Nanoparticles can induce changes in the intracellular metabolism of lipids without compromising cellular viability Ewa Przybytkowski, Maik Behrendt, David Dubois and Dusica Maysinger Department of Pharmacology and Therapeutics, McGill University, Montre ´ al, Canada Introduction Quantum dots (QDs) are colloidal semiconductor nanoparticles (NPs) with unique luminescence charac- teristics and wide biological and industrial applications [1,2]. They could become attractive tools for imaging in basic research and, eventually, in medicine [3]. How- ever, some QDs can be harmful to cells, particularly if Keywords fat oxidation; hypoxia; lipid droplets; nanoparticles; quantum dots Correspondence D. Maysinger, Department of Pharmacology and Therapeutics, McGill University, 3655 Promenade Sir-William-Osler, Montre ´ al, QC, Canada, H3G 1Y6 Fax: (514) 398 6690 Tel: (514) 398 1264 E-mail: dusica.maysinger@mcgill.ca (Received 5 June 2009, revised 17 August 2009, accepted 24 August 2009) doi:10.1111/j.1742-4658.2009.07324.x There is growing concern about the safety of engineered nanoparticles, which are produced for various industrial applications. Quantum dots are colloidal semiconductor nanoparticles that have unique luminescence char- acteristics and the potential to become attractive tools for medical imaging. However, some of these particles can cause oxidative stress and induce cell death. The objective of this study was to explore quantum dot-induced metabolic changes, which could occur without any apparent cellular dam- age. We provide evidence that both uncoated and ZnS-coated quantum dots can induce the accumulation of lipids (increase in cytoplasmic lipid droplet formation) in two cell culture models: glial cells in primary mouse hypothalamic cultures and rat pheochromocytoma PC12 cells. Glial cells treated with CdTe quantum dots accumulated newly synthesized lipids in a phosphoinositide 3-kinase-dependent manner, which was consistent with the growth factor-dependent accumulation of lipids in PC12 cells treated with CdTe and CdSe ⁄ ZnS quantum dots. In PC12 cells, quantum dots, as well as the hypoxia mimetic CoCl 2 , induced the up-regulation of hypoxia- inducible transcription factor-1a and the down-regulation of the b-oxida- tion of fatty acids, both of which could contribute to the accumulation of lipids. On the basis of our results, we propose a model illustrating how nanoparticles, such as quantum dots, could trigger the formation of intra- cellular lipid droplets, and we suggest that metabolic measurements, such as the determination of fat oxidation in tissues, which are known sites of nanoparticle accumulation, could provide useful measures of nanoparticle safety. Such assays would expand the current platform of tests for the determination of the biocompatibility of nanomaterials. Abbreviations DIV 8, day (in vitro) 8; FAS, fatty acid synthase; FFA, free fatty acid; HIF-1a, hypoxia-inducible factor-1a; HIFs, hypoxia-inducible transcription factors; LD, lipid droplet; NP, nanoparticle; PEG, polyethylene glycol; PI3K, phosphoinositide 3-kinase; PSN, penicillin ⁄ streptomycin ⁄ neomycin; QD, quantum dot; ROS, reactive oxygen species; SCD-1, stearoyl-coenzyme A desaturase-1; SREBP-1, sterol regulatory element binding protein-1. 6204 FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS their surface is not fully protected or if they degrade within the biological environment. We have studied the effects of QDs on living cells and have reported their internalization, the intracellular production of reactive oxygen species (ROS) and the damage to mul- tiple cellular sites induced by various QDs [4–6]. We, as well as others, have shown that the degree of inter- nalization of QDs and various other NPs is dependent on their size, surface charge, concentration in the med- ium and duration of exposure [4,7–11]. We have also shown the release of free cadmium from QDs contain- ing a CdTe core [6]. However, this could not explain fully their harmful effects. We have postulated that intracellular ROS formation or interactions with cellu- lar structures (mitochondria in particular) could con- tribute to the observed cytotoxicity [4–7]. Although coating of NPs with ZnS or polyethylene glycol (PEG) commonly prevents some undesirable effects [6,12], the long-term stability of these materials in the biological milieu is not well understood [13]. The purpose of this study was to investigate the more subtle effects of QDs which could occur without any evident morphological cellular damage or cell death. In particular, we explored QD-induced changes in lipid metabolism. Using well-characterized in vitro cell model systems [primary mouse hypothalamic cultures and pheochromocytoma cells (PC12)], we have provided evidence that poorly fluorescent CdTe NPs, without ZnS capping but with cysteamine coating, as well as highly fluorescent CdSe ⁄ ZnS NPs, capped with ZnS and coated with cysteamine on the surface, can induce the accumulation of lipids in cytoplasmic lipid droplets (LDs). LDs are macromolecular lipid assemblies consisting of neutral lipids, such as triacylglycerols, diacylglyce- rols, cholesterol esters and cholesterol, surrounded by a monolayer of phospholipids [14]. Many cell types are able to store excess fat as cytoplasmic