Heat Shock Protein 27 Protects Against Atherogenesis via an Estrogen-Dependent Mechanism
Role of Selective Estrogen Receptor Beta Modulation
Objective— We recently identified HSP27 as an atheroprotective protein that acts extracellularly to prevent foam cell formation and atherogenesis in female but not male mice, where serum levels of HSP27 were increased and inversely correlated with degree of lesion burden. In the current study we sought to determine whether estrogens are required for the observed atheroprotective benefits of HSP27 as well as its extracellular release.
Methods and Results— In vitro estrogens prompted the release of HSP27 from macrophages in an ERβ specific manner that involved exosomal trafficking. Ovariectomy nullified the previously recognized attenuation in aortic lesion area in HSP27o/eapoE−/− mice compared to apoE−/− mice. Supplementation with 17β-estradiol resulted in a >15× increase in uterine weight and attenuation of atherogenesis in all mice, although HSP27o/eapoE−/− had 34% less lesion burden compared to apoE−/− mice. Mice treated with the ERβ-specific agonist, DPN had no effect on uterine weight but a 28% decrease in aortic lesion area in HSP27o/eapoE−/− compared to apoE−/− mice. HSP27 serum levels showed a similar gradual increase with E2 and DPN replacement treatment but did not change in untreated mice.
Conclusions— The extracellular release of and atheroprotection provided by HSP27 is estrogen dependent.
Previously, we showed in human coronary arteries that Heat Shock Protein 27 (HSP27) expression is lower in atherosclerotic plaques compared to lesion-free arterial segments, suggesting higher levels of this protein may confer protection from disease.1 More recently, we demonstrated that HSP27 is an atheroprotective protein that acts in the extracellular space to reduce foam cell formation and atherogenesis by binding scavenger receptor-A on the surface of macrophages and preventing the uptake of acLDL as well as inflammatory cytokine release. Moreover, when fed a high-fat diet for 4 weeks apoE null (apoE−/−) mice that overexpress HSP27 (HSP27o/eapoE−/−) show a reduction in aortic atherosclerotic plaque area—however, only in female and not male mice.2 Interestingly, female HSP27 overexpressing mice have significantly higher levels of serum HSP27 than do their male counterparts, and serum HSP27 levels show a strong inverse correlation with atherosclerotic lesion area. These sex-specific effects of HSP27 hint that the function or release of this intriguing protein may in some way be hormonally modulated.
Given our previous observations that estrogens cause HSP27 secretion, and that HSP27 physically associates with ERβ but not ERα,3 we sought to determine the role for these receptors in the release of HSP27 both in vitro and in vivo. As described herein, we now report that the release of HSP27 from macrophages is preferentially induced via specific ERβ stimulation (not ERα) and induction of HSP27 release in vivo via the ERβ-specific modulator DPN increases serum HSP27 levels to a level comparable to 17β-estradiol (E2)—yet unlike E2 attenuates atherogenesis without causing unwanted uterine hypertrophy.
Materials and Methods
Chemicals and Reagents
All chemicals and reagents were purchased from Sigma unless otherwise noted. Diarylpropionitrile (DPN) and propyl pyrazole triol (PPT) and ICI 182 780 (ICI) were obtained from Tocris Bioscience.
All animals used in this study were approved by the University of Ottawa Animal Care and Veterinary Service Committee. For all surgical procedures, mice were placed under a 3% isoflurane/oxygen mixture and buprenorphine was administered twice daily for 3 days postsurgery for analgesia. Female mice 6 weeks of age were incised dorsally and ovaries removed bilaterally. Mice were allowed to recover for 1 week, at which point pellets containing either E2 (0.25 mg, 60-day release) or diarylpropionitrile (DPN; 0.25 mg, 60-day release) were implanted subcutaneously (Innovative Research America) using a trochar (7 to 9 mice per group).
