Proteomic analysis of endothelial cold-adaptation
© Zieger et al; licensee BioMed Central Ltd. 2011
Received: 8 August 2011
Accepted: 22 December 2011
Published: 22 December 2011
Skip to main content
© Zieger et al; licensee BioMed Central Ltd. 2011
Received: 8 August 2011
Accepted: 22 December 2011
Published: 22 December 2011
Understanding how human cells in tissue culture adapt to hypothermia may aid in developing new clinical procedures for improved ischemic and hypothermic protection. Human coronary artery endothelial cells grown to confluence at 37°C and then transferred to 25°C become resistant over time to oxidative stress and injury induced by 0°C storage and rewarming. This protection correlates with an increase in intracellular glutathione at 25°C. To help understand the molecular basis of endothelial cold-adaptation, isolated proteins from cold-adapted (25°C/72 h) and pre-adapted cells were analyzed by quantitative proteomic methods and differentially expressed proteins were categorized using the DAVID Bioinformatics Resource.
Cells adapted to 25°C expressed changes in the abundance of 219 unique proteins representing a broad range of categories such as translation, glycolysis, biosynthetic (anabolic) processes, NAD, cytoskeletal organization, RNA processing, oxidoreductase activity, response-to-stress and cell redox homeostasis. The number of proteins that decreased significantly with cold-adaptation exceeded the number that increased by 2:1. Almost half of the decreases were associated with protein metabolic processes and a third were related to anabolic processes including protein, DNA and fatty acid synthesis. Changes consistent with the suppression of cytoskeletal dynamics provided further evidence that cold-adapted cells are in an energy conserving state. Among the specific changes were increases in the abundance and activity of redox proteins glutathione S-transferase, thioredoxin and thioredoxin reductase, which correlated with a decrease in oxidative stress, an increase in protein glutathionylation, and a recovery of reduced protein thiols during rewarming from 0°C. Increases in S-adenosylhomocysteine hydrolase and nicotinamide phosphoribosyltransferase implicate a central role for the methionine-cysteine transulfuration pathway in increasing glutathione levels and the NAD salvage pathway in increasing the reducing capacity of cold-adapted cells.
Endothelial adaptation to mild-moderate hypothermia down-regulates anabolic processes and increases the reducing capacity of cells to enhance their resistance to oxidation and injury associated with 0°C storage and rewarming. Inducing these characteristics in a clinical setting could potentially limit the damaging effects of energy insufficiency due to ischemia and prevent the disruption of integrated metabolism at low temperatures.
Hypothermia is utilized in several clinical situations. Moderate-to-deep hypothermia (32°-15°C) is the primary method for delaying ischemic injury in patients during cardiovascular [1, 2], neurovascular  or trauma surgery . More severe hypothermia (< 10°C) is used to protect isolated tissues and organs for transplantation. Therapeutic mild hypothermia (34°-32°C) improves the neurological recovery of patients following a cardiac arrest [5, 6], stroke , or traumatic brain injury . In each situation, the reduction in temperature decreases the metabolic demand for energy and slows the progression of injury. However, hypothermia also disrupts metabolic integration and impairs important functional processes, particularly when it is severe. For example, cold-induced vascular injury is characterized by a loss of endothelial cell-cell contact [9, 10], a release of inflammatory cytokines [11, 12], an increase in the expression of adhesion molecules [11, 12], impaired vasoactivity [13, 14] and cell death . Unfortunately, both the protective and harmful mechanisms induced by hypothermia vary with temperature and are poorly understood and therefore its clinical utility remains limited.
Many organisms, from prokaryotes to plants and animals adapt to a range of temperatures and remain viable. Poikilotherms and mammalian hibernators adapt to cold and become resistant to injury from more prolonged and severe hypothermia and ischemia. Interestingly, the cells of non-hibernating mammals adapt to mild-moderate hypothermia in tissue culture [16–19] in what is likely a conserved response to cold and several mechanisms of cold-induced changes in gene expression have been identified. These include a generalized inhibition of transcription and translation, an increase in transcription of some RNA-binding chaperones, alternative splicing of pre-mRNAs and preferential translation of mRNAs that have cold-inducible internal ribosome entry sites (IRESs) . In our studies of cold-adaptation, human coronary artery endothelial cells (HCAECs) cultured at 25°C become progressively more resistant over time to 0°C-injury and in particular to the oxidative stress induced by exposure to 0°C and rewarming . The molecular basis of the adaptation remains largely unknown but the resulting protection at 0°C is due, in part, to the sequestration of catalytically active iron . The protection may also be associated with an increase in intracellular glutathione, an important antioxidant and signaling molecule of the cell, at 25°C. Glutathione (GSH) reacts directly with free radicals, participates in the reductive detoxification of hydrogen peroxide and organic peroxides, serves as a co-factor in the enzymatic breakdown of xenobiotics and reacts with protein thiols to form mixed disulfides (P-SSG) under conditions of mild oxidative stress . Protein glutathionylation is a reversible modification that provides a mechanism for protecting proteins from irreversible oxidative damage and for regulating protein function and thereby many diverse cellular processes such as cytoskeletal organization, ionic homeostasis and the expression of genes involved in antioxidant defenses . Understanding the process of cold-adaptation and the characteristics of the protected state in human cells may be beneficial to the development of new clinical strategies for improved ischemic and hypothermic protection.