LDs. How- ever, under physiological conditions, LDs are found mostly in tissues involved directly in energy meta- bolism, such as adipocytes, liver and muscles [14]. LDs are much more than simply blobs of fat segre- gated from the hydrophilic milieu of the cytoplasm. They are organelles with a particular structure and organization [15]. They are formed when free (uneste- rified) fatty acids (FFAs) from exogenous or endoge- nous sources are available inside the cells. Such FFAs are either esterified to form complex lipids, which are then stored in droplets or become part of the cellular membrane, or are oxidized in mitochondria for energy production. The formation and maintenance of LDs are complex, dynamic and highly regulated processes [15,16]. The formation of LDs induced by NPs could be par- tially explained by a reduction in fat oxidation, which occurs in parallel with an increase in LD number. As the measurement of fat oxidation is relatively simple to perform and gives clear objective values, it has the potential for broad application in the assessment of the metabolic effect described in this study. Given that high levels of cytoplasmic LDs present in nonadipose tissues are considered to be harmful, such assays would expand the current platform of tests for the determination of nanomaterial biocompatibility. Excess fat in nonadipose cells may be involved in several human pathologies, such as fatty liver, obesity, athero- sclerosis and type 2 diabetes, and may contribute to the development of insulin resistance and lipotoxic tis- sue damage [17]. The accumulation of neutral lipids in cytoplasmic LDs occurs following exposure to mito- chondrial toxins [18], during chronic viral infections [19,20], in response to protease inhibitors [21] and during hypoxia [22–24]. In this study, we also showed that exposure to colloidal semiconductor NPs leads to an increased expression of hypoxia-inducible transcription factor- 1a (HIF-1a). The family of hypoxia-inducible tran- scription factors (HIFs) regulates the adaptation to hypoxic conditions, which is critical for cell survival during decreased availability of oxygen in tissues [25,26]. Hypoxia and hypoxia-related signaling have been associated with major pathologies, such as car- diovascular disease, stroke and cancer [27]. The sig- naling for hypoxia was of interest in this study because QDs, which are redox-active NPs, can release Cd 2+ and induce the intracellular formation of ROS, and, as such, make good candidates for hypoxia mimetics. In addition, signaling induced by hypoxia promotes alterations in cellular metabolism. HIFs bind to DNA, forming heterodimers composed of one oxygen-regulated a subunit (HIF-1a, HIF-2a and HIF-3a) and one stable b subunit [28]. In normoxia, the oxygen-dependent hydroxylation of critical proline within the a subunit promotes its degradation by the ubiquitin–proteosome system. At low oxygen concen- trations, HIF-a subunits become more stable and thus can participate in the transcription of target genes, initiating the hypoxic response [25,26,28]. It is well accepted that HIF-a subunits can also be stabi- lized at normoxia by nonhypoxic stimuli, such as ROS, divalent metals and some mitochondrial meta- bolites [29,30]. In this study, we used CoCl 2 as a con- trol for the induction of signaling for hypoxia [31,32]. QDs were much poorer inducers of HIF-1a than was CoCl 2 . In addition, the induction of HIF-1a by QDs was not correlated with the accumulation of lipids, E. Przybytkowski et al. Nanoparticle-induced metabolic changes FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS 6205 thus suggesting that the two phenomena are most probably independent. In the present study, we used hypothalamic glial cells as a model system because the hypothalamus is a brain structure that is not completely protected by the blood–brain barrier and thus can be accessed by xeno- biotics and NPs [33]. Various other tissues, such as liver, kidney, spleen and bone marrow, are known sites of NP accumulation in vivo [34,35]. The evaluation of fat oxidation in these tissues could complement current toxicological assays for the safety screening of NPs and other nanomaterials. Results Short- and long-term effects of uncoated CdTe QDs on mouse primary hypothalamic cultures The effects of green, positively charged CdTe QDs with cysteamine surfaces [4,6,7] were investigated in primary mouse hypothalamic cultures. Mixed neural cultures were obtained from 5-day-old animals, and experiments were initiated at day (in vitro) 8 (DIV 8), when neural cells were fully differentiated (Fig. 1A). A few neurons with small cell bodies were visible on top of supporting glia. In this study, we focused on glial cells. To examine the short-term effects, the cultures were exposed to QDs (0–20 lgÆmL )1 ) for a period of 24 h in serum-free Neurobasal A medium with supplements. Cells responded to QD treatment by forming multiple cytoplasmic LDs (Fig. 1B–D). The number of LDs increased with increasing concentration of QDs (Fig. 1E). Within this time period, cells exposed to relatively low concentrations of QDs (5 lgÆmL )1 ) were viable, and only the highest concentration (20 lgÆmL )1 ) induced cell detachment and cell death (data not shown). To examine the effects of long-term exposure, primary