- Email: [email protected]
Low preoperative HSP70 atrial myocardial levels correlate significantly with high incidence of postoperative atrial fibrillation after cardiac surgery
Cardiovascular Surgery, Vol. 10, No. 3, pp. 228–232, 2002 2002 The International Society for Cardiovascular Surgery Published by Elsevier Science Ltd. All rights reserved 0967-2109/02 $22.00
PII: S0967-2109(01)00138-7
www.elsevier.com/locate/cardiosur
Low preoperative HSP70 atrial myocardial levels correlate significantly with high incidence of postoperative atrial fibrillation after cardiac surgery Kyriakos St. Rammos*, George J. Koullias†, Moustafa O. Hassan‡, Nikolaos P. Argyrakis*, Christos G. Voucharas*, Steven J. Scarupa† and Tomas G. Cowte† *Cardiothoracic Surgery, AHEPA General Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece; †Pinnacle Health System Residency in General Surgery, Harrisburg, PA, USA and ‡Department of Surgery, University of Pittsburgh, Pittsburgh, PA, USA Purpose of the study. Atrial fibrillation after cardiac surgery is still a frequent encountered complication and has been associated with increased hospital length of stay and numerous postoperative complications. The pathogenesis of atrial fibrillation involves an overall sequence of perioperative events, collectively termed as ischemia-reperfusion injury. Heat-shock proteins have been found to provide increased protection during ischemia-reperfusion as well as increased postischemic cardiac functional recovery. We sought to determine whether preoperative atrial heat shock levels were correlated with the appearance of postoperative atrial fibrillation Basic methods. Preoperative atrial myocardial samples obtained just before cannulation from 101 patients were used to detect immunohistochemically the expression of heat-shock proteins. The derived results were compared statistically with the incidence of postoperative atrial fibrillation, its time of appearance, duration and resistance to administered antiarrhythmics. Principal findings. The overall incidence of postoperative atrial fibrillation was 22.3%. Of these patients, 58.3% had no detectable heat shock proteins in their cytoplasm, in sharp contrast with 100% of the patients with no atrial fibrillation who were positive for heat shock proteins (p⬍0.01). Four percent of our patient group had prolonged atrial fibrillation (defined as duration ⬎48 h). These patients had significantly less (p⬍0.01) nuclear heat shock protein expression compared with the non-atrial fibrillation group. However, the difference of the heat shock protein expression between the prolonged atrial fibrillation and the rest of the atrial fibrillation patients was not significant (p = 0.891). Conclusions. Our results indicate that patients with low preoperative atrial heat shock protein expression have a significantly greater incidence of postoperative atrial fibrillation. Heat shock protein expression did not, however, correlate with the onset of atrial fibrillation and the resistance to administered medications. Heat shock protein preoperative induction as a measure of myocardial preconditioning may potentially decrease the incidence of postoperative atrial fibrillation. 2002 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd. All rights reserved Keywords: HSP70 atrial myocardial levels, atrial fibrillation
Introduction Correspondence to Kyriakos St. Rammos, 29 Karolou Diehl Str, Thessaloniki 546 23, Greece. Tel./Fax: +30-31-222-021; e-mail: [email protected]
228
Atrial fibrillation (AFIB) is a common dysrhythmia following coronary artery bypass surgery (CABG) [1, 2]. Postoperative AFIB increases the duration of hospitalization and has in numerous studies been CARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al.