Mouse Atherosclerosis Model
Mice overexpressing HSP27 (HSP27o/e) were generated at Imperial College London as previously described.4,5 HSP27o/e mice were crossed more than 8 generations onto an apoE−/− C57BL6 background to generate HSP27o/eapoE−/− and apoE−/− littermates. Genotyping was done in all animals for verification of both HSP27 and apoE−/− using PCR as previously described.2,4 Overexpression of HSP27 was verified using immunocytochemistry (using an antibody that recognizes both HSP25/HSP27 and an antibody that recognizes the HA-tag on the HSP27 protein) and real-time PCR (using a UPL probe and primers specific for HSP27). One week after pellet implantation (at 8 weeks of age), animals were placed on a high-fat diet containing 1.25% cholesterol and 15.8% fat (Harlan Teklad) for 4 weeks. At euthanasia, animals were anesthetized under isoflurane, and whole blood was collected through cardiac puncture. Hearts were perfused with phosphate-buffered saline (PBS) followed by 10% neutral buffered formalin (NBF) via the left ventricle, and the heart and aorta were removed and immersed in NBF overnight. Adventitial tissue was removed, and the aorta was opened longitudinally and stained with oil red O. Images were captured with a video camera and en face atherosclerotic lesions were analyzed by 2 observers using Image-Pro software (Media Cybernetics) and expressed as a percent lesion area per total area of the aortic arch. Uteri were removed and dissected of all fat and connective tissue before wet weight was obtained and expressed as a fraction of total body weight at sacrifice (see Table, first column). Serum was collected and analyzed for total cholesterol (Wako Pure Chemical Industries Ltd; see Table, second column) and IL-1β (Calbiochem).
HSP27 Serum Levels
Plasma levels of HSP27 were measured using an ELISA kit specific to human HSP27 (QIA119, Calbiochem). A total of 5 μL of serum from HSP27o/e apoE−/− and apoE−/− mice was diluted 1:10 in dilution buffer and assayed according to the manufacturer’s protocol. A standard curve of known amounts of HSP27 was constructed with each assay. This assay was found to have no cross-reactivity with mouse HSP25.
THP-1 human macrophages (American Type Culture Collection, ATCC; Manassas, Va) were cultured in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (FBS, Wisent), sodium pyruvate, glutamine, penicillin-streptomycin, fungizone, and β-mercaptoethanol (Invitrogen). Differentiation was induced with 100 nmol/L phorbol myristate acetate for 72 hours before any treatment was initiated. Twenty-four hours before treatment, cells were placed in phenol-red free media containing 10% charcoal-stripped FBS, and treated with E2 (100 nmol/L), ICI (10 nmol/L), PPT (10 nmol/L), or DPN (10 nmol/L) overnight unless otherwise stated. Analysis of conditioned media for HSP27 secretion was performed after verification of cell viability using the LDH release assay (CytoTox 96 Non-Radioactive Cytotoxicity Assay; Promega) and equal cell number per well. LDH release (a reflection of cell viability) did not differ between any of the treatments and was similar to the cell culture media alone (ie, without cells).
For Western blotting, 40 μL of conditioned media was loaded onto a 10% SDS-PAGE gel and separated at 120V using gel electrophoresis. Protein was then transferred to a PVDF membrane (BioRad) for 2 hours at 60V. Membranes were then subjected to western blotting using antibodies to HSP27 (Chemicon, mouse monoclonal, 1:200).
Analysis of Exosomal Secretion of HSP27
Human THP-1 macrophages were grown in culture as described above. Before stimulation with either E2 (100 nmol/L) or DPN (10 nmol/L), cells were treated for 1 hour in the presence of 15 nmol/L dimethylamiloride (DMA), which inhibits the secretion of exosomes by inhibition of the H+/Na+ and Na+/Ca2+ exchanger.6,7 Cells were then stimulated, in the presence of DMA, with ligand overnight. The conditioned media was collected and analyzed for HSP27 secretion as described above.
Human THP-1 cells were grown on cover slips, and treated with 100 nmol/L E2-BSA conjugated to FITC (Sigma) for 4 hours. Cells were fixed with BD Cytofix as directed by the manufacturer and blocked with 2% bovine serum albumin (BSA) for 2 hours. Antibodies against human HSP27 (mouse monoclonal, 1:200) and ERβ (rabbit polyclonal, 1:500) (Chemicon) were incubated overnight at 4°C. Visualization of substrates was done using secondary antibodies AlexaFluor 594 and 350, respectively. Negative controls included incubation with control IgG and secondary antibody alone. Coverslips were mounted with Dako Fluorescent mounting media (Dako Cytomation). Cells were visualized with an Olympus FluoView FV1000 confocal microscope (Olympus America Inc) at 100× magnification, using sequential scanning of each fluorophore to reduce any potential nonspecific excitation of the different fluorophores. Additionally, cells were visualized after staining with only one fluorophore and no bleed-through into the opposite channel was observed.