The aim of the present study was to characterize the cold-adapted state and the potential role of GSH in endothelial protection at 0°C. We performed quantitative proteomic analyses of human coronary artery endothelial cells collected before and after 72 h of culture at 25°C and classified the changes in protein abundance using the DAVID Bioinformatics Resource [22, 23]. The most significant biological processes associated with cold-adaptation were translation, glycolysis, RNA processing, actin filament-based processes, and mRNA metabolic processes. Evidence of a regulated decrease in energy-consuming processes such as protein, DNA and fatty acid synthesis and suppression of cytoskeletal dynamics suggest that the cold-adapted state is hypometabolic. There was also evidence of a complementary increase in antioxidant protection as there were significant changes in categories relevant to GSH such as oxidoreductase activity and cell redox homeostasis. Among the many changes in cold-adapted endothelial cells were increases in the abundance of thioredoxin, thioredoxin reductase and glutathione S-transferase, which correlated with increases in their respective activities, increases in protein glutathionylation, recovery of reduced protein thiols and diminished oxidative stress following rewarming from 0°C.
Changes in protein expression following cold-adaptation as determined by LC/MS-based label-free protein quantitative analysis.
Peptide ID confidence
Number of proteins
Number significant changes
Maximum absolute fold change
Number increasing at 25°C
Number decreasing at 25°C
Selected GOa biological process classifications of proteins differentially expressed in HCAECs following cold-adaptation.
# of Proteins
Gene ontology categories
Primary metabolic process
1.4 × 10-3
Macromolecule metabolic process
1.9 × 10-3
3.0 × 10-6
1.7 × 10-3
9.9 × 10-2
5.1 × 10-12
Protein metabolic process
3.9 × 10-5
Macromolecular complex assembly
2.2 × 10-3
Ribonucleoprotein complex assembly
5.3 × 10-5
1.1 × 10-9
mRNA metabolic process
3.6 × 10-7
2.8 × 10-28
1.1 × 10-2
1.8 × 10-2
Unfolded protein bindingf
5.0 × 10-5
3.3 × 10-2
3.5 × 10-2
Structure-specific DNA binding
2.5 × 10-2
1.5 × 10-5
Actin filament-based process
2.5 × 10-8
3.3 × 10-2
5.1 × 10-2
Cell redox homeostasis
6.4 × 10-3
8.9 × 10-2
Two parameters used in DAVID, P-value and fold-enrichment, quantify the extent to which a given category is overrepresented within the dataset when compared to the precentage of similarly categorized genes from the entire genome. Categories with smaller P-values (< 0.05) and larger fold-enrichments (≥ 1.5) are considered to be of greater interest . The most enriched groups from Table 2 are translation, glycolysis, RNA processing, actin filament-based processes and mRNA metabolic processes (P-values of 2.8 × 10-28, 5.1 × 10-12, 1.1 × 10-9, 2.5 × 10-8 and 3.6 × 10-7, and fold-enrichment of 10.2, 21.6, 4.2, 6.0, and 4.3, respectively; see Table 2). Categories that may be relevant to GSH activity, such as oxidoreductase activity, response to stress and cell redox homeostasis appeared to be less enriched and of lower significance (fold-enrichment of 1.8, 1.4, and 6.7 and P-values of 5.1 × 10-2, 8.9 × 10-2 and 6.4 × 10-3, respectively) and therefore potentially of reduced importance. These categories show mixed changes in protein abundance, with increase-to-decrease ratios of 1.8-to-1, 1-to-1.5, and 1.5-to-1.
The molecule nicotinamide adenine dinucleotide or NAD plays a central role in linking the energy producing catabolic reactions of the cell with the maintenance of redox homeostasis . Significant changes in NAD metabolic proteins were identified in this study. Cells adapted to 25°C had an increase in the level of nicotinamide phosphoribosyltransferase (Nampt), the rate-limiting enzyme in NAD biosynthesis from nicotinamide  (see Additional file 1). There was also an increase in glycolytic enzymes that use NAD as an electron acceptor or cofactor. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which reduces NAD+ to NADH, increased at 25°C, as did lactate dehydrogenase (LDHA, LDHAL6A and LDHAL6B), which reduces pyruvate to lactate and regenerates NAD+(Figure 2). Interestingly, cold-adaptation increased the abundance of S-adenosylhomocysteine hydrolase (SAHH), an NAD+ dependent enzyme and component of the methionine-cysteine transulfuration pathway that generates cysteine used to synthesize GSH . Inosine monophosphate dehydrogenase (IMPDH2), the NAD+-dependent and rate-limiting enzyme of de novo guanine nucleotide synthesis , decreased at 25°C.
The phosphorylated and reduced form of NAD, NADPH, provides the reducing equivalents for GSH and thioredoxin reduction that bolsters the cell's resistance to oxidative stress. Several NADP(H)-dependent enzymes changed significantly in abundance during cold-adaptation. Aldehyde dehydrogenase (ALDH1A1), an NADP+-dependent enzyme, increased at 25°C (Figure 2). Its function in the cell is to oxidize toxic aldehydes produced during lipid peroxidation and thereby prevent their adduct formation with proteins or DNA . Conversely, several proteins decreased in abundance, including ribonucleotide reductase (RRM1), a thioredoxin-dependent enzyme  that catalyzes the formation of deoxyribonucleotide from ribonucleotide ; and fatty acid synthase (FAS), which catalyzes the formation of long-chain fatty acids from acetyl CoA, malonyl CoA and NADPH.