associated with an increased incidence in postoperative stroke, ventricular arrhythmias and the need for a permanent pacemaker [3]. Numerous efforts to determine which patients are likely to develop postoperative AFIB have identified multiple preoperative, intraoperative and even postoperative factors and have led to modifications in several intraoperative practices [4]. Nevertheless, the incidence of AFIB after CABG remains in the vicinity of 25% [4]. One of the recently identified endogenous mechanisms of preoperative and intraoperative myocardial protection is the presence and the possible induction of the heat shock protein (HSP70) family. HSP70 expression is stimulated by ischemia [5]. Furthermore, up-regulation of HSP70 in myocardial cells by adenovirus mediated gene transfer and HSP70 product overexpression in transgenic animals has been shown to greatly increase their resistance to ischemia and improve their post-ischemic recovery [6, 7]. HSP70 proteins function as molecular chaperones facilitating intramyocardial protein assembly, folding, transport and translocation as well as preventing denaturation of intramyocardial enzymes [8–10]. Resident macrophages within myocardial cells containing high HSP70 levels produce significantly lower quantities of TNF-alpha and provide increased protection during ischemia-reperfusion (I/R) as well as echocardiographically proven improved cardiac functional recovery [11]. Local reductions in atrial partial oxygen pressure during cardiopulmonary bypass and postoperative downregulation of atrial intracellular Ca++ pumps have been implicated in the appearance of postoperative AFIB and decreased intramyocardial quantities of HSP70 have been found in both the above biochemical instances [12, 13]. In view of the above, we have examined the expression of HSP70 in right atrial cells of patients undergoing open heart surgery. We have correlated these results with the incidence, the timing of appearance, the intensity, and the resistance to administered medications of postoperative AFIB.
Methods One hundred and one (n = 101) of the patients that underwent open heart surgery using cardiopulmonary bypass at the Department of Cardiothoracic Surgery at the AHEPA University Hospital and operated on by the senior author (KR) were included in the present study. The same cardiac anesthesiologist, perfusionist and operating room team participated in the performance of all these procedures. All patients underwent cardiopulmonary bypass with moderate hypothermia (33–34°C), antegrade crystalloid cardioplegic arrest (St Thomas I) and local hypothermia with ice-slash; cardioplegic solution was readministered every 20–30 min. EpidemiologCARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
ical data of these patients and the incidence of postoperative AFIB are presented in Table 1. The study was approved by the Hospital Ethical Committee and informed consent was obtained from all the above patients preoperatively. Tru-cut biopsy at the right atrial cannulation site as soon as the pericardium was opened yielded approximately 2–3 mm3 of atrial tissue. This tissue sample was immediately placed and fixed in a 10% neutral formalin solution and was subsequently embedded in paraffin. Two 5 µm thick parallel tissue sections from each patient were subjected to a three-step immunoperoxidase immunohistochemical staining technique for the detection and expression of the two major HSP70s using a rabbit polyclonal antibody which identifies HSP72 and HSP73 (DAKO, Califoria, USA, code no. A500). The above method was carried out using DAB as the final chromogen agent. Negative controls were prepared by replacing the primary antibody with normal swine serum and sections of previously known strongly positive tonsil squamous epithelium were used as positive controls. All immunohistochemical material was evaluated blindly by two histopathologists. The intensity of staining was graded semiquantitatively as ‘0’ when no staining was present, ‘1’ when staining was present in up to 10% of cells, ‘2’ when 10–50% of the cells were positive and ‘3’ when the majority (⬎50%) of cells were positive for HSP70. The above grading system was applied separately to the cytoplasm and the nucleus and therefore for each slide a ‘cytoplasmic’ and a ‘nuclear’ HSP70 score were obtained.
Statistical analysis Statistical analysis was performed with the use of the commercially available ‘Primer of Biostatistics’ biostatistical software by Stanton A. Glantz, PhD (McGraw-Hill Health Professions Division, 1997). Student’s t-test and Mann–Whitney Rank sum test were used to evaluate differences between cytoplasmic and/or nuclear HSP70 scores of patients with or without postoperative AFIB. On occasion, analysis of variance (ANOVA) and Spearman correlation was used to also detect and estimate correlations between different variables. The p⬍0.05 value was considered the level of statistically significant difference.