mouse hypothalamic cultures were exposed to 5 lgÆmL )1 of QDs for 4 days. The cultures were viable and contained multiple cytoplasmic LDs (Fig. 1G). The number of LDs in control, untreated glial cells increased gradually with aging of the cultures. However, cells treated with QDs for 4 days contained more and much larger droplets than those found in the corresponding controls (Fig. 1F versus Fig. 1G). A B C D E F G Fig. 1. CdTe QDs trigger the formation of lipid droplets in glial cells from primary mouse hypothalamic cultures. (A–D, F, G) Representative photomicrographs of primary mouse hypothalamic cultures. (A) Phase contrast photomicrograph taken at DIV 9. (B–D) Photomicrographs of cultures stained with Oil Red O to visualize neutral lipids at DIV 9. (B) Control untreated cultures. (C, D) Cultures treated for 24 h (DIV 8 to DIV 9) with 5 lgÆmL )1 (C) and 10 lgÆmL )1 (D) of CdTe QDs. (E) The mean number of lipid droplets per microscopic field evaluated as described in Materials and methods. The data represent the mean and SEM from two independent experiments (n = 10). (F, G) Photomicro- graphs of cultures stained with Oil Red O at DIV 12. (F) Control untreated cultures. (G) Cultures treated with 5 lgÆmL )1 of CdTe QDs for 4 days (DIV 8 to DIV 12). The scale bars correspond to 50 l m. Nanoparticle-induced metabolic changes E. Przybytkowski et al. 6206 FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS LD formation triggered by CdTe QDs in glial cells depends on the de novo synthesis of lipids and phosphoinositide 3-kinase (PI3K) signaling Cytoplasmic LDs contain neutral lipids, including triacylglycerols, diacylglycerols and cholesterol esters. To verify whether de novo fat synthesis is involved in LD formation during treatment with QDs, hypotha- lamic cultures were exposed to CdTe QDs (5 lgÆmL )1 ) in the presence of the fatty acid synthase (FAS) inhibi- tor cerulenin (5 lgÆmL )1 ). FAS is responsible for the synthesis of FFAs, which can then be esterified to form components of cell membranes or lipids stored in LDs. Cerulenin, an antifungal antibiotic isolated from Cephalosporium caerulens, irreversibly binds to FAS, thereby inhibiting its activity [36]. Treatment with 5 lgÆmL )1 of cerulenin for 24 h had little effect on glial cell viability and resulted in the disappearance of LDs (Fig. 2), suggesting that these cells carried out lipogenesis (de novo synthesis of lipids) in control cultures. Interestingly, cerulenin also inhibited LD formation in cultures treated with 5 lgÆmL )1 CdTe QDs for 24 h (Fig. 2C–E). These results suggest that increased LD formation triggered by low dose (5 lgÆmL )1 ) CdTe QDs in glial cells involves de novo lipid synthesis. The PI3K ⁄ Akt signaling pathway is best known for its role in the maintenance of cell survival, but is also responsible for the up-regulation of glucose metabo- lism and the induction of lipogenesis [37–40]. We hypothesized that signaling via the PI3K ⁄ Akt pathway was involved in the promotion of the formation of LDs in cells treated with QDs. To test this hypothesis, primary mouse hypothalamic cultures were exposed to the PI3K inhibitor LY294002 (50 lm), alone or in combination with CdTe QDs (5 lgÆmL )1 ). Treatment with LY294002 completely inhibited LD formation in glial cells under control conditions (Fig. 3A, B, E), and A B C D E Fig. 2. Formation of LDs induced by QDs in glial cells from primary mouse hypothalamic cultures depends on the de novo synthesis of lipids. (A–D) Representative photomicro- graphs of primary mouse hypothalamic cultures stained with Oil Red O. (A) Control cultures at DIV 9. Cultures treated for 24 h with 5 lgÆmL )1 of cerulenin (B), 5 lgÆmL )1 of CdTe QDs (C) and both 5 lgÆmL )1 of CdTe QDs and 5 lgÆmL )1 of cerulenin (D). The scale bars correspond to 50 lm. (E) The quantification of the lipid droplet number from (A) to (D). The data represent the mean and SD from three independent fields. ***P < 0.001. E. Przybytkowski et al. Nanoparticle-induced metabolic changes FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS 6207 inhibited the formation of LDs in QD-treated cells (Fig. 3C–E), indicating that the formation of LDs was dependent on signaling via PI3K. Glial cell survival was not markedly compromised by the inhibition of PI3K signaling with LY294002 in a fully supplemented medium and within the tested time period (Fig. 3B, D). QDs induce the formation of LDs in PC12 cells in a growth factor-dependent manner without compromising cellular viability Rat pheochromocytoma PC12 cells are commonly used as a model cell line to study trophic factor signaling, cell death by trophic factor deprivation [41–43] and effects of NPs [4,6]. The cells were treated with two different types of QD in two different culture condi- tions: in full medium and in serum-free medium (buffered with 10 mm Hepes). We used CdTe QDs, which contain a CdTe core and have no protective shell on the surface, as well as CdSe ⁄ ZnS QDs (i.e. CdSe core and ZnS shell) [44]. The latter are much less harmful to cells [6,12]. Indeed, CdSe ⁄ ZnS QDs were not toxic to PC12 cells during 24 h of exposure in the presence or absence of serum, but unprotected CdTe QDs were nontoxic only in the presence of serum (Fig. 4A, B). Serum albumin can reduce QD entry into the nucleus [7], and serum trophic factors can support cell survival by signaling via the PI3K ⁄ Akt pathway [38,45]. PC12 cells treated in the presence of serum with both CdTe and CdSe ⁄ ZnS QDs contained more LDs than did untreated con- trols (Fig. 4C–E, K). The effect was dose dependent and was much more pronounced with uncoated CdTe QDs (Fig. 4K). The activation of PI3K ⁄ Akt was necessary for LD formation in PC12 cells (Fig. 4F–H). A B C D E Fig. 3. Formation of LDs induced by CdTe QDs in glial cells from primary mouse hypo- thalamic cultures depends on the PI3K signaling pathway. (A–D) Representative photomicrographs of primary mouse hypo- thalamic cultures stained with Oil Red O. (A) Control cultures at DIV 11. (B) Cultures treated with 50 l M LY294002 for 2 days (DIV 9 to DIV 11). (C) Cultures treated with 5 lgÆmL )1 of CdTe QDs for 3 days (DIV 8 to DIV 11). (D) Cultures treated with 5 lgÆmL )1 of CdTe QDs and 50 lM LY294002 for 3 days (DIV 8 to DIV 11). The scale bars correspond to 50 lm. (E) The quantification of the lipid droplet number from (A) to (D). The data represent the mean and SD from three independent fields. **P < 0.01. Nanoparticle-induced metabolic changes E. Przybytkowski et al. 6208 FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS As QDs are redox-active NPs (effective electron donors and acceptors) which can release Cd 2+ and are able to induce intracellular formation of ROS in PC12 and MCF 7 cells [4,6], they make good candidates for hypoxia mimetics [27]. Thus, we used the known hypoxia mimetic, CoCl 2 , and tested whether it also induced the formation of LDs in PC12 cells in a growth factor-dependent manner. PC12 cells treated with 100 lm CoCl 2 for 24 h contained more and larger LDs than the control (Fig. 4I versus Fig. 4C, K). Cells treated with CoCl 2 in the absence of serum did not contain LDs, suggesting that the activation of PI3K ⁄ Akt by trophic factors is also necessary for LD formation induced by this hypoxia mimetic (Fig. 4J). A B C D E H G F K I J Fig. 4. Uncoated and coated QDs trigger the formation of LDs in PC12 cells in a growth factor-dependent manner without compromising cellular viability. Pheochromocytoma PC12 cells were treated with QDs in fully supplemented medium (A, C–E, I and K) and in serum-free medium (B, F–H, J). (A, B) Percentage cell survival relative to control (determined with Alamar blue assay) after exposure to two different concentrations of QDs in fully supplemented medium (A) and in serum-free medium (B). Data represent the mean and SEM from two inde- pendent experiments. Representative photomicrographs of cells stained with Oil Red O. (C, F) Control untreated cells. (D, G) Cells treated with 20 lgÆmL )1 of CdTe QDs for 24 h. Cells treated with 20 lgÆmL )1 of CdSe ⁄ ZnS QDs (E, H) and with 100 lM CoCl 2 (I, J) for 24 h. The scale bars correspond to 50 lm. ***P < 0.001; *P < 0.05. (K) Number of lipid droplets found in PC12 cells treated with two different con- centrations of QDs in fully supplemented medium. The data represent the mean and SEM from two independent experiments. E. Przybytkowski et al. Nanoparticle-induced metabolic changes FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS 6209 QDs can induce signaling implicated in the response to hypoxia and can reduce the rate of fat oxidation in PC12 cells We hypothesized that QDs and CoCl 2 induce the accu- mulation of lipids in cytoplasmic LDs by the activa- tion of HIF-1a. Transcription factors involved in signaling for hypoxia are known to stimulate glucose metabolism [26] and to promote the accumulation of lipids [23,24]. PC12 cells were incubated with QDs or CoCl 2 for 24 h, and the expression of HIF-1a protein was analyzed by western blotting. QDs caused the up-regulation of HIF-1a, but only when added at higher concentration (20 lgÆmL )1 ) and to serum-free medium (Fig. 5A, B). Thus, it is unlikely that HIF-1a is involved directly in the accumulation of lipids trig- gered by QDs under the conditions of uncompromised cell survival in the presence of serum. On the basis of the results obtained in this study showing that glial cells treated with QDs accumulate newly synthesized lipids, we further hypothesized that NP-induced lipid accumulation could be a result of the down-regulation of b-oxidation of FFAs. Saturated fatty acid palmitate (C16:0), synthesized de novo,is elongated, desaturated and esterified to form other FFAs and eventually more complex lipids (including triacylglycerols stored in LDs), or can be transported to mitochondria where it is oxidized. The down-regula- tion of the b-oxidation of palmitate in mitochondria would provide more FFAs available for esterification and storage in LDs. PC12 cells were treated with QDs for 24 h and the rate of fat oxidation was measured. The oxidation of exogenous palmitate was decreased in cells treated with QDs or CoCl 2 in a dose-dependent manner in both the presence and absence of serum (Fig. 6A, B). In the presence of serum, fat oxidation decreased by 25–40% after treatment with CdTe QDs and by 20% after treatment with 20 lgÆmL )1 of CdSe ⁄ ZnS QDs. In the absence of serum, the effect of QDs was even more pronounced, as fat oxidation decreased by 40–50% after treatment with CdTe QDs and by 19–36% after treatment with CdSe ⁄ ZnS QDs (Fig. 6A). These