Thirteen out of 17 proteins associated with actin filament (F-actin)-based processes were up-regulated in cold-adapted cells (Figure 7). Although there were no significant changes in the abundance of cytosolic actin at 25°C (see Additional file 1), there were significant decreases in actin nucleation and cross-linking proteins (ARPC3  and FLNA ), increases in actin filament severing proteins (CFL1 and DSTN ) and in proteins that inhibit polymerization by binding to monomeric G-actin (CAP1 , TMSB10  and TMSL3). There was also a decrease in septin 2, a GTP-binding and actin-organizing protein . Cold-adaptation increased several other proteins that associate with actin, including the membrane-associated small GTPase RHOQ; Rho GDP dissociation inhibitor β (ARHGDIB), a chaperone protein that inactivates Rho GTPases ; CDC42, a Rho GTPase that regulates actin cytoskeletal remodeling; the actin-bundling protein fascin (FSCN1); and myosin light chain subunits (MYL6, MYL6B, MYL12A and MYL12B), which regulate actinomyosin contractility.
Other changes included a cold-induced increase in the type III intermediate filament (IF) proteins vimentin (VIM) and peripherin (PRPH) that form a third class of cytoskeletal network (Figure 7). There were changes in the abundance of proteins (CASL, TLN1, PLEC1, and FERMT3) that link the cytoskeleton to integrin, which is a membrane receptor that binds the cell to proteins of the extracellular matrix, and there was a significant decrease in β-catenin (CTNNB1), a cytoplasmic component of the adherens cell-cell junction.
Differential proteomic analyses of endothelial cells collected before or after cold-adaptation at 25°C demonstrated significant changes in the abundance of proteins that regulate redox homeostasis or have oxidoreductase activity. The increase in thioredoxin and thioredoxin reductase correlated with the increase in their respective activities and with the recovery of reduced protein thiols following rewarming from 0°C, conditions that generate oxidative stress and endothelial cell damage [16, 50, 51]. Protein disulfide isomerases catalyze thiol oxidation; therefore, the decrease in PDIA1 and PDIA4 abundance in cold-adapted cells may have contributed to the recovery of PSH with rewarming. Additional adaptations are potentially beneficial to the maintenance of redox homeostasis. For example, preliminary studies showed that there is an increase in tyrosine hydroxylase, the rate-limiting enzyme in catecholamine synthesis, at 25°C . Catecholamines such as dopamine attenuate hypothermia-induced reactive oxygen species and the loss of -SH reducing equivalents at 4°C in human umbilical vein endothelial cells .
Endothelial cold-adaptation at 25°C is accompanied by a significant increase in intracellular GSH . The two enzymes required for GSH synthesis, gamma glutamylcysteine synthetase and glutathione synthetase, were not identified in our proteomic study. However, there was a significant increase in S-adenosylhomocysteine hydrolase (SAHH; Figure 2), a component of the methionine-cysteine transsulfuration pathway that generates cysteine for GSH synthesis. This pathway is up-regulated in response to oxidative stress  and accounts for up to 50% of GSH production in some tissues . We did not observe an increase in GSSG/tGSH, which is indicative of oxidative stress, with cold-adaptation, but this does not rule out the possibility of localized changes in GSSG/tGSH that could have had a stimulatory effect on tGSH synthesis at 25°C. The increase in protein glutathionylation following rewarming from 0°C matched the higher levels of tGSH in cold-adapted cells  and the increase in abundance and activity of glutathione S-transferase. GST conjugates GSH to proteins under conditions of oxidative stress . Protein glutathionylation is a reversible process that precludes irreversible disulfide cross-linking ; therefore, glutathionylation is potentially a protective mechanism in cold-adapted cells.
The increase in GSH/GST and Trx/TrxR activities requires the availability of reducing equivalents from NADPH. Cold-adapted cells had a significant increase in Nampt, the rate-limiting enzyme in NAD synthesis from nicotinamide . NAD kinase, which phosphorylates NAD+(H) to NADP+(H), was not identified in our study. The enzymes of the pentose phosphate pathway (PPP) that reduce NADP+ to NADPH, glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase, were unchanged in abundance. However, the decrease in phosphofructokinase (PFK), the rate-limiting enzyme in the glycolytic pathway, potentially redirects glucose catabolism from glycolysis to the alternative PPP in a manner similarly induced by down-regulating the activity of other downstream glycolytic enzymes such as GAPDH or triosephosphate isomerase (TPI) . The proteomic data indicate that there may be two additional mechanisms in cold-adapted cells to bolster NADPH availability. Aldehyde dehydrogenase (ALDH1A1- Figure 2), which oxidizes the lipid peroxidation products 4-hydroxynonenal and malondialdehyde using NADP+ as a cofactor , increased at 25°C. Lipid peroxidation is an important mechanism of injury during 0°C storage and rewarming [50, 51]. An increase in ALDH1A1 would therefore potentially protect cell membranes from oxidative damage and simultaneously regenerate NADPH for GSSG and Trx reduction. A second approach to ensuring NADPH availability relates to the apparent decrease in NADPH-dependent biosyntheses at 25°C. Ribonucleotide reductase (RRM1), a Trx-dependent enzyme  that catalyzes deoxyribonucleotide formation and controls the rate of DNA synthesis at 37°C , decreased significantly at 25°C (Figure 2). Inosine monophosphate dehydrogenase (IMPDH2), the rate limiting enzyme in de novo guanine nucleotide synthesis, also decreased at 25°C. Inhibition of IMPDH2 decreases guanine nucleotide pools and interrupts both DNA and RNA synthesis . Down-regulating these two enzymes may therefore suppress DNA synthesis independently of the Q10 effects  and increase the availability of NADPH for GSSG and Trx reduction. A significant decrease in the NADPH-dependent fatty acid synthase (Figure 2), and presumably fatty acid synthesis, at 25°C may have a similar NADPH-sparing effect and thereby contribute to the protective phenotype of cold-adapted cells.