Results Patients with no postoperative AFIB Normal sinus rhythm or sinus tachycardia was observed in 77 out of 101 patients (77.7%). The average cytoplasmic HSP70 score was 2.692±0.057 and the average nuclear HSP70 score was 1.846±0.091. All (100%) of these patients had a 229
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al. Table 1 Preoperative and intraoperative patient data. No statistically significant differences wee observed between patients that had postoperative atrial fibrillation and the rest of the group Patient groups
AFIB (n = 24)
Non-AFIB (n = 77)
Statistical significance
Age Male Female No. of grafts Valve replacements MVR AVR Combined valve and CABG NYHA class IDDM HTN Preop EF Smoking Aortic Cx time CPB time Degree of hypothermia
18 6 3.0 3/24 2 1 1/24 III+0.1 8/24 19/24 42.1% 19/24 58.7±1.48 103.84 33.2±0.2°C
50 27 2.55 10/77 4 6 6/77 III+0.39 15/77 35/77 53.08% 68/77 69.6±0.94 126.15 33.7±19°C
NS NS NS NS NS NS NS NS NS NS NS NS P⬍0.08 NS NS
moderate to intense cytoplasmic HSP70 staining (Figure 1). HSP70 appeared to be located mainly in the cytoplasm, the nucleus and the nucleoli of atrial myocardial cells. Endothelial cells, adjacent macrophages and lymphocytes emerged exempt of HSP70. Patients with postoperative AFIB Atrial fibrillation (AFIB) was seen in 24 out of 101 patients (22.3%) (Table 2). The average cytoplasmic HSP70 score in these patients was 1.857±0.169 and the average nuclear score was 0.714±0.176. HSP70 seem to spare predominantly the nucleus in AFIB patients (Figure 2). In 4/24 of these patients (16.6%) HSP70 was not detected in their cytoplasm and in 14/24 (58.3%) of this group HSP70 was not Figure 2 Histological section of a right atrial biopsy in a patient with postoperative atrial fibrillation. Nuclei are spared from staining and a mild granular HSP70 presence is observed in the cytoplasm (X40)
detected in the nucleus (Figure 3). In both of the above scores (cytoplasmic and nuclear) the difference of the HSP70 scores between the patients with and without postoperative AFIB was statistically sigTable 2 Medication and medication combinations used for atrial fibrillation treatment in our study Medications for atrial fibrillation in 24 patients
Figure 1 Representative immunohistochemical microscopy of right atrial section of a patient with no atrial fibrillation after CABG. Nuclei and nucleoli are intensely stained. The cytoplasm shows intense granular HSP70 staining (X40)
230
Digoxin (1) alone Amiodarone (2) alone Diltiazem (3) alone Propranolol (4) alone (1) and (2) (1) and (3)
CARDIOVASCULAR SURGERY
1 2 1 – 18 9
JUNE 2002 VOL 10 NO 3
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al.
meters and the patient’s HSP70 nuclear or cytoplasmic scores.
Discussion
Figure 3 Right atrial microscopy of a patient with postoperative atrial fibrillation. No HSP70 was detected on either the nuclei or the cytoplasm (X40)
nificant (p⬍0.001 — 98 degrees of freedom, t = 7.746 and ⫺4.944, respectively) (Figure 4). Patients with prolonged (⬎48 h) AFIB Prolonged — as defined above — atrial fibrillation was encountered in 4/24 AFIB patients (16.6%). Statistical comparison of these four patients with the rest of the AFIB group revealed that patients with prolonged AFIB had a significantly lower (p⬍0.01) nuclear HSP70 expression. However, the difference in the cytoplasmic HSP70 expression between these four patients and the rest of the AFIB group was not significant (p⬍0.891). Medications for atrial fibrillation Table 2 shows the incidence, kind and number of medications used for the treatment of atrial fibrillation in the AFIB group. No statistically significant associations were observed between these para-
Figure 4 Graphic presentation of cytoplasmic (white) and nuclear (black) HSP70 staining scores between non-AFIB, AFIB and prolonged (⬎48 h) AFIB patients
CARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