All data represent mean±SEM, except as specifically stated. Each experiment was conducted at least 3 times. Statistical analysis was performed with 1-way ANOVA by using SigmaStat 3.5 software. Differences were considered significant at probability value <0.05 and is denoted by an asterisk.
In Vitro HSP27 Is Released Into the Extracellular Space via Estrogen Stimulation
Given our previous observations that HSP27 in the extracellular space is atheroprotective and its release from cells appears to be induced via estrogen, we sought to determine which estrogen receptor (ERα or ERβ) preferentially promotes its extracellular release. Human macrophages in culture were treated with E2 in the presence or absence of the ER antagonist ICI. Treatment with E2 caused a dose-dependent release of HSP27 into the conditioned media, and upon treatment with ICI HSP27 levels reverted to near baseline (Figure 1A). Macrophages treated with E2 or compounds that specifically stimulate ERα or ERβ (PPT and DPN, respectively) produced varying effects on extracellular HSP27 levels. Stimulation with PPT for 4 hours did not increase HSP27 secretion beyond baseline, whereas DPN treatment resulted in extracellular HSP27 levels equivalent to E2 (Figure 1B). After 24 hours of DPN treatment HSP27 secretion induced higher extracellular HSP27 levels than E2 (Figure 1B, bottom). Conditioned media samples from all treatment groups were tested for cell viability using a lactose dehydrogenase (LDH) assay and no differences were found among the treatment groups (data not shown), thereby indicating that HSP27 release is not simply a nonspecific byproduct of cell death. Taken together, these in vitro results imply that ERβ is primarily responsible for the extracellular release of HSP27 from macrophages and perhaps can result in levels that even surpass those found with E2 alone.
Estrogen Is Required In Vivo for HSP27-Mediated Atheroprotection
Previously we demonstrated that after 4 weeks of a high-fat diet female (but not male) mice overexpressing HSP27 had 35% smaller atherosclerotic lesions compared to nonoverexpressing mice.2 We therefore asked whether estrogens are required for this observed effect by performing bilateral ovariectomy to remove all endogenous estrogens from 6-week-old female apoE−/− and HSP27o/eapoE−/− mice. After 4 weeks of high-fat atherogenic diet, the aortic arch was examined and lesion area measured using en face oil red O staining. In the absence of estrogen, HSP27o/eapoE−/− mice had similar lesion areas compared to apoE−/− mice (11.6±0.9% versus 11.7±1.3% of total arch area; P=0.99; n=8 mice/group; Figure 2A and 2B). To examine whether this atheroprotection could be restored with replacement of estrogen, both groups of mice (apoE−/− and HSP27o/eapoE−/−) were implanted with a subcutaneous time-release E2 pellet (0.25 mg) one week before commencing a high-fat diet. According to data provided by the manufacturer, these pellets release approximately 4.2 μg of estrogen per day for the duration of the study. The first major finding in all E2-treated mice was the greater than 15-fold increase in uterine weight (P≤0.05, Table)—thereby indicating the strong ERα component of E2 therapy. Second, aortic atherosclerotic lesion size was remarkably reduced in both groups of E2-treated mice, with HSP27o/eapoE−/− mice having 34% less lesion burden compared to apoE−/− mice (2.5±0.4% versus 3.7±0.4% of total arch area; P≤0.05). Therefore, although estrogen therapy was effective in reducing atherosclerosis in the absence of HSP27, it was also able to recapitulate the atheroprotective qualities of HSP27, in the overexpressing mice, demonstrating that indeed this effect is estrogen dependent.
Given that macrophages in vitro release HSP27 into the extracellular space on stimulation with DPN (an ERβ-specific agonist) but not PPT (an ERα-specific agonist), we next asked whether DPN could affect the HSP27-mediated atheroprotection in our ovariectomized mouse model. Before commencement of a high-fat diet, DPN pellets (0.25 mg) were implanted subcutaneously in a manner identical to that used for the E2 pellets. Unlike E2-treated mice, the uterine weight did not change in DPN treated mice (Table)—consistent with previous reports that DPN has minimal cross-reactivity with ERα at the doses used in the present study and hence lacks the undesirable effects on reproductive tissue. Moreover, aortic lesions from HSP27o/eapoE−/− mice treated with DPN were 28% smaller than DPN-treated apoE−/− mice (8.8±0.5% versus 12.2±0.8%, P≤0.05)—a result that was remarkably congruent with our original observations in ovary-intact mice (supplemental Figure I). Hence, these data confirm that stimulation of ERβ in vivo can reproduce the predicted HSP27 atheroprotective effects.