Cold-adapted cells had a significant increase in lactate dehydrogenase (LDHA, LDHAL6A and LDHAL6B), which reduces pyruvate to lactate. This step diverts pyruvate from further oxidation by mitochondrial respiration and regenerates NAD+ to facilitate more glycolysis. The reduction of pyruvate despite the availability of oxygen is known as the Warburg effect or aerobic glycolysis and has been shown to decrease mitochondrial respiration, mitochondria-derived ROS and apoptosis [66, 67]. In summary, the observed changes in the abundance of glycolytic proteins may be a cold-induced adaptation that promotes NADPH reduction while maintaining ATP generation and minimizing mitochondria-generated oxidative stress.
Cold-adaptation induced significant changes in the abundance of proteins that regulate gene expression. Histones and histone modifications, such as methylation or acetylation, play an important role in regulating transcription by modifying chromatin structure . The histone H2A.V increased significantly at 25°C. In Arabidopsis and in budding yeast cells, incorporation of the closely related variant H2A.Z  into nucleosomes is a requirement for direct temperature-induced changes in gene expression . Whether or not H2A.V plays a similar role in promoting the expression of cold-responsive endothelial genes is unknown.
Histone methylation mediates short-term chromatin condensation and transcriptional repression (DNA methylation affects the long-term silencing of genes) . In addition to its role in generating cysteine for GSH synthesis, S-adenosylhomocysteine hydrolase (SAHH) is critical to histone/DNA methylation and gene silencing because it eliminates S-adenosylhomocysteine (SAH), a by-product and potent inhibitor of S-adenosylmethionine (SAM)-dependent methyltransferase reactions  (SAM is the methyl donor). The increase in SAHH observed at 25°C may therefore promote histone/DNA methylation and potentially gene silencing . There is also experimental evidence that increases in Nampt, which we observed at 25°C, promotes histone deacetylation and consequently gene silencing . The level of NAD and Nampt in mouse fibroblasts was shown to regulate the activity of Sir2, a protein deacetylase . Similarly, the overexpression of Nampt in smooth muscle cells enhanced cell resistance to oxidative stress and increased the activity of SIRT1, the human homolog of Sir2 . SIRT1 regulates nucleosome and chromatin structure by deacetylating histones H3 and H4 and mediates stress resistance, apoptosis and inflammatory responses by deacetylating the transcription factors FOXO, p53 and NF-κB . The increase in the expression of SAHH and Nampt at 25°C, and their capacities to modulate both transcriptional repression and resistance to oxidative stress, suggests that these proteins may have pivotal roles in the adaptation to cold.
Additional findings point to an overall decrease in transcription and translation at 25°C. There were significant decreases in the abundance of proteins associated with transcription initiation, mRNA elongation, processing and nonsense-mediated decay, and stress granule assembly. The inhibition of IMP dehydrogenase has been shown to disrupt DNA and RNA synthesis  and its decrease at 25°C potentially limits global transcription and translation. The increase in transcriptional repressors ENO1 and HNRNPDL at 25°C complement earlier findings that showed increases in the repressors prohibitin, L3MBTL2 and Zinc finger protein 224 . Although ENO1 (α-enolase) is recognized as a glycolytic protein, an alternate translation product of the ENO1 mRNA is a repressor of the c-myc gene . The myc family of proteins are transcription factors that regulate ribosome biogenesis, protein synthesis, cell growth and proliferation . Although we did not identify c-myc, N-myc was identified and it had the largest fold-change decrease (-2.38) of any protein detected in this study (Additional file 1; N-myc was excluded from the analysis because of its classification as a priority 4 protein [24, 26]). However, the decrease in N-myc abundance and increase in ENO1 match the apparent decrease in ribonucleoprotein complex biogenesis (11 out of 11 proteins decreased), translation (38 out of 40 decreased) and unfolded protein binding (9 out of 9 decreased). By contrast, cold-adaptation increased the abundance of RBM3, a cold-inducible RNA-binding protein that increases the efficiency of global protein synthesis at 37°C and at mild hypothermic temperatures . It has been suggested that RBM3 may prevent a disproportionate cold-induced decrease in translation . The fact that one third of the changes in protein abundance at 25°C were increases suggests that RBM3 may have an important role in facilitating these increases.
