Dynamics of heat shock protein 60 in endothelial cells exposed to cigarette smoke extract.

Heat shock protein 60 (HSP60), expressed on the surface of endothelial cells (ECs) stressed by e.g. oxidized LDL or mechanical shear, was shown to function as an auto-antigen and thus as a pro-atherosclerotic molecule. The aim of this study was to determine whether cigarette smoke chemicals can lead to the activation of the "HSP60 pathway." It was also our aim to elucidate the dynamics of HSP60 from gene expression to endothelial surface expression and secretion. Here we show for the first time that the exposure of human umbilical vein endothelial cells (HUVECs) to cigarette smoke extract (CSE) results in an up-regulation of HSP60 mRNA. Live cell imaging analysis of a HSP60-EYFP fusion protein construct transfected into ECs revealed that mitochondrial structures collapse in response to CSE exposure. As a result, HSP60 is released from the mitochondria, transported to the cell surface, and released into the cell culture supernatant. Analysis of HSP60 in the sera of healthy young individuals exposed to secondhand smoke revealed significantly elevated levels of HSP60. Cigarette smoking is one of the most relevant risk factors for atherosclerosis. Herein, we provide evidence that cigarette smoke may initiate atherosclerosis in the sense of the "auto-immune hypothesis of atherosclerosis."

Introduction

It is generally accepted that atherosclerosis is an inflammatory disease of the arterial vessel wall [1,2]. Our previous observations suggest that inflammation is also an essential factor in the incipient stages of atherosclerosis [2]. ECs exposed to stress caused by classical atherosclerosis risk factors show surface expression of heat shock protein 60 (HSP60). Infections during childhood resulting in the formation of anti-microbial heat shock protein antibodies, and a cross-reactivity of these antibodies to human heat shock protein 60 (due to high sequence homologies) are the basis for an auto-immune reaction against ECs. It has been shown that atherosclerosis risk factors such as oxidized low-density lipoproteins and shear stress lead to the translocation of HSP60 to the EC surface, where it serves as an auto-antigen and signalling molecule [3-5]. Anti-HSP60 antibody-mediated cellular cytotoxicity has been observed, and in several human studies, anti-HSP60 antibody serum titres were correlated with atherosclerosis and other autoimmune diseases [2].

In a study of healthy young adults, cigarette smoking was found to be the most relevant risk factor for early atherosclerosis, indicating the impact of this risk factor in atherosclerosis initiation [6]. Experimental microarray-based studies demonstrated that cigarette smoke extract (CSE) provokes a massive heat shock response in ECs at the transcriptome level [7]. The effects of cigarette smoke chemicals on HSP60 at the cellular level (particularly the dynamics and surface expression) as well as the dependence of serum HSP60 levels on the smoking status in young individuals have not been studied to date. Based on previous results on other risk factors for early atherosclerosis, this study was designed to address the hypothesis that the risk factor smoking, too, may induce in stressed ECs an HSP60 response ranging from activation of transcription to the translocation of HSP60 from the mitochondria to the cytosol, and from there to the EC surface. It has previously been reported that soluble HSP60 activates macrophages [8]. In this study, we wanted to determine if soluble HSP60 also binds to macrophages. To ascertain the in vivo relevance of this investigation, we examined not only the EC culture supernatants but also serum samples from healthy young smokers, passive smokers (secondhand smokers) and non-smokers, for the presence of soluble HSP60.

Materials and methods

For details on Materials and methods, see the Online Supplement.

Results

HSP60 mRNA expression and dynamics of HSP60 protein in endothelial cells exposed to CSE

First, the effect of CSE on endothelial HSP60 mRNA expression was analyzed after different incubation times. Fig. 1A shows that an up-regulation of hspd1 mRNA occurs as early as 3 h after exposure to 8% CSE and peaks at 7 h. The transient character of this upregulation is indicated by the drop in the levels of HSP60 mRNA after 24 h.

Fig. 1

CSE induces the upregulation of HSP60 mRNA, causes HSP60 release from mitochondria, and leads to HSP60 surface expression. A. Real-time PCR analysis of hspd1 mRNA in response to exposure of HUVECs to 8% CSE over different time periods. Images B to E show the effects of treating HUVECs with 16% CSE over different time periods on HSP60-EYFP localisation. White arrows indicate contracted mitochondria; orange arrows indicate cells where HSP60-EYFP was released from mitochondria. Images F, H, and J show control HUVECs; images G, I, and K show HUVECs treated with 8% CSE for 24 hours. Images F, G: Phase contrast analysis (PhC); images H, I: HSP60 immunofluorescence analysis; images J, and I: overlay of PhC and IF. Black arrows indicate a plasma membrane-associated localisation of HSP60. All experiments were repeated three times. Shown are either representative images or mean values of a representative experiment performed in duplicates ±SD. Asterisks indicate significant differences (determined by ANOVA; Posthoc Bonferroni) of the groups compared to untreated control cells. (*p < 0.05; **p < 0.01; ***p < 0.001).