Extracellular HSP27 Release In Vivo Is Estrogen Dependent
To determine whether estrogen is required for the release of HSP27 into the extracellular space in vivo, serum levels from HSP27o/eapoE−/− mice from all treatment groups were examined for HSP27 expression at baseline, post-ovariectomy (ie, before hormone supplementation), as well as 2 weeks and 4 weeks after commencement of a high-fat diet. After ovariectomy and throughout the course of the 4-week atherogenic diet, HSP27 levels were similar to baseline in ovariectomized mice not receiving hormone treatment (Figure 3, solid line). Conversely, E2 supplementation caused a greater than 4-fold increase in HSP27 serum levels after 4 weeks compared to baseline (Figure 3, hatched line; P≤0.05). Similarly, the increase in serum HSP27 levels in DPN-treated mice closely mimicked those observed in E2 treated mice (Figure 3, gray line). These data demonstrate that, in line with our previously reported observations, estrogen is required for the release of HSP27 into the serum, possibly via stimulation of ERβ.
HSP27 Secretion Involves Exosomes
Given that other members of the HSP family are known to be secreted via an exosomal pathway, we next examined whether exosomes are involved in HSP27 secretion in response to stimulation.6 Macrophages were treated with an inhibitor of exosomes (dimethylamiloride, DMA) before stimulation with either E2 or DPN, after which conditioned media was collected and examined for HSP27 secretion. Inhibition of the exosomal pathway significantly decreased the level of HSP27 secreted into the media in response to either E2 and DPN (Figure 4A and 4B). Treatment with inhibitors of other cellular pathways, including inhibitors of the Golgi (brefeldin A), protein synthesis (cyclohexamide), transcription (actinomycin D), or ABC transporters (glybenclamide) did not alter the secretion of HSP27 in response to E2 (data not shown). Upon treatment with a cell-impermeable E2 (E2-BSA), HSP27 was localized to vesicle-like structures at the cell membrane, where it colocalized with ERβ (Figure 4C). Therefore, inhibition of exosomal trafficking in macrophages specifically abrogates the estrogen-induced secretion of HSP27.
Although there remains considerable debate regarding the potential role of estrogens in preventing atherosclerosis, the current study brings forth new ideas that help generate enthusiasm for hormonal modification of atherogenesis.8 We learn that estrogen is required for the acute protection from atherosclerosis provided by HSP27.2 After ovariectomy, when estrogen levels are essentially nonexistent, HSP27 overexpression no longer results in the release of HSP27 into the serum and protection from atherogenesis. Replacing estrogenic stimulation with either E2 or the ERβ-specific modulator DPN, restores the release of HSP27 into the serum and the resultant protection from lesion development.
Next, we show both in vitro and in vivo that ERβ can be specifically stimulated to enable the release of HSP27, demonstrating a novel mechanism of targeting HSP27 to exit the cell where it may protect against atherosclerosis. In ovariectomized mice DPN treatment recapitulated the previously observed 35% decrease in lesion area in HSP27-overexpressing mice compared to apoE−/− mice (see supplemental Figure I). Furthermore, DPN caused HSP27 to be secreted into the extracellular space both in vivo into the serum as well as in vitro from macrophages. In contrast, specific stimulation of ERα via PPT had a more modest effect on HSP27 release. The stimulation of ERβ may represent a novel molecular target for activation of HSP27 secretion for protection of atherogenesis. Although E2 also restored the increased release of HSP27 into the serum, its antiatherogenesis effects did not appear to be dependent on HSP27. Estrogen reduced aortic lesion area in both apoE−/− and HSP27o/eapoE−/− mice by more than 3-fold, clearly demonstrating the protective nature of E2 in the vessel wall. This reduction in lesion burden in both groups of E2-treated mice can be partially attributed to ERα, which among its many actions in the vessel wall (eg, vasodilation, reduced smooth muscle cell activation)9,10 contributes to a 45% reduction in total serum cholesterol (eg, OVX+E2 mice: 876.5±82.8 versus OVX mice: 1599.2±118.2 mg/dL; P≤0.05; Table).11 It is believed that ERα is necessary for the prevention of atherosclerosis at advanced stages of disease; indeed, studies have shown that deletion of ERα abrogates the protective nature of E2 on advanced atherosclerosis and vascular injury.11–13 However, recent evidence suggests that ERα may not be involved in estrogen-mediated protection from early lesion development, highlighting the importance of other estrogen receptor mechanisms in this initial stage of disease.14,15 Although in our current study estrogen was remarkably