The response to stress proteins (Figure 6) that increased or decreased significantly with cold-adaptation are associated with processes like chromatin binding, transcription, DNA repair, pre-mRNA processing, regulation of translation, protein folding and cell adhesion. Five of the identified proteins are found in the ER, which is the site where secretory and cell surface proteins are synthesized, folded and modified . The decrease in the abundance of these ER-resident proteins after 72 h at 25°C is additional evidence that protein synthesis and protein folding is repressed in the cold-adapted state. Excess unfolded or misfolded proteins in the ER activates the Unfolded Protein Response or UPR. The UPR initiates translational repression, increases the transcription of ER chaperone genes and increases the degradation of misfolded proteins to re-establish the equilibrium between protein synthesis and protein folding reactions . It is not known if ER function is impaired at 25°C or if translational repression due to hypothermia is initiated by the UPR, but evidence of UPR activation in the tissues of hibernating mammals has been reported . Activation of the UPR is also a potential mechanism for increasing GSH in cold-adapted cells  because of the important role that GSH plays in regulating redox homeostasis during the oxidative folding of proteins . ER stress-activated transcription factors ATF4 and Nrf2 increase GSH synthesis [81, 82] and increases in endothelial GSH have been linked to hyperoxic- or cycloheximide-induced suppression of translation .
YB-1 and CSDA are two nucleic acid-binding stress proteins that contain a central cold-shock domain, a 100-amino acid sequence that derives its name from its resemblance to bacterial cold-shock proteins . In homeothermic organisms, cold-shock domain (CSD) proteins have been associated with the transcriptional regulation of stress-response, growth factor, and proliferation-associated genes . In bacteria, cold-shock proteins function as mRNA chaperones that facilitate translation . Their expression levels increase during the acclimation to cold and then later decline below pre-cold-shock levels . It is this archaic function of CSD proteins that may have special relevance to the endothelial adaptation to cold.
Cold-adapted cells had significantly higher levels of 14-3-3 proteins zeta/delta (Figure 6, 7) and theta (Figure 5). 14-3-3's are a family of conserved proteins that control many diverse processes by binding to specific phosphorylated sites on target proteins to induce changes in the target conformation or to change the target's interactions with other proteins . In plants, 14-3-3's help regulate processes such as photosynthesis, ATP production and nitrate reduction in response to changing environmental factors including cold stress. In mammalian cells, 14-3-3's regulate responses to growth factors, including the stimulation of glycolysis, inhibition of apoptosis and inhibition of transcription .
The adaptation of endothelial cells to cold induced significant changes in the abundance of cytoskeletal proteins that implicate diminished microtubular and actin networks, an expanded intermediate filament network and modified cell-cell and cell-extracellular matrix interactions. The actin cytoskeleton underlies the plasma membrane to provide support, resist tension and maintain cell shape. However, it is a highly dynamic structure that responds to environmental stimuli and rapidly reorganizes to change cell shape, form membrane protrusions, (such as lamellipodia and filopodia for cellular locomotion), or form stress fibers (bundles of radially oriented actin-filaments cross-linked with myosin motor proteins) for retracting the plasma membrane. The proteomic findings suggest that there is a significant reorganization of the actin cytoskeleton in response to the reduced temperature. Decreases in actin nucleation and cross-linking proteins (ARPC3  and FLNA ), increases in actin filament severing proteins (CFL1 and DSTN ) and in proteins that bind to G-actin to inhibit polymerization (CAP1  and TMSB10 ), allude to a net disassembly of the actin cytoskeleton at 25°C. Disassembly is further indicated by the increase in RhoGDI2, which promotes actin depolymerization , and by the decrease in septin 2, whose depletion leads to dissociation of actin stress-fibers .
Microtubules consist of α- and β-tubulin heterodimers and form robust structures that are resistant to bending or compression . Microtubules organize the cytoplasm, position the nucleus and organelles, transport organelles and make up the mitotic spindle . The decrease in α-tubulins, microtubule associated proteins and chaperonins at 25°C indicate a decrease in the microtubule network (11 out of 12 microtubule-associated proteins decreased). Our preliminary studies showed a matching increase in the microtubule stabilizing protein LIS1 at 25°C , which points to a decrease in microtubule catastrophe events and diminished microtubule dynamics . One consequence of these changes may be a reduction in GTP utilization because microtubule assembly/disassembly requires the hydrolysis of tubulin-bound GTP to GDP . Such a reduction would match the apparent decrease in translation, which is another GTP-utilizing process, and the decrease in IMPDH2 (Figure 2), which is the rate-limiting enzyme in de novo GMP synthesis.
In contrast to diminished microtubular and actin-filament networks, there was a cold-induced increase in type III intermediate filaments (IFs; Figure 7) that form a third class of cytoskeletal network. Cells at 25°C developed a significant increase in vimentin, a major IF that provides mechanical support for the cell and takes part in additional processes such as the regulation of cell adhesion to the extracellular matrix. A second IF, peripherin, also increased. IFs are highly dynamic structures that self-assemble without expenditure of GTP or ATP required by tubulin or actin polymerization, respectively. It has been theorized that since actin dynamics may account for up to 50% of ATP turnover in resting cells, a change in cytoskeletal structure to a less dynamic form may reflect a decrease in the energy state of the cell or possibly represent a form of ATP preservation for more crucial activities . The substitution of an expanded IF network for diminished actin filament and microtubule networks may therefore be a possible sign that cold-adapted cells are in an energy-reduced or energy-conserved state.