results strongly suggest that: (a) QDs can induce changes in cellular lipid metabolism with- out affecting cellular viability; and (b) QD-induced A B Fig. 5. Uncoated and coated QDs increase the expression of HIF- 1a in PC12 cells. Pheochromocytoma PC12 cells were treated with QDs for 24 h in fully supplemented medium (A) and in serum-free medium (B). After treatment, HIF-1a protein levels were analyzed by western blot. A B Fig. 6. Uncoated and coated QDs decrease the rate of b-oxidation of fatty acids in PC12 cells. Pheochromocytoma PC12 cells were treated with QDs for 24 h in fully supplemented medium (A) and in serum-free medium (B). After treatment, fatty acid oxidation was measured using [1- 14 C]palmitate as a substrate. The data represent the mean and SEM from two independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. Nanoparticle-induced metabolic changes E. Przybytkowski et al. 6210 FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS accumulation of lipids in cytoplasmic LDs could be explained, in part, by the down-regulation of fat oxida- tion triggered by these NPs. Discussion This study shows that exposure of cells to CdTe and CdSe NPs can affect the intracellular metabolism of lipids and induce HIF-1a-mediated signaling. The results suggest that QDs, together with trophic factors, promote the accumulation of lipids in cytoplasmic LDs, in part by down-regulating the oxidation of de novo-synthesized fatty acids (Fig. 7). PC12 cells, as most cell types, require trophic factors for survival and differentiation [41–43]. When placed into culture, they will not grow or survive for extended periods of time without trophic factors in the cellular medium. Serum, a mixture of proteins isolated from the blood (in this study, from the blood of bovine fetus), is a source of trophic factors. The PI3K ⁄ Akt signaling pathway is activated on the cytoplasmic side of the plasma membrane when various trophic factors involved in the regulation of cell growth, survival and proliferation bind to their receptors on the cell surface [38,42,43,45,46]. Stimulation of this signaling pathway also enhances metabolism, resulting in an increase in glucose uptake and an up-regulation of glycolysis [37,38]. In addition, PI3K signaling has been shown to be involved in the up-regulation of lipogenic enzymes, such as FAS, most probably via sterol regulatory element binding protein-1 (SREBP-1) transcription factor [39,40]. Thus, in many cell types, signaling via PI3K ⁄ Akt ensures the supply of substrate for lipid synthesis and enhances the activities of lipogenic enzymes, setting the stage for the accumulation of lipids. Consistent with this, we have observed a small number of LDs in glial cells and in PC12 cells grown in a fully supplemented medium (control conditions). Interestingly, when glial cells from primary mouse hypothalamic cultures were exposed to QDs, the num- ber and ⁄ or size of LDs increased markedly. LDs have been recognized recently as ubiquitous dynamic organ- elles, which communicate with other cellular compart- ments and participate in important functions, such as transport and communication between different vesi- cles and compartments inside the cell [47]. Some of these functions are probably dependent on the pres- ence of specific proteins on the surface of droplets [48]. It has been shown previously that QDs can cause distortion and ⁄ or damage to cellular membranes, which are composed mostly of complex lipids [49]. This could result in the release of FFAs, which then would be available for esterification and the formation of triacylglycerols (the main components of LDs). We considered such a possibility; however, the inhibition of PI3K with LY294002 and the inhibition of FAS with cerulenin caused the disappearance of droplets and prevented the formation of new droplets during exposure to NPs when trophic factors were present in the medium. These findings indicate that glial cells (from primary mouse hypothalamic cultures) accumu- lated newly synthesized lipids when exposed to NPs. This was also true for PC12 cells treated with NPs, as they accumulated lipids mainly in the presence of serum. Thus, our results suggest that the fatty acids necessary for LD formation during exposure to NPs are synthesized de novo by the cells, rather than being released from the damaged membranes. However, NP interaction with particular membrane domains, such as Trophic factors QDs/CoCl 2 PI3K/AKT ROS LY294002 Serum withdrawa l SREBP-1 FAS (Lipogenesis) Cerulenin HIF-1α FFA ? LD Esterification FFA oxidation in mitochondrion (Fat utilization) products of FFA (Fat storage) Fig. 7. A model illustrating how colloidal semiconductor nanoparti- cles, such as QDs, could trigger the formation of intracellular lipid droplets. Activation of the PI3K ⁄ Akt pathway by trophic ⁄ growth factors creates the metabolic state, in which cells are able to syn- thesize FFAs. These newly synthesized FFAs are stored in LDs in the form of triacylgycerols or are oxidized in mitochondria. We hypothesize that QDs interfere with this processes by down-regu- lating fat oxidation. As a result, more FFAs become available for esterification and storage in LDs. The PI3K ⁄ Akt signaling pathway stimulates lipogenesis via SREBP-1. FAS is the enzyme responsible