protective against lesion development, E2 increased uterine weight in both murine groups by more than 15-fold and raised serum IL-1β levels an average of 6-fold (Table). In contrast, stimulation with DPN did not increase uterine weight or circulating IL-1β, and yet produced an increase in serum HSP27 levels similar to those observed in the E2-treated HSP27-overexpressing mice. Interestingly, the in vitro data suggests that ERβ stimulation by DPN may produce a more sustained effect on HSP27 secretion than E2. Why DPN results in enhanced in vitro release of macrophage HSP27 at 24 hours posttreatment is unclear, but perhaps the lower potency of this ERβ ligand causes a less brisk release of HSP27 and hence less negative feedback compared to E2.16 It is intriguing to speculate why DPN failed to reduce atherogenesis in apoE−/− mice that do not overexpress HSP27, as certainly these mice are capable of expressing the endogenous mouse homolog, HSP25. From previous studies, we know that endogenous vessel wall expression of HSP27 (or HSP25) diminishes within with progression of atherosclerosis.1,17,18 As well, it is recognized that although the binding affinity of DPN is highly specific for ERβ compared to ERα, the ERβ transactivation potency of DPN is considerably less than that of E2 or other selective estrogen receptor modulators (SERMs).19 Therefore it is likely that the dose of DPN used in the present study was insufficient to activate ERβ in the presence of only modest levels of HSP25 in the nonoverexpressing mice (ie, in comparison to the high levels found in the transgenic mice). Higher levels of DPN may be required to demonstrate an ERβ-mediated atheroprotection in the setting of normal or reduced HSP27 (or HSP25) expression, much like those achieved by other ERβ-specific agonists.20,21
Another novel finding from the present study is that HSP27 is secreted from macrophages via the exosomal pathway in response to hormonal stimulation. Interestingly, although this is the same pathway employed by other HSPs (eg, HSP60, HSP70),6,22 this often results in a proinflammatory response, whereas HSP27 release primarily promotes antiinflammatory signaling.2,23,24 Confocal microscopy revealed that HSP27 exists near the cell membrane, where it colocalizes with both E2 and ERβ, supporting a role for membrane-like vesicles mediating the release of HSP27. Perhaps this vesicle-associated pool of HSP27 is responsible for the extracellular HSP27 that can be detected as early as 1 hour poststimulation.2 Moreover, although we had previously found HSP27 to physically associate with ERβ, this is the first evidence that these 2 proteins interact in macrophages and that this interaction may occur at the cell surface, enabling HSP27 to be released.1,3 Of note, treatment of macrophages with the global estrogen receptor antagonist ICI 182,780 did not completely abrogate HSP27 secretion, insinuating that there may receptor-independent pathways involved in this release mechanism.
In summary, we have presented novel data on the integral role of estrogens in promoting the release of HSP27 from cells, where it may function as an atheroprotective factor. Moreover, we show that although E2 may enact the release of HSP27 from cells, it has HSP27-independent effects that attenuate atherogenesis. Although it is well recognized that E2 induces the expression of HSP27, the current studies provide novel information outlining that its release from cells is modulated by E2.25,26 Use of an ERβ specific modulator to release HSP27 from cells may be equally effective in raising serum HSP27 levels, and more importantly does not produce the unwanted side effect of marked uterine hypertrophy. Future studies are now focusing on several key opportunities for the development of HSP27 as a potential agent to maintain vessel wall homeostasis. For example, new information from the current study demonstrating that HSP27 is present on the cell membrane and released via the exosomal pathway in response to stimulation by E2 and DPN may allow us to initiate nonhormonal strategies to promote the release of HSP27 from cells. Alternatively, emphasis is being placed on testing other specific ERβ modulators to promote HSP27 release and atheroprotection.20 Finally, the development of novel HSP27 mimics might allow for direct therapies that surpass the estrogenic stimulation required to promote its release from cells. By directly introducing HSP27 or truncated forms into the extracellular space to males and females, the interaction with SR-A and reduction of cholesterol uptake into the vessel wall may be targeted. Studies testing these innovative formulations of recombinant HSP27 are underway both pre- and post-induction of atherosclerosis.
We acknowledge the expertise of the Animal Care and Veterinary Services at the University of Ottawa Heart Institute.