Several studies in the literature have pointed to the cytoprotective role of HSP70 from ischemia and reperfusion myocardial injury after actual or experimental ischemia or open heart surgery [8–10, 14, 15]. The HSP70 family of proteins contains at least four members; HSP72, HSC73, BiP and GRP 76. From these, myocardial cells contain for certain the first two; HSP72 is inducible following myocardial injury. HSC73 appears to be constitutively expressed and its levels seem to be dependent on previous injury, use of medications or are genetically determined. Our study is the first report that demonstrated a repeated low nuclear and cytoplasmic atrial myocardial expression of HSP72/HSC73 in preoperative specimens of patients that had AFIB after open heart surgery. Low HSP70 levels were significantly correlated with the occurrence of postoperative AFIB. HSP70 presence, however, did not correlate with the time of AFIB onset, nor the duration or the resistance of AFIB reversal to different administered medications. Although our study patient sample (n = 101) is of adequate size to value any statistically significant results, two issues in our methodology deserve further elucidation. First the performance of immunohistochemistry. Immunohistochemical techniques do not allow for accurate quantitative measurements, thus limiting the precision of the detection of HSP70 levels. Despite this, we feel that immunohistochemistry is a timely, simple and easily reproducible method that allows for a widely accepted semiquantitative detection of HSP70. At the same time, it preserves tissue architecture and in this sense provides valuable biochemical and pathological data regarding the intracellular distribution of HSP70. Second, we chose to determine and correlate the occurrence of postoperative AFIB, only with the preoperative and not the postoperative levels of HSP70. We did this because there appears to be considerable controversy in the literature regarding the time frame of the post stress induction of HSP70 synthesis, In one study, nuclear transcription assays have demonstrated increased synthesis of HSP72mRNA as early as 30 min after an ischemic insult [16] and in others [17, 18] this time period was 2–4 h and highly dependent on the degree of hypothermia/hypoxia. Also we chose to obtain atrial myocardial tissue for HSP70 determination because it has been clearly demonstrated that in cardial surgical patients, intramyocardial HSP70 levels do not correlate with either lymphocyte or serum HSP70 levels [15]. The polyclonal antibody used in the present study detects the inducible (HSP72) as well as the consti231
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al.
tutively expressed forms (HSC73) of the HSP70 proteins. Assessing the preoperative intracellular level might, theoretically, reflect only the variations in the constitutively expressed form (HSC73). This preliminary study was not designed to detect intraoperative variations in the levels of various HSP70 members during cardiac surgery. This has been demonstrated by several well designed studies [8–10, 15], which prove in an experimental and in a clinical setting that HSP70 preconditioned hearts demonstrate improved post-ischemic ventricular performance and reduction in infarct size [19]. Our study intended to provide an answer to a specific question: do patients that undergo cardiac surgery with cardiopulmonary bypass and who have postoperative AFIB have a different HSP70 profile than patients that do not suffer any postoperative rhythm disturbances? Our data indicate that patients with low preoperative HSP70 atrial myocardial content have a significantly greater incidence of postoperative AFIB. Furthermore, patients with no postoperative rhythm disturbances have, as a group, statistically significantly higher HSP70 expression than the AFIB patients. Myocardial preoperative preconditioning in humans is still far from clinical application. Data derived from this study demonstrate that in the future it is not unreasonable to speculate that induction of intracellular intramyocardial HSP70 complex as a preoperative measure of cardiac preconditioning might have a potential preventive and therapeutic beneficial effect in the decrease of postoperative atrial fibrillation.
Acknowledgements We would like to thank Ms S. Roeder RN and Ms D. Crowther for their valuable help in the preparation and editing and typing of the manuscript.
References 1. Lauer, M. S., Eagle, K. A., Buckley, M. J. et al., Atrial fibrillation following coronary artery bypass surgery. Prog. Cardiovasc. Dis., 1989, 31, 367–378. 2. Frost, I., Molgaard, H., Christiansen, E. H. et al., Atrial fibrillation and flutter after coronary artery bypass surgery: epidemiology, risk factors, and preventive trials. Int. J. Cardiol., 1992, 36, 253–261.
