Introduction

The contact between the blood of the patient and the artificial surface of the dialyzer leads to the activation of both cellular and plasmatic components [1], causing disturbances in lipid metabolism, oxidative stress, complement activity, and apoptosis [2, 3]. These hemodialysis (HD)-related complications may, at least in part, contribute to accelerated atherosclerosis in HD patients [2, 3]. It has also been suggested that heat shock proteins (HSP) may play a role in atherogenesis, but there is currently a lack of data on their function in patients with end stage renal disease.

Heat shock proteins are distributed in all prokaryotes and cellular compartments of eukaryotic organisms [4]. Their structure, preserved throughout evolution, is characterized by a high homogeneity among species [4]. They act as chaperones, preventing protein degradation in stress conditions, such as heat shock, toxicity, hypoxia, or inflammation [5]. Cell damage and necrosis cause the release of HSPs into the circulation, where they play a protective role. Their increased concentrations in the blood indicates the loss of cellular integrity and may induce immune tolerance [6–8].

The role of HSPs in atherogenesis is only partially understood [9]. According to the danger theory, the circulating HSPs function merely as markers of endothelial damage, released upon stimulation or stress factors [6, 7]. Alternatively, the hypothesis of molecular mimicry states that the highly conserved HSP structure may lead to cross-reactivity between, for example, bacterial and human HSPs, as well as auto-reactivity against its own antigens [10, 11], thus triggering the autoimmune disturbances observed in atherosclerosis [9, 11]. On the other hand, immunization with HSPs activates regulatory T cells that become able to cross-recognize own HSPs, thus inducing immune tolerance [8].

At the present time there is a lack of data on the circulating HSPs and anti-HSPs in patients with chronic kidney disease—both on those treated conservatively and on those already on dialysis. Although accelerated atherosclerosis has been detected in children with end stage renal disease, there are as yet no data on HSP engagement in that process. We have evaluated the levels of soluble (s) Hsp60, Hsp90alpha, anti-Hsp60, and anti-Hsp70, already described in patients with cardiovascular disease and hypertension, in the pediatric population in order to determine whether hemodialysis has an impact on HSPs and, if so, what kind of influence is that. Our aim was to assess the concentrations of selected HSPs and antibodies against them in the serum samples from the children on hemodialysis and to analyze whether they could be considered markers of atherosclerosis and if there is any relationship between those parameters and the known risk factors for atherosclerosis, such as lipid disturbances, endothelial activation, and inflammation.

Patients and methods

The patient cohort studied consisted of 17 patients (nine girls, eight boys; median age 14 years, interquartile range 12.5–17.5 years) on maintenance HD with polysulfone (PS) membranes (median time of therapy 2.1 years, interquartile range 1.4–2.5 years). Among these patients, the causative factors for chronic renal failure were: chronic glomerulonephritis (six patients), chronic pyelonephritis (six patients), and urinary tract malformations (five patients). The control group consisted of 15 patients (seven girls, eight boys; median age 14.5 years, interquartile range 12.5–16 years) diagnosed for urinary tract abnormalities or urolithiasis, with normal kidney function. The HD sessions (3.5–4 h) were performed three times weekly using bicarbonate dialysate; the blood flow ranged from 200 to 500 ml/min. The membrane area was between 1.0 and 1.6 m2, and the dialyzers were not reused. The water, purified by re-osmosis, was regularly checked for contamination. All patients were on a stable anticoagulation regimen using low-molecular-weight heparin. None of the patients showed clinical evidence of infection, had malignancies, or took antibiotics, corticosteroids, or immunosuppressive therapy. Informed consent was obtained from the subjects and their parents, if necessary. The research project was approved by the University Ethics Committee, according to the Helsinki declaration.

Blood samples were drawn from the afferent line of the first-use dialyzer prior to starting an HD session and at the end of the session. For the control subjects, blood was drawn from the peripheral vein. The samples were centrifuged at 4°C for 10 min, and the serum samples were stored at −20°C until assayed. The serum concentrations of Hsp60, Hsp90alpha, anti-Hsp60, anti-Hsp70 and E-selectin were evaluated by commercially available enzyme-linked immunosorbent assay (ELISA) kits (Stressgen; R&D Systems, Abingdon, UK). Each sample was measured in duplicate, and the arithmetic mean was considered to be the final result. Standards and serum samples were transferred to 96 well micro-plates pre-coated with either the mouse monoclonal antibodies to human Hsp60, Hsp90alpha, or E-selectin or those to recombinant human Hsp60 and Hsp70. In the case of HSPs, the captured proteins were first incubated with the Hsp60 or Hsp90alpha antibody conjugated to horseradish peroxidase and then with the appropriate substrate (tetramethylbenzidine). When anti-HSPs were evaluated, the captured antibodies against Hsp60 or Hsp70 were incubated with a polyclonal goat antibody specific for human immunoglobulin (Ig)G, IgA, and IgM, conjugated with hydrogen peroxidase. The assay was then developed with tetramethylbenzidine. Finally, the reaction was stopped with the acid solution and the absorbance measured at 450 nm with the correction wavelength set at 620 nm. The measurements were performed according to the manufacturer’s instructions, and the results were calculated by reference to the standard curves. Limits of detection for the parameters were as follows: Hsp60, 3.1 ng/ml; Hsp90alpha, 0.06 ng/ml; anti-Hsp60, 2.8 ng/ml; anti-Hsp70, 6.8 ng/ml; sE-selectin, 0.1 ng/ml.