Sources of Funding
This work was supported by operating grants (MOP80204 and IGO9448) from the Canadian Institutes of Health Research (CIHR) as well as the Heart and Stroke Foundation of Ontario (HSFO; T6335). KR was supported by a studentship jointly funded by the CIHR Institute of Gender and Health and the Ontario Women’s Health Council. JS was the recipient of a HSFO McGloin fellowship and TS received a HSFO Summer Medical studentship. EOB holds a Research Chair (URC #57093) that is jointly funded by CIHR and Medtronic.
Received January 27, 2009; revision accepted August 24, 2009.
Miller H, Poon S, Hibbert B, Rayner K, Chen Y-X, O'Brien ER. Modulation of estrogen signaling by the novel interaction of heat shock protein 27, a biomarker for atherosclerosis, and estrogen receptor β. Arterioscler Thromb Vasc Biol. 2005; 25: e10–e14.
Rayner K, Chen YX, McNulty M, Simard T, Zhao X, Wells DJ, de Belleroche J, O'Brien ER. Extracellular release of the atheroprotective heat shock protein 27 Is mediated by estrogen and competitively inhibits acLDL binding to scavenger receptor-A. Circ Res. 2008; 103: 133–141.
Gupta S, Knowlton AA. HSP60 trafficking in adult cardiac myocytes: role of the exosomal pathway. Am J Physiol Heart Circ Physiol. 2007; 292: H3052–H3056.
Savina A, Furlan M, Vidal M, Colombo MI. Exosome Release Is Regulated by a Calcium-dependent Mechanism in K562 Cells. J Biol Chem. 2003; 278: 20083–20090.
Xing D, Nozell S, Chen YF, Hage F, Oparil S. Estrogen and mechanisms of vascular protection. Arterioscler Thromb Vasc Biol. 2009; 29: 289–295.
Sader MA, Celermajer DS. Endothelial function, vascular reactivity and gender differences in the cardiovascular system. Cardiovasc Res. 2002; 53: 597–604.
Billon-Gales A, Fontaine C, Filipe C, Douin-Echinard V, Fouque MJ, Flouriot G, Gourdy P, Lenfant F, Laurell H, Krust A, Chambon P, Arnal JF. The transactivating function 1 of estrogen receptor alpha is dispensable for the vasculoprotective actions of 17beta-estradiol. Proc Natl Acad Sci U S A. 2009; 106: 2053–2058.
Pare G, Krust A, Karas RH, Dupont S, Aronovitz M, Chambon P, Mendelsohn ME. Estrogen receptor-alpha mediates the protective effects of estrogen against vascular injury. Circ Res. 2002; 90: 1087–1092.
Villablanca A, Lubahn D, Shelby L, Lloyd K, Barthold S. Susceptibility to early atherosclerosis in male mice is mediated by estrogen receptor alpha. Arterioscler Thromb Vasc Biol. 2004; 24: 1055–1061.
Martin-Ventura JL, Duran MC, Blanco-Colio LM, Meilhac O, Leclercq A, Michel JB, Jensen ON, Hernandez-Merida S, Tunon J, Vivanco F, Egido J. Identification by a differential proteomic approach of heat shock protein 27 as a potential marker of atherosclerosis. Circulation. 2004; 110: 2216–2219.
Park HK, Park EC, Bae SW, Park MY, Kim SW, Yoo HS, Tudev M, Ko YH, Choi YH, Kim S, Kim DI, Kim YW, Lee BB, Yoon JB, Park JE. Expression of heat shock protein 27 in human atherosclerotic plaques and increased plasma level of heat shock protein 27 in patients with acute coronary syndrome. Circulation. 2006; 114: 886–893.
Harrington WR, Sheng S, Barnett DH, Petz LN, Katzenellenbogen JA, Katzenellenbogen BS. Activities of estrogen receptor alpha- and beta-selective ligands at diverse estrogen responsive gene sites mediating transactivation or transrepression. Mol Cell Endocrinol. 2003; 206: 13–22.
Lancaster GI, Febbraio MA. Exosome-dependent trafficking of HSP70: a novel secretory pathway for cellular stress proteins. J Biol Chem. 2005; 280: 23349–23355.
De AK, Kodys K, Yeh BS, Miller-Graziano C. Exaggerated human monocyte IL-10 concomitant to minimal TNF-alpha induction by heat-shock protein 27 (Hsp27) suggests Hsp 27 is primarily an antiinflamatory stimulus. J Immunol. 2000; 165: 3951–3958.