232
3. Creswell, L. L., Schuessler, R. B., Rosenbloom, M. et al., Hazards of postoperative atrial arrhythmias. Ann. Thorac. Surg., 1993, 56, 539–549. 4. Malhew, J. P., Parks, R., Savino, J. S. et al., Atrial fibrillation following coronary artery bypass graft surgery: predictors, outcomes, and resource utilization. JAMA, 1996, 276, 300–306. 5. Yellon, D. M. and Marber, M. S., HSP 70 in myocardial ischemia. Experientia, 1994, 50, 1075–1084. 6. Mestril, R., Giordano, F. J., Conde, A. G. et al., Adenovirusmediated gene transfer of a heat shock protein 70 (HSP 70) protects against simulated ischemia. J. Mol. Cell Cardiol., 1996, 28, 2351–2358. 7. Plumier, J. C., Ross, B. M., Currie, R. W. et al., Transgenic mice, expressing the human heat shock protein 70 have improved post-ischemic myocardial recovery. J. Clin. Invest., 1995, 95, 1854–1860. 8. Currie, R. W., Karmazyn, M., Kloe, M. et al., Heat shock response is associated with enhanced postischemic ventricular recovery. Circ. Res., 1988, 63, 543–549. 9. Donnelly, T. J., Sievers, R. E., Vissern, R. L. et al., Heat shock protein induction in the rabbit after brief cardiac ischemia. J. Clin. Invest., 1991, 87, 139–147. 10. Amrani, M., Corbett, J., Allen, N. J. et al., Induction of heatshock proteins enhances myocardial and endothelial functional recovery after prolonged cardioplegic arrest. Ann. Thorac. Surg., 1994, 57, 157–160. 11. Meng, X., Banerjee, A., Ao, L. et al., Inhibition of myocardial TNF-alpha production by heat-shock. A potential mechanism of stress-induced cardioprotection against postischemic dysfunction. Ann. NY Acad. Sci., 1999, 874, 69–82. 12. Loncar, R., Flesche, C. W. and Deussen, A., Regional myocardial heat-shock protein (HSP 70) concentrations under different blood flow conditions. Eur. J. Pathol., 1998, 437(1), 98–103. 13. O’Brien, P. J., Bo, L. I., Locke, M. et al., Compensatory opregulation of cardiac SRCa+2-pump by heat-shock counteracts SRCa+2-channel activation by ischemia/reperfusion. Mol. Cell Biochem., 1997, 173(1-2), 135–143. 14. Taggart, D. P., Bakkenist, C. J., Riddolph, S. C. et al., Induction of myocardial heat-shock protein 70 during cardiac surgery. J. Pathol., 1997, 182, 362–366. 15. Demidov, O. N., Vadim, V. T., Svistov, A. S. et al., Heat-shock proteins in cardiosurgery patients. Eur. J. Card.-Thor. Surg., 1999, 16, 444–448. 16. Sharma, H. S., Snoecks, L. H. E. H., Saseen, L. M. A. et al., Immunohistochemical localization of heat-shock protein-72 in the procine heart: enhanced transcription during ischemic preconditioning. Circulation, 1993, 88, 156. 17. Benjamin, I. J., Kroger, B. and Williams, R. S., Activation of the heat-shock transcription factor by hypoxia in mammalian cells. Proc. Natl Acad. Sci. USA, 1990, 87, 6263–6267. 18. Iwaki, U., Chi, S., Dillman, W. H. et al., Induction of HSP70 in cultured rat neonatal cardiomyocytes by hypoxia and metabolic stress. Circulation, 1993, 87, 2023–2032. 19. Iliodromitis, E. K., Karavolias, G. K., Bofilis, E. et al., Enhanced protection of heat-shock in myocardial infarction: inhibition. Cardiovasc. Drugs Ther., 1999, 13(3), 223–236. Paper accepted 30 November 2001
CARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
PII: S0967-2109(01)00138-7
www.elsevier.com/locate/cardiosur
Low preoperative HSP70 atrial myocardial levels correlate significantly with high incidence of postoperative atrial fibrillation after cardiac surgery Kyriakos St. Rammos*, George J. Koullias†, Moustafa O. Hassan‡, Nikolaos P. Argyrakis*, Christos G. Voucharas*, Steven J. Scarupa† and Tomas G. Cowte† *Cardiothoracic Surgery, AHEPA General Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece; †Pinnacle Health System Residency in General Surgery, Harrisburg, PA, USA and ‡Department of Surgery, University of Pittsburgh, Pittsburgh, PA, USA Purpose of the study. Atrial fibrillation after cardiac surgery is still a frequent encountered complication and has been associated with increased hospital length of stay and numerous postoperative complications. The pathogenesis of atrial fibrillation involves an overall sequence of perioperative events, collectively termed as ischemia-reperfusion injury. Heat-shock proteins have been found to provide increased protection during ischemia-reperfusion as well as increased postischemic cardiac functional recovery. We