To define the dynamics of HSP60 in ECs exposed to CSE, live cell imaging analyses were performed. Figs. 1B–E and the video file (Online Supplement) revealed that CSE leads to significant changes in mitochondrial morphology and alters cellular HSP60 localization (CSE was added at time point 0, live cell imaging was conducted for 180 min). The major findings were (i) disruption of the filamentous phenotype of mitochondria (for comparison also see Fig. 1H) 60 min after addition of CSE (indicated by a white arrow in Fig. 1C), and (ii) partial disintegration of mitochondria, leading to HSP60 release from mitochondria and its cytosolic appearance (orange arrows in Fig. 1E). In order to test whether, in addition to causing a release of HSP60 from the mitochondria into the cytosol, CSE also causes translocation of HSP60 to the cell surface, ECs were incubated with 8% CSE for 24 h, followed by intra-cellular and membrane immuno-staining for HSP60. Figs. 1F–K show that, in contrast to the intracellular filamentous staining pattern in the controls, CSE-treated cells show a membrane-associated distribution of HSP60 (black arrows indicate surface HSP60).

HSP60 is released into the culture supernatant in response to CSE treatment of EC

To address the question whether surface HSP60 is released into the culture supernatant, an ELISA-based quantification of soluble HSP60 in the supernatant of ECs treated with different concentrations of CSE and controls was conducted at time points 0, 3, 6, 12, 24, and 48 h. As early as 3 h after treatment with 32% CSE, HSP60 was detected in the supernatant (Fig. 2A). We have previously shown that treating ECs with 32% CSE leads to massive membrane damage [9]. Thus, it can be hypothesized that membrane damage is the cause of rapid release of HSP60 into the supernatant as observed in the present study. However, in the previous study, we also found that the breakdown of plasma membrane integrity of cells exposed to 16% CSE did not occur before 16 h of exposure, and 8% CSE was not toxic to ECs throughout the entire exposure time [9]. In the present study, however, HSP60 was released even before 16 h of treatment with 16% CSE, with no damage to membranes, and also after treatment with 8% CSE. Thus, it appears that HSP60 can be released even from cells with apparently intact membranes. The data presented in Fig. 2A suggest that there has to be an as yet unknown mechanism that triggers HSP60 release from ECs.

Fig. 2

Cigarette smoke chemicals increase the level of soluble HSP60 in endothelial cell culture supernatants and human serum. Image A shows the level of soluble HSP60 in the supernatant of HUVECs incubated with CSE concentrations for the times indicated. The experiment was repeated three times. Shown are mean values of a representative experiment performed in quadruplicates + SD. Asterisks indicate significant differences of the groups compared to the control. Data were analyzed using ANOVA and Posthoc Bonferroni. *p < 0.05; **p < 0.01; ***p < 0.001. The box blot in B shows the results of a study of 56 healthy volunteers (non-smokers, passive smokers and smokers). The smoking status was correlated with serum levels of soluble HSP60. Differences between the groups were calculated by ANOVA (Post-hoc Bonferroni). n.s—not significant.

Increase in serum HSP60 levels after exposure to secondhand smoke

In order to test whether the above in vitro data can be extrapolated to the in vivo situation, we analyzed the serum HSP60 levels in 56 young healthy volunteers, comprising active smokers, passive smokers, and non-smokers. As can be seen in Fig. 2B, only secondhand smoking leads to a significant increase in serum HSP60 levels compared to the control.

Anti-oxidants do not attenuate CSE-caused HSP60 release from HUVECs

To investigate whether oxidants or radicals present in cigarette smoke may be responsible for stress-mediated release of HSP60, ECs were pre-incubated with vitamins C and E, as well as will N-acetyl cysteine (NAC). None of these anti-oxidants reduced the extent of CSE-mediated HSP60 release from ECs (see Online Supplement Fig. 4). These in vitro findings are supported by the in vivo observations that dietary intake of antioxidants (fruits, vegetables) did not differ between the groups, and had no impact on serum HSP60 levels in another human study (see Online Supplement Fig. 6), suggesting that, anti-oxidant intake has no influence on the levels of serum HSP60, which are increased by secondhand smoke.

Discussion

Based on the “autoimmune hypothesis of atherosclerosis” [2], in this study we analyzed the effect of the hydrophilic fraction of cigarette smoke on HSP60 expression and dynamics of cellular localization in ECs. CSE treatment of ECs led to a transcriptional up-regulation of the nuclear encoded hspd1 gene (HSP60) as early as 3 h after CSE exposure (Fig. 1A). The transient wave of hspd1 mRNA levels (upregulation after 3 and 7 h and the drop after 24 h) suggests a tight regulation of the gene, as has been reported in previous studies [3,10]. The regulatory pattern may help restore mitochondrial HSP60 pools after a stress-induced HSP60 release.

