, Exploring the effect ofmild hypothermia on CHO cell productivity, Biochem. Eng. J. 60(2011) 1–8.

[24] C Delphin, D Bouvier, M Seggio, E Couriol, Y Saoudi, E Denarier,C Bosc, O Valiron, M Bisbal, I Arnal, MAP6-F is a temperaturesensor that directly binds to and protects microtubules from

cold-induced depolymerization, J. Biol. Chem. 42 (2012) 35127–35138.

[25] J C Bruusgaard, K Liestøl, K Gundersen, Distribution of myonucleiand microtubules in live muscle fibers of young, middle-aged,and old mice, J. Appl. Physiol. 100 (2006) 2024–2030.

[26] G Liu, W M Grant, D Persky, V M Latham, R H Singer, J Condeelis,Interactions of elongation factor 1α with F-actin and β-actinmRNA: implications for anchoring mRNA in cell protrusions, Mol.Biol. Cell 13 (2002) 579–592.

[27] C Bosc, A Andrieux, D Job, STOP proteins, Biochemistry 42 (2003)12125–12132.

[28] A C Lim, S-Y Tiu, Q Li, R Z Qi, Direct regulation of microtubuledynamics by protein kinase CK2, J. Biol. Chem. 279 (2004) 4433–4439.

[29] S Danno, K Itoh, T Matsuda, J Fujita, Decreased expression ofmouse rbm3, a cold-shock protein, in sertoli cells of cryptorchidtestis, Am. J. Pathol. 156 (2000) 1685–1692.

[30] Y Hua, K Hisano, Y Morimoto, Effect of mild and moderatehypothermia on hypoxic injury in nearly pure neuronal culture,J. Anesth. 24 (2010) 726–732.

[31] S Liu, J F Chen, Strategies for therapeutic hypometabothermia,J. Exp. Stroke Transl. Med. 5 (2012) 31–42.

[32] J Van Rijn, J Van den Berg, J Kipp, D Schamhart, R Van Wijk, Effectof hypothermia on cell kinetics and response to hyperthermiaand X rays, Radiat. Res. 101 (1985) 292–305.

[33] R P Pulis, B M Wu, N M Kneteman, T A Churchill, Conservation ofphosphorylation state of cardiac phosphofructokinase duringin vitro hypothermic hypoxia, Am. J. Physiol.-Heart Circ. Physiol.279 (2000) 2151–2158.

[34] L Canevari, A Console, E A Tendi, J B Clark, T E Bates, Effect ofpostischaemic hypothermia on the mitochondrial damageinduced by ischaemia and reperfusion in the gerbil, Brain Res.817 (1999) 241–245.

[35] S A Chappell, G C Owens, V P Mauro, A 50 leader of Rbm3, a coldstress-induced mRNA, mediates internal initiation of translationwith increased efficiency under conditions of mild hypothermia,J. Biol. Chem. 276 (2001) 36917–36922.

[36] Z Xia, X Zheng, H Zheng, X Liu, Z Yang, X Wang, Cold-inducibleRNA-binding protein (CIRP) regulates target mRNA stabilizationin the mouse testis, FEBS letters 586 (2012) 3299–3308.

[37] Y Liu, W Hu, Y Murakawa, J Yin, G Wang, M Landthaler, J Yan,Cold-induced RNA-binding proteins regulate circadian geneexpression by controlling alternative polyadenylation, Sci. Rep. 3(2013) 2054–2065.

[38] K E Cullen, K D Sarge, Characterization of hypothermia-inducedcellular stress response in mouse tissues, J. Biol. Chem. 272(1997) 1742–1746.

[39] T Kaneko, K Kibayashi, Mild hypothermia facilitates the expres-sion of cold-inducible RNA-binding protein and heat shockprotein 70.1 in mouse brain, Brain Res. 1466 (2012) 128–136.

[40] J M Matz, M J Blake, H Tatelman, K P Lavoi, N J Holbrook,Characterization and regulation of cold-induced heat shockprotein expression in mouse brown adipose tissue, Am. J.Physiol.-Regul., Integr. Comp. Physiol. 269 (1995) 38–47.

[41] E Laios, I M Rebeyka, C A Prody, Characterization of cold-inducedheat shock protein expression in neonatal rat cardiomyocytes,Mol. Cell. Biochem. 173 (1997) 153–159.

[42] M Frink, S Flohé, M van Griensven, P Mommsen, F Hildebrand,Facts and fiction: the impact of hypothermia on molecularmechanisms following major challenge, Mediat. Inflamm. 201(2012) 762840–762852.

[43] J J Patel, J C Utting, M L Key, I R Orriss, S E B Taylor, P Whatling,T R Arnett, Hypothermia inhibits osteoblast differentiation andbone formation but stimulates osteoclastogenesis, Exp. Cell Res.318 (2012) 2237–2244.

[44] K Furukawa, K Ohsuye, Effect of culture temperature on arecombinant CHO cell line producing a C-terminal α-amidatingenzyme, Cytotechnology 26 (1998) 153–164.