sought to determine whether preoperative atrial heat shock levels were correlated with the appearance of postoperative atrial fibrillation Basic methods. Preoperative atrial myocardial samples obtained just before cannulation from 101 patients were used to detect immunohistochemically the expression of heat-shock proteins. The derived results were compared statistically with the incidence of postoperative atrial fibrillation, its time of appearance, duration and resistance to administered antiarrhythmics. Principal findings. The overall incidence of postoperative atrial fibrillation was 22.3%. Of these patients, 58.3% had no detectable heat shock proteins in their cytoplasm, in sharp contrast with 100% of the patients with no atrial fibrillation who were positive for heat shock proteins (p⬍0.01). Four percent of our patient group had prolonged atrial fibrillation (defined as duration ⬎48 h). These patients had significantly less (p⬍0.01) nuclear heat shock protein expression compared with the non-atrial fibrillation group. However, the difference of the heat shock protein expression between the prolonged atrial fibrillation and the rest of the atrial fibrillation patients was not significant (p = 0.891). Conclusions. Our results indicate that patients with low preoperative atrial heat shock protein expression have a significantly greater incidence of postoperative atrial fibrillation. Heat shock protein expression did not, however, correlate with the onset of atrial fibrillation and the resistance to administered medications. Heat shock protein preoperative induction as a measure of myocardial preconditioning may potentially decrease the incidence of postoperative atrial fibrillation. 2002 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd. All rights reserved Keywords: HSP70 atrial myocardial levels, atrial fibrillation
Introduction Correspondence to Kyriakos St. Rammos, 29 Karolou Diehl Str, Thessaloniki 546 23, Greece. Tel./Fax: +30-31-222-021; e-mail: [email protected]
228
Atrial fibrillation (AFIB) is a common dysrhythmia following coronary artery bypass surgery (CABG) [1, 2]. Postoperative AFIB increases the duration of hospitalization and has in numerous studies been CARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al.
associated with an increased incidence in postoperative stroke, ventricular arrhythmias and the need for a permanent pacemaker [3]. Numerous efforts to determine which patients are likely to develop postoperative AFIB have identified multiple preoperative, intraoperative and even postoperative factors and have led to modifications in several intraoperative practices [4]. Nevertheless, the incidence of AFIB after CABG remains in the vicinity of 25% [4]. One of the recently identified endogenous mechanisms of preoperative and intraoperative myocardial protection is the presence and the possible induction of the heat shock protein (HSP70) family. HSP70 expression is stimulated by ischemia [5]. Furthermore, up-regulation of HSP70 in myocardial cells by adenovirus mediated gene transfer and HSP70 product overexpression in transgenic animals has been shown to greatly increase their resistance to ischemia and improve their post-ischemic recovery [6, 7]. HSP70 proteins function as molecular chaperones facilitating intramyocardial protein assembly, folding, transport and translocation as well as preventing denaturation of intramyocardial enzymes [8–10]. Resident macrophages within myocardial cells containing high HSP70 levels produce significantly lower quantities of TNF-alpha and provide increased protection during ischemia-reperfusion (I/R) as well as echocardiographically proven improved cardiac functional recovery [11]. Local reductions in atrial partial oxygen pressure during cardiopulmonary bypass and postoperative downregulation of atrial intracellular Ca++ pumps have been implicated in the appearance of postoperative AFIB and decreased intramyocardial quantities of HSP70 have been found in both the above biochemical instances [12, 13]. In view of the above, we have examined the expression of HSP70 in right atrial cells of patients undergoing open heart surgery. We have correlated these results with the incidence, the timing of appearance, the intensity, and the resistance to administered medications of postoperative AFIB.