for the synthesis of palmitate, the precursor of FFAs. The inhibition of signaling by trophic factors, removal of trophic factors or inhibi- tion of fat synthesis result in the down-regulation of lipogenesis and the disappearance of LDs. QDs can also induce the up-regula- tion of HIF-1a, most probably via the production of ROS. QDs and the hypoxia mimetic CoCl 2 down-regulate the oxidation of FFAs and induce the accumulation of lipids. Further studies are needed to clarify the relationship between HIF-1a-mediated signaling and the metabolism of lipids in cells exposed to nanoparticles. E. Przybytkowski et al. Nanoparticle-induced metabolic changes FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS 6211 caveolae, which consist of small invaginations in plasma membranes, containing the protein caveolin and a particular lipid content, could contribute to the accumulation of lipids. It has been shown recently that caveolae can act as regulatory sites for the synthesis and trafficking of triacylglycerols [50]. Moreover, it is well documented that NGF signaling via the trk recep- tor in PC12 cells involves caveolin [51], and that cave- olin associates with LDs [52]. Therefore, NPs may induce the accumulation of lipids in PC12 cells by disturbing the function of cellular membranes on the cell surface, as well as by disturbing cellular functions inside the cells. Several mechanisms could explain our results: (a) NPs could up-regulate de novo lipogenesis; (b) NPs could increase the esterification of FFAs; (c) NPs could down-regulate the b-oxidation of FFAs; and ⁄ or (d) NPs could modulate the expression of proteins involved in the retention of lipids in LDs, such as lip- ases or other LD-associated proteins. We have shown that at least one of these mechanisms is implicated in lipid accumulation triggered by NPs in PC12 cells. Exposure to ‘degenerated’ (uncoated and uncapped) CdTe NPs, as well as fluorescent CdSe ⁄ ZnS NPs, significantly down-regulated the b-oxidation of FFAs, making them available for esterification and storage in LDs. The rate of esterification itself was not altered by treatment with QDs (data not shown). However, the mechanisms involved in the down-regulation of fat oxidation by NPs require further investigation. Fat accumulation and the down-regulation of the b-oxidation of FFAs were induced in PC12 cells by exposure to both NPs and the hypoxia mimetic CoCl 2 . Several studies have shown that the accumulation of lipids may occur during hypoxia [22–24]. Intermittent hypoxia induced hyperlipidemia in mouse liver via signaling through SREBP-1 and via up-regulation of stearoyl-coenzyme A desaturase-1 (SCD-1) [22]. Hyp- oxic conditions also enhanced the synthesis of neutral lipids in human macrophages via the up-regulation of lipogenesis (increase in SCD-1 activity) and also via the down-regulation of the b-oxidation of fatty acids [23]. Hypoxia has also been shown to induce the for- mation of LDs in various tumors [24]. Both NP and CoCl 2 treatment induced the up-regulation of HIF-1a in PC12 cells. However, the up-regulation of HIF-1a by NP exposure was detectable only in serum-free con- ditions, whereas LDs were produced mainly in the presence of serum. These findings suggest that the up-regulation of HIF-1a may not be necessary for the accumulation of lipids induced by NPs in trophic factor-supplemented medium, or that changes in its levels were too subtle to be detected by western blot- ting [53]. Further studies are needed to clarify the rela- tionship between HIF-1a-mediated signaling and the metabolism of lipids in cells exposed to NPs. HIF- mediated signaling not only induces the expression of genes involved in cellular adaptation to low oxygen [26], but also alters the cell’s response to various stres- sors [27]. In this regard, it could be predicted that exposure to NPs could also modify the cellular responses to various xenobiotics. The lipid accumulation induced in PC12 cells by two types of NP was concentration dependent and largely preceded the manifestation of cell death. QDs are redox-active NPs (effective electron donors and accep- tors), which can induce the formation of ROS [4,35]. Several studies have shown that QDs and other NPs can generate ROS, particularly if they are exposed to light [54–58]. However, we and others have shown that exposure to QDs causes the intracellular production of ROS with and without illumination [4,6,59,60]. Both exogenously and endogenously produced ROS cause an imbalance in the cellular redox state, resulting in oxidative stress. It is possible that the accumulation of fat in cytoplasmic LDs in nonadipose cells, such as glia from the hypothalamus, may be an early sign of dam- age and ⁄ or oxidative stress induced by certain types of NP. There is evidence that NPs, such as fullerenes and carbon nanotubes, may also produce ROS in vitro [54] and in vivo [57]. If so, these NPs could probably induce changes in lipid metabolism and ⁄ or the up-regulation of HIFs. We have also tested NPs which are consid- ered to be safe and which do not induce oxidative stress in cells (gold NPs and latex beads) for their abil- ity to induce LD formation. We did not observe any increase in LD formation in PC12 cells exposed to these NPs. These findings corroborate results from studies