The lipid profile [total cholesterol, high-density lipoprotein (HDL)-cholesterol (CHOL), low-density lipoprotein (LDL)-CHOL, triglycerides] was assessed in all patients (Applied BioSystems, Foster City, CA). A marker of inflammation, high-sensitivity C-reactive protein (hs-CRP), was assessed by nephelometry (Dade Behring, Milton Keynes, UK).

Statistical analysis

The results are expressed as median values and interquartile ranges. The differences between all groups as well as between the HSP values during the HD sessions were evaluated using nonparametric tests (Mann–Whitney U test; Wilcoxon). The relations between the parameters were assessed by simple regression analysis. The statistical analysis was performed using the package Statistica ver. 7.1 (StatSoft, Tulsa, OK). A p value < 0.05 was considered to be significant.

Results

The median values and the interquartile ranges of the examined parameters are shown in Table 1, and the statistically significant results of a simple regression analysis are presented in Table 2.

Table 1 The median values and interquartile ranges of the examined parameter concentrations in children and young adults on hemodialysis before and after a single HD session and in the control subjects
Full size table
Table 2 Statistically significant associations between the examined parameters assessed by a simple regression analysis before and after HD session with a polysulfone membrane
Full size table

Before the HD session

The median values of Hsp60 before the HD session were lower than those in the controls, whereas the median values of Hsp90alpha, anti-Hsp60, and sE-selectin were higher in the HD patients than in the control group (Table 1). No such difference was seen when the anti-Hsp70 and hsCRP concentrations were analyzed.

After the HD session

A single HD session on a polysulfone dialyzer had a statistically significant influence on all analyzed HSPs: the median values of Hsp60 and Hsp90alpha increased (p < 0.001; p < 0.05, respectively) and the concentrations of both anti-Hsp60 and anti-Hsp70 decreased (p < 0.01; p< 0.001, respectively). Despite these changes, the Hsp60 concentrations after the HD session remained lower relative to the the controls (the potentially confusing discrepancy between the median values, seen in the Table 1, results from the rank sums being compared, instead of the median values, in the nonparametric tests). In contrast, the values for Hsp90alpha and anti-Hsp60 after HD were still higher relative to the values for the control group. The only observed difference was for anti-Hsp70, the levels of which became lower than those in the controls after the HD session.

The HD session had no impact on hsCRP and sE-selectin concentrations.

Discussion

Our study reports on the profile of serum HSPs in children and young adults on chronic HD. We have shown for the first time that chronic HD, as well as a single HD session, may modify the levels of selected HSPs and antibodies against these in that population.

The concentrations of Hsp60 in the patient group were lower than those of the control group both before and after HD. Although there are no data on the circulating Hsp60 level in patients on HD, such a decrease suggests a dysfunction of this protein in that group and an impaired protection against stressing factors. This decreased level may also have an impact on the enhanced apoptosis observed in HD patients [12, 13]. In vitro investigations have shown that diminished cytosolic Hsp60 concentrations alone can precipitate apoptosis by activating the proapoptotic proteins Bax and Bak and by suppressing the antiapoptotic Bcl-2 [14]. In addition, the correlations between Hsp60 and the lipid profile may suggest that the former has a possible role in atherogenesis, although the observation that these associations appear only after HD weakens the significance of these results. As the levels of Hsp60 are known to be higher in adults and elderly of the general population, a comparison with pediatric patients on HD using data from these age groups seems risky [15]. On the other hand, Xu et al. [15] have reported a correlation between Hsp60 levels and the thickness of the intima-media, which may imply that the former has a role in atherosclerosis.

We found that a single HD session significantly raised the concentration of Hsp60. However, the reasons for such an increase may be multiple. A number of in vitro and in vivo investigations have confirmed that overexpression of Hsp60 on the endothelial cells can be provoked by shear stress [16], and elevated Hsp70 expression on lymphocytes collected from adults after a HD session on the polysulfone membranes has been reported [12]. Extracellular HSPs appear in the circulation in the case of stress [17], indicating the loss of cellular integrity, and a HD session, even with a biocompatible membrane, is no doubt a stressful condition. Moreover, according to Grooteman et al. [18], during the HD session, circulating cells are more prone to damage than endothelial ones. Therefore, the post-dialytical increase in Hsp60 is a possible consequence of the interaction between blood cells and the dialyzer membrane. Indeed, any cell undergoing stress during the HD session, including erythrocytes and degranulated leukocytes, may be the source of Hsp60, which makes the exact origin of the circulating Hsp60 difficult to specify. The actual degree of increase can only be diminished by the impact of the post-dialysis decrease in plasma volume; therefore, this potentially confounding variable should be taken into account.