Cytoskeletal changes such as those noted above are frequently mediated through changes in cell-cell and cell-matrix adhesion and our study indicates that similar processes are occurring with cold-adaptation. Cdc42, a Rho GTPase that regulates actin cytoskeletal remodeling in response to environmental signals, increased significantly at 25°C. Activation of Cdc42 preserves the integrity of endothelial cell-cell junctions , prevents membrane retraction  and induces filopodia formation . These properties may contribute to the improved ability of cold-adapted endothelial cells to reverse intercellular gap formation during rewarming from 0°C . The increase in Cdc42 in our study coincided with significant increases in fascin, a protein that organizes actin filaments into the parallel bundles that are incorporated into filopodial protrusions ; in myosin light chain subunits (MYL6, MYL6B, MYL12A and MYL12B) that regulate actinomyosin contractility and filopodia formation ; and in fibronectin 1 and thrombospondin 1 (Figure 6), which are secreted components of the extracellular matrix. The association of fascin with actin is regulated by the binding of integrin, a transmembrane receptor, to the extracellular matrix and to thrombospondin and fibronectin in particular . In addition, there were changes in the abundance of proteins that link integrin to the cytoskeleton. For example, CasL, which mediates signals between integrin and possibly vimentin , and talin 1, which links integrin to actin , increased significantly. By contrast, plectin, a linker protein between IFs and microtubules, microfilaments or integrins , and kindlin-3, which is a co-activator of integrin , decreased at 25°C. Cold-adaptation also caused changes in components of cell-cell complexes. There were significant decreases in β-catenin, which links the adherens cell-cell junction to F-actin . Our previous findings showed decreases in the membrane receptor EphA3, which is a member of the Eph receptor family of receptor tyrosine kinases that induce changes in cell shape and movement upon the binding of ephrin ligands from an adjacent cell . Overall, these findings suggest that cytoskeletal remodeling during cold-adaptation may be stimulated by increases in fibronectin and thrombospondin secretion, integrin binding and cdc42 activation, and by decreases in adherens junction and ephrin/ephrin receptor cell-cell signaling.
Organisms that successfully adapt to cold change the lipid composition of their cell membranes to maintain membrane fluidity and function with decreasing temperature. Some cold-induced membrane changes observed in fish are an increase in the proportion of unsaturated to saturated fatty acids, an increase in fatty acid chain length and an increase in the proportion of phosphatidylethanolamine to phosphatidylcholine . These changes are brought about by adjustments in lipid metabolism, including de novo synthesis, elongation and desaturation of fatty acids; de novo phospholipid synthesis and desaturation; and phospholipid catabolism and turnover in the membrane . In our study, few lipid metabolic proteins were identified so there is little information about potential modifications to endothelial cell membranes in response to cooling. Fatty acid synthase was significantly down regulated in cold-adapted cells (Figure 2), which suggests that de novo fatty acid synthesis is not essential. Interestingly, there was a cold-induced increase in phospholipase A2 receptor 1 (PLA2R1 was a priority 4 protein; see Additional file 1), which binds the lipolytic enzyme phospholipase A2 (PLA2). PLA2s hydrolyze phospholipids at the sn-2 position, releasing free fatty acids and lysophospholipids from the membrane. The sn-2 position is frequently occupied by polyunsaturated fatty acids (PUFA), so PLA2 activity potentially regulates membrane unsaturation and fluidity. In support of this hypothesis, Zhang et al. demonstrated that inhibition of Ca2+-independent PLA2 activity increased PUFA in membrane phosphatidylcholines . PLA2 activity at hypothermic temperatures is well-known [104–106]. It is not known, however, if PLA2R1 has a stimulatory or inhibitory effect on endothelial PLA2s.
In addition to a direct effect of cooling on membrane fluidity, there may be an indirect effect that is caused by cold-induced oxidative stress and peroxidation of unsaturated lipids. The lipid peroxidation products 4-hydroxynonenal and malondialdehyde have been shown to increase the rigidity of cell membranes [107, 108] and have been measured following cold exposure and rewarming [50, 51]. The increase in peroxiredoxin, aldehyde dehydrogenase and glutathione S-transferase in cold-adapted cells potentially attenuate lipid peroxidation in endothelial membranes and may help preserve membrane fluidity during cooling.