with biological NPs and latex beads in macro- phages, reviewed by D’Avila et al. [61]. Thus, it appears that biological and artificial NPs which cause oxidative stress and ⁄ or the release of specific cytokines are potentially hazardous and are the prime candidates for LD induction. The majority of in vitro tests designed for the assess- ment of NP safety are based on viability assays, the peroxidation of membrane lipids, the depletion of cellular glutathione or the secretion of inflammatory mediators [35]. The results from the present study sug- gest that metabolic measurements, such as the determi- nation of changes in fat oxidation, could be used as an additional sensitive test for the evaluation of NP safety ⁄ biocompatibility. Metabolic measurements, especially those related to mitochondrial function and nonhypoxic induction of metabolic effects normally observed with hypoxia, such as changes in oxygen Nanoparticle-induced metabolic changes E. Przybytkowski et al. 6212 FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS consumption, are already being considered as impor- tant criteria for the evaluation of drug-induced toxicity [62–64]. The accumulation of lipids in cytoplasmic LDs in glial cells from primary mouse hypothalamic cultures has not been reported previously. Changes in lipid metabolism induced by metallic ions and NPs in hypo- thalamus, liver, kidney, spleen or bone marrow could compromise organismal homeostasis. Thus, metabolic measurements, such as the determination of the rate of fat oxidation, in these organs or tissues, which are also known sites of NP accumulation in vivo [34,35,65], could provide useful measures of NP biocompatibility and safety. Materials and methods Materials The sources of the reagents were as follows: phenol red-free RPMI, DMEM and Neurobasal A media, fetal bovine serum, B27 supplement, l-glutamine, penicillin ⁄ streptomycin ⁄ neomycin (PSN) cocktail, Ca 2+ ⁄ Mg 2+ -free NaCl ⁄ P i , 0.25% trypsin ⁄ EDTA, Hepes buffer, trypan blue solution and 545 amino (PEG) QDs from Invitrogen (Burlington, Canada); Oil Red O, Harris hematoxylin, polyornithine and laminine from Sigma (Oakville, Canada); formalin from Fisher Scientific (Nepean, Canada); Alamar blue reagent from Biosource (Montreal, Canada); paraformal- dehyde (PFA) from BDH Laboratories (Poole, UK); glia-specific rabbit GFAP antibody (Z0334) from Doko Cytomation (Glostrup, Denmark); anti-HIF-2a IgG (NB100-122SS) from Novus Biologicals (Littleton, CO, USA); Texas red goat anti-rabbit IgG-conjugated second- ary antibody (111-075-045) from Jackson ImmunoResearch (West Grove, PA, USA); Hoechst Dye 33342 (H1399) from Molecular Probes (Carlsbad, CA, USA); Aqueous Mount mounting medium from ScyTec (Hornby, Canada); and VECTASHIELD medium from VECTOR Laboratories Inc. (Burlingame, CA, USA). Nanoparticles CdTe and CdSe ⁄ ZnS NPs were prepared and characterized as originally described by Gaponik et al. [66] and modified by Lovric et al. [7]. CdTe NPs were prepared as described by Lovric et al. [7]; they contain a CdTe core, their dia- meter is 2.8 nm and they have an emission maximum at 535 nm. There was no ZnS cap on the CdTe core and cysteamine was attached directly to the surface [7]. CdSe ⁄ ZnS NPs were prepared as described by Hoshino et al. [44]; they contain a CdSe core, their diameter is 2.4 nm and they have an emission maximum of 518 nm. The CdSe core was capped with one layer of ZnS to which cysteamine was attached, and thus the particle size, measured by the dynamic light scattering method, was about 10 nm [44]. QDs were added to the cellular media in different amounts and for different lengths of time, as indicated in the figures. Primary mouse hypothalamic cultures All experiments were conducted with the approval of the McGill University Animal Care Committee. Mouse (strain 129T2 ⁄ SV) hypothalamus was obtained by dissection at postnatal day 5 and freed from the meninges. Tissue pooled from at least six animals was stored in ice-cold sterile Ca 2+ ⁄ Mg 2+ NaCl ⁄ P i . The tissues were dissociated mechan- ically by gentle pipetting using sterile 1 mL pipette tips, and digested with 0.25% trypsin ⁄ EDTA at 37 °C for 10 min. Dissociated cells were resuspended in DMEM med- ium containing 10% fetal bovine serum and PSN cocktail. Cells in suspension were seeded onto polyornithine- and laminine-coated 12 mm 2 glass microscope slide coverslips and incubated in a 95% air ⁄ 5% CO 2 atmosphere at 37 °C for 1 h. The coverslips with attached cells were placed in DMEM medium containing 10% fetal bovine serum and PSN for 24 h in 24-well plates (Corning, Nepean, Canada). The next day, the cells were washed twice with pre-warmed Neurobasal A medium and, finally, Neurobasal A medium supplemented with B27, l-glutamine and PSN was added to the cultures. The cultures were maintained at 37 °Cina 95% air ⁄ 5% CO 2 atmosphere. Experiments were initiated at DIV 8. Cell culture Rat pheochromocytoma (PC12, ATCC # CRL-1721) cells were cultured at 37 °C in phenol red-free RPMI medium containing 2 mm glutamine and 10% fetal bovine serum. For survival experiments, cells were seeded onto 24-well plates (25 000 cells per well) and, for LD staining, onto 12 mm 2 glass coverslips (13 000 cells per coverslip). Treat- ments with NPs were performed in