The first clinical implication of that increase may be the overproduction of proinflammatory cytokines and reactive oxygen species (ROS) by macrophages and monocytes during HD [19–21]. Secondly, the increase in the Hsp60 concentration may act as a repetitive immunization, triggering the activation of regulatory T cells and inducing immune tolerance [8]. Such a hypothesis could, at least partially, explain the phenomenon of the patients on chronic hemodialysis, in whom immune overactivity becomes silent once they have started the HD therapy.

Contrary to the situation with Hsp60, concentrations of antibodies against Hsp60 increased relative to the controls both before and after HDs. There are no data available on human anti-Hsp60 in patients on HD. Although increased titers of antibodies against mycobacterial Hsp65 have been observed in HD subjects with vascular events [22], it has been shown that the reactions against bacterial and human HSPs are not similar [23]; consequently, the interpretation of the role of the cross-reactive responses against HSPs may be difficult [24]. Nevertheless, high levels of anti-Hsp60 have been found in patients with coronary heart disease and atherosclerosis [25]. Ghayour-Mobarhan et al. [26] observed increased anti-Hsp60 and anti-Hsp70 titers in dyslipidemic patients, with or without coronary heart disease. The regression analysis carried out in our investigation also revealed several correlations between anti-Hsp60 and the lipid profile before and after HD. Taken together, the high antibody titers observed in our HD children may suggest a state of autoimmune overactivity triggered by unfavorable conditions, such as lipid disorders, endothelial dysfunction, and increased cardiovascular risk, all of which are common in that population [27].

The significance of the anti-Hsp60 and anti-Hsp70 concentrations, which fall after the HD session, as well as the reason for that process, remain unknown. We can only speculate whether this could be a consequence of a simple adsorption on the dialyzer surface or of an antigen-antibody reaction, such as the creation of HSP–anti-HSP complexes [28, 29]. Indeed, such complexes are able to activate the complement, which is in accordance with the phenomenon observed in the course of a HD session [30, 31]. The high anti-Hsp60 and anti-Hsp70 titers may also suppress Hsp60 concentrations, which was confirmed by the inverse correlation in our investigation and, probably, act in a detrimental manner.

From this point of view, every HD session would be a period during which this proportion is ameliorated by the transient decrease in anti-HSP levels and the increase in HSP concentrations. Given this situation, reduced antibody concentrations would strengthen the protection against autoimmunity, probably already triggered by the increase in HSP concentration. On the whole, a HD session may have a beneficial, albeit transient, effect on HSP–anti-HSP relations. Whether these changes may influence the overall immune status of the patient remains unknown, since other elements, such as the complement or cytokine activity, have not been analyzed in our study.

The concentration of Hsp90alpha remained increased both before and after HD. The elevation in its concentration during the session may support the previous assumption of its protective role and is concordant with observations suggesting a strict connection between Hsp90alpha and oxidative stress. Experimental data have shown that oxidative stress stimulates vascular smooth muscle cells to produce Hsp90alpha [32]. Additionally, the inverse correlation between Hsp90alpha and cholesterol, observed before HD, may confirm the protective role of that HSP in atherosclerosis.

The increased concentrations of soluble E-selectin, a molecule released only upon stimulation, both before and after HD, confirms the endothelial dysfunction characteristic for HD [3, 33]. The fact that a single polysulfone HD had no impact on the sE-selectin levels is in agreement with our previous observations on sVCAM-1, which is also generated after endothelial stimulation [34]. One possible explanation may be the hypothesis of Grooteman et al. [18], according to which a single HD session affects the circulating cells rather than the endothelial ones. However, the correlations between sE-selectin, Hsp60, and anti-Hsp70 seem to advocate the hypothesis, already suggested by the association between HSPs and the lipid profile, that they may be considered markers of atherosclerosis.

There were a number of limitations to our study. Due to the fact that there are no available data on HSPs in children and patients on HD, the conclusions based on our results, in particular those concerning the reason for their modified concentrations, have to be assessed carefully. Additionally, the small number of patients in our study, conditioned by the size of the pediatric population on HD, also requires cautious reasoning. Although several parameters were assessed, we did not manage to take into account all of the potential factors influencing the immune system, so our analysis of the impact of HSP on the immune status of the hemodialyzed patients needs further in-depth studies.

Conclusions

We have investigated the dysfunctional profile of HSPs in the population of children and young adults on chronic HD. Based on our results, we suggest the potential usefulness of these proteins as markers of atherosclerosis in the course of chronic kidney disease. This is the first report of such a study in a pediatric HD population. Whether the increase in HSP levels and the decrease in antibody concentrations during HD is a transient favorable event and/or may have a long-term beneficial effect on the immune system is still a speculative idea that needs further investigation on a larger group of patients.