In our fusion protein-based analyses of HSP60 dynamics in ECs, we found that HSP60 release from mitochondria started with a condensation of scattered mitochondria already 60 min after exposure to 8% CSE. This phenomenon increased in intensity over time and culminated in the total condensation of the mitochondrial mass, loss of mitochondrial membrane integrity, and release of HSP60 into the cytosol (indicated by the cytosolic distribution of the EYFP signal). These data show for the first time in (endothelial) cells exposed to cigarette smoke constituents a time-resolved analysis of HSP60 dynamics and release of HSP60 from mitochondria. Further, phase contrast microscopic analyses of HSP60 in ECs treated with 8% CSE revealed a membrane-associated distribution after 24 h (surface expression), and release of HSP60 into the supernatant was demonstrated by ELISA. Since incubation with 8% CSE left the plasma membrane intact and thus was not toxic to endothelial cells [11], it is suggested that HSP60 is actively transferred to the outer layer of the plasma membrane by an unknown mechanism, and then released (passively or actively, for instance, by shedding) into the extracellular space. Both, extracellular membrane-associated and soluble HSP60 may contribute to local and systemic immune reactions [2]. Interestingly, it could be shown that soluble HSP60 binds to macrophages (Online Supplement Fig. 3). Habich et al. [8] have previously demonstrated that HSP60 binds to TOLL-like receptor 4 and cause its activation. Those results further emphasize the role of HSP60 as an immunological danger signal.

We also compared the effects of CSE on HSP60 expression, surface expression, and release with those of other previously analyzed atherogenic risk factors such as oxidized low density lipoprotein, advanced glycation end products (unpublished data) and shear stress [3,10]. Even though 8% CSE reflects only a relatively low and physiologically relevant exposure of cells to cigarette smoke constituents [9], this concentration of CSE exerted the most pronounced effect of HSP60 release compared to other endothelial “stress-factors” that has been observed by our group so far. This appears to be in line with previous results on healthy young adults [6] showing that cigarette smoking was by far the most relevant risk factor for early atherosclerosis (increased intima-media thickness).

Cigarette smoke is a complex mixture of over 4000 different chemicals. As oxidants and radicals play an important role in several smoking-associated cardiovascular pathophysiologies, the use of antioxidants to inhibit their negative influence needs to be discussed [12]. We examined the effect of the antioxidants vitamin C, vitamin E, and NAC on CSE-induced HSP60 release from ECs. None of the above antioxidants was capable of significantly reducing CSE-mediated HSP60 release (see Online Supplement Fig. 4), suggesting that oxidants present in cigarette smoke are not involved in this process. Further, as we were previously able to show that NAC is a potent inhibitor of the adverse effects of metals in cigarette smoke, also metals can be excluded as essential players in HSP60 release. As CSE contains mainly hydrophilic cigarette smoke chemicals, also the fraction of hydrophobic compounds (tar) can be excluded as active principle [13]. In summary, the identity of the HSP60 release-relevant chemical(s) in CSE remains unknown; however, several important and well-known chemicals in cigarette smoke could be excluded.

In order to find out if our in vitro data can be extrapolated to the in vivo situation, we analyzed the serum HSP60 levels in non-smokers, as well as active and passive smokers. Total values of HSP60 protein in the serum of healthy study subjects were found to be very stable, an observation that has also been reported by others [14]. Our analyses revealed that only those breathing in secondhand smoke showed significantly increased serum HSP60 levels when compared to the control group. The study groups were analyzed in an age-matched manner, and other confounding factors such as inflammatory diseases, other acute and chronic diseases, medication, and fruit and vegetable consumption were excluded. In addition, the above observation was verified by determining the soluble serum HSP60 levels in samples of the ARFY study [15], which is also a study of healthy young adults. Those analyses yielded essentially the same results, although the differences did not reach significance (see Online Supplement Fig. 5). Median serum HSP60 levels of non-smokers were 2.5, passive smokers 4.5, and active smokers 3.75 ng/ml (non-smokers, active-, and passive smokers were classified based on the criteria described in Materials and methods). The finding of significantly increased serum HSP60 levels only of passive smokers was unexpected. We would like to speculate that young active smokers, due to regular exposure, may be able to deal with the increased atherogenic burden caused by smoking by upregulation of defence mechanisms due to regular exposure, whereas, due to the occasional character of exposure, the organism of passive smokers is unable to adapt to this noxa, and therefore reacts more severely. If this is true, then exposure to secondhand smoke constitutes a particularly significant risk for early atherosclerosis at a young age.

Funding

This work was supported by the Medizinischer Forschungsfonds Innsbruck (MFI-Project #4302 to DB), the Austrian National Bank (Project #12697 to DB), the Austrian Research Found (FWF grant #14741 to GW) and by the European Initiative to Fight Chlamydial Infection by Unbiased Genomics (ECIBUG; #818496 to GW) from the ERA Network of Pathogenomics.

Disclosures

None.