serum-containing and serum-free medium buffered with 10 mm Hepes (pH 7.4). LD staining and quantification LDs were stained with Oil Red O, as described in detail by Przybytkowski et al. [67]. Briefly, cells were washed with NaCl ⁄ P i and incubated with Oil Red O working solution for 15 min at room temperature. The cells were then washed once with NaCl ⁄ P i , and fixed with 10% formalin for 25 min. Subsequently, the cells were washed again with NaCl ⁄ P i , stained for 2 min with Harris hematoxylin, washed with distilled water and mounted on microscopic slides using Aqueous Mount mounting medium. Photomicrographs were taken from representative fields using an Olympus BX2 microscope (Olympus America Inc., Center Valley, PA, E. Przybytkowski et al. Nanoparticle-induced metabolic changes FEBS Journal 276 (2009) 6204–6217 ª 2009 The Authors Journal compilation ª 2009 FEBS 6213 [...]... plastic tubes containing folded glass fiber filter paper (GF ⁄ B; Whatman International Ltd., Maidstone, UK) saturated with 0.15 mL of 5% KOH The sealed flasks were incubated for 1 h at 37 °C Control blank flasks contained all reagents without cells The reaction was stopped by the injection of 0.3 mL of 3 m sulfuric acid through the serum stopper into each flask with a syringe, and the flasks were then shaken gently... temperature The filters were then removed and placed into scintillation vials containing scintillation liquid (ScintiSafe Plus 50%, Fisher Scientific Canada, Ottawa, Canada) Radioactivity was counted 24 h later, using a liquid scintillation counter (Wallac 1410; GMI, Inc., Ramsey, MN, USA) and an Easy DPM protocol Results were expressed as nanomoles of CO2 released per hour per milligram of cell protein Western... analysis Statistical significance was determined by analysis of variance The differences were considered to be significant at P < 0.05 (*), P < 0.01 (**) and P < 0.001(***) Acknowledgements Financial support for this work was provided in part by the Natural Science and Engineering Research Council of Canada (NSERC) and Canadian Institutes of Health Research (CIHR) References 1 Medintz IL, Uyeda HT, Goldman... irritation potential of quantum dot nanoparticles in epidermal keratinocytes J Invest Dermatol 127, 143–153 Mancini MC, Kairdolf BA, Smith AM & Nie S (2008) Oxidative quenching and degradation of polymer-encapsulated quantum dots: new insights into the long-term fate and toxicity of nanocrystals in vivo J Am Chem Soc 130, 10836–10837 Murphy DJ (2001) The biogenesis and functions of lipid bodies in animals,... Drezek R (2008) Cytotoxicity of nanoparticles Small 4, 26–49 36 Omura S (1976) The antibiotic cerulenin, a novel tool for biochemistry as an inhibitor of fatty acid synthesis Bacteriol Rev 40, 681–697 37 Whiteman EL, Cho H & Birnbaum MJ (2002) Role of Akt ⁄ protein kinase B in metabolism Trends Endocrinol Metab 13, 444–451 38 Jones RG & Thompson CB (2007) Revving the engine: signal transduction fuels... serum-free survival of human breast cancer cells Biochem Cell Biol 85, 301–310 68 O’Brien J, Wilson I, Orton T & Pognan F (2000) Investigation of the Alamar Blue (resazurin) fluorescent dye for the assessment of mammalian cell cytotoxicity Eur J Biochem 267, 5421–5426 69 Averill-Bates DA & Przybytkowski E (1994) The role of glucose in cellular defences against cytotoxicity of hydrogen peroxide in Chinese hamster... Quantum dots and other fluorescent nanoparticles: quo vadis in the cell? Adv Exp Med Biol 620, 156–167 Behrendt M, Sandros MG, McKinney RA, McDonald K, Kriz J, Przybytkowski E, Tabrizian M & Maysinger D (2009) Cell imaging in real time and organelle distribution of fluorescent InGaP ⁄ ZnS nanoparticles Nanomedicine, in press Harush-Frenkel O, Altschuler Y & Benita S (2008) Nanoparticle–cell interactions:... to a final concentration of 10% (v ⁄ v), and the cells were incubated for 2 h in a humidified atmosphere with 5% CO2 at 37 °C Fluorescence was measured in 50 lL aliquots of medium (in duplicate) using a FLUOstar OPTIMA spectrofluorometer (BMG Labtech, Offenburg, Germany) (kexc = 535 nm; kem = 580 nm) one flask were used for protein determination and western blotting Cells from the second and third flasks... of hypoxia-inducible factor 1 J Biol Chem 270, 1230–1237 33 Toni R, Malaguti A, Benfenati F & Martini L (2004) The human hypothalamus: a morpho-functional perspective J Endocrinol Invest 27, 73–94 34 Fischer HC, Liu L, Pang KS & Chan WCW (2006) Pharmacokinetics of nanoscale quantum dots: in vivo distribution, sequestration, and clearance in the rat Adv Funct Mater 16, 1299–1305 35 Lewinski N, Colvin... discarded from these two flasks and replaced with 0.9 mL of fresh MEM containing 0.1% BSA Subsequently, 100 lL of 10 · reaction mix (prepared freshly 1 h in advance and containing 10 mm carnitine, 1 mm palmitate, 4% BSA and 0.5 lCi per culture flask of labeled fatty acid) were added to each flask The flasks were immediately sealed with rubber serum vial stoppers The stoppers were fitted inside with plastic . Nanoparticles can induce changes in the intracellular metabolism of lipids without compromising cellular viability Ewa Przybytkowski,. cancer [27]. The sig- naling for hypoxia was of interest in this study because QDs, which are redox-active NPs, can release Cd 2+ and induce the intracellular