This study was performed using a well-validated label-free mass spectrometry-based method [24, 26, 109, 110] that is rapid, sensitive and has a three- to four-fold higher protein dynamic range (protein mass and abundance) than two-dimensional gel electrophoresis (2-DE) . Almost one third of the 1089 proteins that were identified significantly increased or decreased in abundance during cold-adaptation and 181 proteins of high identification confidence were included in the functional analysis. This number greatly exceeds the number of proteins identified and analyzed in previous 2-DE-based hypothermia studies [112, 113]. Ten of 26 proteins identified by Baik et al.  changed significantly in our study, including the glycolytic proteins pyruvate kinase M2, phosphoglycerate kinase and glyceraldehyde-3-phosphate dehydrogenase; cytoskeletal proteins α-tubulin, annexin A1 and vimentin; chaperones HSP90β and HSPA8; and the ER-resident protein disulfide isomerase. With the exception of PDI and HSPA8, the changes in abundance were in the same direction as in our study. Kumar et al.  observed similar cold-induced increases in the abundance of vimentin, aldehyde dehydrogenase 1A1 and GAPDH as were reported in this study. They also reported decreases in tubulins (β-tubulin rather than α-tubulin) and annexins (A4 rather than A1) . In addition, Kumar et al.  identified a decrease in α-enolase (we noted an increase) and an increase in the cytoskeletal chaperone CCT2 (we reported decreases in CCT4 and 5).
Central to the proteomic changes is the evidence of a regulated decrease in energy consuming processes such as protein synthesis, DNA synthesis, fatty acid synthesis and cytoskeletal dynamics. The net decrease in proteins associated with translation, ribonucleoprotein complex biogenesis, RNA processing, mRNA metabolism and protein folding at 25°C is consistent with a generalized cold-induced inhibition of protein synthesis described by others . Increases in transcriptional repressors and changes in the abundance of histone and non-histone chromatin-binding proteins are possible indicators of a concurrent decrease in transcription. Changes in the expression of cold-shock proteins (RBM3 increased, CSDA and YB-1 decreased) and the increase in histone H2A.V at 25°C potentially counteract the generalized decrease in gene expression by enhancing translation efficiency and the direct responsiveness of genes to cold, respectively. This may be an important requirement for increasing the synthesis of a subset of proteins at reduced temperatures- one third of identified proteins increased in abundance in our study. Two enzymes we have identified, SAHH and Nampt, may play pivotal roles in mediating adaptation to cold because each induces transcriptional repression and increases the cell's reducing capacity. SAHH facilitates SAM-dependent DNA and histone methylation and consequently gene silencing  while up-regulating GSH synthesis. Nampt regulates histone deacetylation and transcriptional repression  while increasing NAD synthesis. The ability to down-regulate cellular energy demand while increasing the cell's reducing capacity would be clinically beneficial where the application of cold-ischemia would otherwise create energy insufficiency, uncouple metabolic integration and generate oxidative stress and injury. Whether or not a similar adaptive response occurs in other cell types, such as myocardial or neural cells, or following an ischemic insult, such as during the application of therapeutic hypothermia, remain important questions.
Proliferating human coronary artery endothelial cells (Clonetics; passages 4 to 8) were seeded at a density of 1.4 × 104 viable cells/cm2 (6-well culture plates, 12.5 cm2, 25 cm2 or 75 cm2 culture flasks; Falcon, Franklin Lakes, New Jersey) and were maintained at 37°C (95% air + 5% CO2) in culture medium (EGM2-MV; Clonetics). Cells formed a confluent monolayer after 3 to 4 days of culture and were then used in experiments. Cells were adapted to cold at 25°C for up to 72 h in a water-jacketed incubator (3.8% CO2 + 96.2% air) and the culture medium was replaced with fresh medium every 24 h. Controls consisted of confluent cells collected at the start of the adaptation period or following continued maintenance at 37°C for up to 72 h. For cold storage, cells were washed with Hank's balanced salt solution, flasks received endothelial basal medium (EBM-2, Clonetics) and were immersed in a circulating water bath at 0°C.
A large scale quantitative proteomics experiment was performed using a well-validated label-free mass spectrometry-based method [24, 26, 109, 110]. Cold-adapted and control cells were prepared as previously described . Proteins were extracted from the tissue culture cells using 8 M urea, reduced by triethylphosphine, alkylated by iodoethanol, and digested by trypsin . Tryptic peptides (~20 μg/injection) were analyzed using Thermo-Fisher linear ion-trap mass spectrometer (LTQ) coupled with a Surveyor HPLC system. C-18 reverse phase column (i.d. = 2.1 mm, length = 50 mm) was used to separate peptides with a flow rate of 200 μL/min. Peptides were eluted with a gradient from 5 to 45% acetonitrile developed over 120 min and data were collected in the triple-play mode (MS scan, zoom scan, and MS/MS scan). The acquired data were filtered and analyzed by a proprietary algorithm that was developed and described by Higgs, et al. [26, 109]. Database searches against the IPI (International Protein Index) human database was carried out using both the X!Tandem and SEQUEST algorithms. Protein quantification was also carried out using the same proprietary algorithm described previously [24, 26, 110].
IPIs of Priority 1 and 2 proteins were converted to UniProt Accession numbers using the EMBL-EBI PICR Service, http://www.ebi.ac.uk/Tools/picr/. Accession numbers were then uploaded to the DAVID (Database for Annotation, Visualization and Integrated Discovery) Bioinformatics Resources 6.7 website http://david.abcc.ncifcrf.gov/tools.jsp and the proteins were classified and organized according to their GO Biological Process, Molecular Function, Cell Component and SP-PIR Keyword annotations using the Gene Functional Annotation Tool.
Confluent HCAEC from T-75 flasks were rinsed twice with HBSS and scraped in 1.5 ml of 5% TCA in 10 mM HCl with 5 mM EDTA. Protein pellets and non-protein supernatants were separated by centrifugation and stored at -80°C until analysis of protein thiols (PSH) and mixed disulfides (PSSG) using the method described by Nagy  with some modifications. The protein pellets were extracted with diethyl ether to eliminate TCA, redissolved in 0.25 ml 6 M GuHCl, and pH adjusted to 8 by addition of 0.5 ml phosphate buffer with 5 mM EDTA, pH 8.6. Insoluble material was removed by centrifugation and supernatants were used to determine the levels of PSH and PSSG. For PSH, 100 μl of the sample was reacted with 0.9 ml of 0.1 mM 5,5'-dithio-bis-(2-nitrobenzoic acid) (DTNB, 1 mM stock in methanol) in 125 mM sodium phosphate buffer containing 6.25 mM EDTA, pH 8.6. Absorbance was measured at 412 nm after 30 minutes at room temperature. GSH standards (50-500 μM) were similarly measured and PSH-levels in samples were calculated directly from the standard curve.
PSSG levels were determined in the reaction mix comprised of 0.1 ml sample, 0.9 ml of 125 mM sodium phosphate buffer, pH 7.5 (with 6.25 mM EDTA) containing 0.1 mM DTNB, 0.03 mM NADPH and 20 μl of glutathione reductase (25 units per ml). The absorbance was measured at 412 nm after 10 minutes of incubation at room temperature. PSSG levels in samples were calculated from a GSH standard curve (0.5 - 10 μM) run under same conditions.
Confluent cells were washed two times in HBSS and then lysed in 500 μL of 125 mM sodium phosphate buffer with 6.3 mM EDTA and 2% Triton X-100. Protein was removed from lysates by precipitation with 500 μL of metaphosphoric acid (10% w/v) and centrifugation at 14,000 × g for 5 minutes. The supernatant was neutralized with 5 N sodium hydroxide and stored at -80°C until t-GSH and GSSG was measured using the method of Ishii . To determine t-GSH, 100 μL of the neutralized supernatant was incubated at 25°C with 800 μL of 0.3 mM NADPH in 125 mM sodium phosphate buffer with 6.3 mM EDTA, 100 μL of 6 mM 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) and 20 μL of 25 U/mL GSH reductase. After precisely 15 minutes, the optical density was measured at 412 nm. To determine GSSG, GSH in samples was derivatized by adding 2.5 μL of 4-vinylpyridine (95%) per 250 μL solution, mixing vigorously for 1 min and incubating samples at room temperature for 30 min. Samples were then centrifuged at 14,000 × g for 5 min and GSSG was measured in supernatants by the procedure described for t-GSH. The measured optical densities were converted to concentrations using standard curves prepared with GSH (0.5-10 μM) or GSSG (0.25-5 μM).
Confluent HCAECs from T-75 flasks were scraped and resuspended in 0.15-0.5 ml of assay buffer and cell lysates were prepared by sonicating samples two times on ice (10 seconds, amplitude of 40) using an Ultrasonic Processor. Insoluble materials were removed by centrifugation at 13,000 rpm for 20 minutes at 4°C and supernatants were used to measure enzyme activities and total protein concentrations (Micro Bicinchoninic Assay Kit; Pierce Thermoscientific). Thioredoxin (Trx) activity in samples was determined using an assay kit (Redoxica, Littlerock, AK) and following the manufacturer's recommendations. Briefly, Trx activity was determined by measuring the rate of depletion of NADPH at 340 nm in the presence of oxidized insulin and thioredoxin reductase (TrxR). Reaction blanks with each sample were run without the insulin substrate to determine Trx-independent NADPH depletion. Positive controls consisting of thioredoxin supplied with the kit were run to validate the assay. Reaction rates were calculated from the slopes of logarithmic plots of A340 nm versus time (in min). The difference in reaction rates between the samples and corresponding reaction blanks represented Trx activity and was expressed as nmoles of Trx activity/min/mg protein.
TrxR activity was measured using an assay kit (Cayman Chemical Company, Ann Arbor, Michigan, USA). The assay is based on measuring the rate of reduction of DTNB (5,5'-dithiobis(2 dinitrobenzoic acid)) to a colored product by TrxR, in the presence of NADPH. TrxR-independent reduction of DTNB was determined by adding aurothiomalate (ATM), a specific inhibitor of TrxR, to the reaction mixture. TrxR activity was determined by measuring the difference in the rate of increase in absorbance at 412 nm between samples with or without ATM and was expressed in nmoles/min/mg proteins, using an extinction coefficient of 6.35 mM-1.
GST activities were assessed using an assay kit from Cayman and following the manufacturer's protocol. GST activity was quantified by measuring the rate of conjugation of reduced glutathione by GST to CDNB (1-chloro-2,4 dinitrobenzene) at 340 nm and was expressed in nmoles of conjugated CDNB/min/mg protein, using an extinction coefficient of 0.0053 μM-1.
International Protein Index
Universal Protein Resource
European Molecular Biology Laboratory
European Bioinformatics Institute
Protein Identifier Cross Reference
Database for Annotation, Visualization and Integrated Discovery
Swiss Prot- Protein Information Resource
We thank Heather Richardson for her help preparing the figures. We gratefully acknowledge the support of the Methodist Research Institute and IU Health in funding this study.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.