Methods One hundred and one (n = 101) of the patients that underwent open heart surgery using cardiopulmonary bypass at the Department of Cardiothoracic Surgery at the AHEPA University Hospital and operated on by the senior author (KR) were included in the present study. The same cardiac anesthesiologist, perfusionist and operating room team participated in the performance of all these procedures. All patients underwent cardiopulmonary bypass with moderate hypothermia (33–34°C), antegrade crystalloid cardioplegic arrest (St Thomas I) and local hypothermia with ice-slash; cardioplegic solution was readministered every 20–30 min. EpidemiologCARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
ical data of these patients and the incidence of postoperative AFIB are presented in Table 1. The study was approved by the Hospital Ethical Committee and informed consent was obtained from all the above patients preoperatively. Tru-cut biopsy at the right atrial cannulation site as soon as the pericardium was opened yielded approximately 2–3 mm3 of atrial tissue. This tissue sample was immediately placed and fixed in a 10% neutral formalin solution and was subsequently embedded in paraffin. Two 5 µm thick parallel tissue sections from each patient were subjected to a three-step immunoperoxidase immunohistochemical staining technique for the detection and expression of the two major HSP70s using a rabbit polyclonal antibody which identifies HSP72 and HSP73 (DAKO, Califoria, USA, code no. A500). The above method was carried out using DAB as the final chromogen agent. Negative controls were prepared by replacing the primary antibody with normal swine serum and sections of previously known strongly positive tonsil squamous epithelium were used as positive controls. All immunohistochemical material was evaluated blindly by two histopathologists. The intensity of staining was graded semiquantitatively as ‘0’ when no staining was present, ‘1’ when staining was present in up to 10% of cells, ‘2’ when 10–50% of the cells were positive and ‘3’ when the majority (⬎50%) of cells were positive for HSP70. The above grading system was applied separately to the cytoplasm and the nucleus and therefore for each slide a ‘cytoplasmic’ and a ‘nuclear’ HSP70 score were obtained.
Statistical analysis Statistical analysis was performed with the use of the commercially available ‘Primer of Biostatistics’ biostatistical software by Stanton A. Glantz, PhD (McGraw-Hill Health Professions Division, 1997). Student’s t-test and Mann–Whitney Rank sum test were used to evaluate differences between cytoplasmic and/or nuclear HSP70 scores of patients with or without postoperative AFIB. On occasion, analysis of variance (ANOVA) and Spearman correlation was used to also detect and estimate correlations between different variables. The p⬍0.05 value was considered the level of statistically significant difference.
Results Patients with no postoperative AFIB Normal sinus rhythm or sinus tachycardia was observed in 77 out of 101 patients (77.7%). The average cytoplasmic HSP70 score was 2.692±0.057 and the average nuclear HSP70 score was 1.846±0.091. All (100%) of these patients had a 229
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al. Table 1 Preoperative and intraoperative patient data. No statistically significant differences wee observed between patients that had postoperative atrial fibrillation and the rest of the group Patient groups
AFIB (n = 24)
Non-AFIB (n = 77)
Statistical significance
Age Male Female No. of grafts Valve replacements MVR AVR Combined valve and CABG NYHA class IDDM HTN Preop EF Smoking Aortic Cx time CPB time Degree of hypothermia
18 6 3.0 3/24 2 1 1/24 III+0.1 8/24 19/24 42.1% 19/24 58.7±1.48 103.84 33.2±0.2°C
50 27 2.55 10/77 4 6 6/77 III+0.39 15/77 35/77 53.08% 68/77 69.6±0.94 126.15 33.7±19°C
NS NS NS NS NS NS NS NS NS NS NS NS P⬍0.08 NS NS
moderate to intense cytoplasmic HSP70 staining (Figure 1). HSP70 appeared to be located mainly in the cytoplasm, the nucleus and the nucleoli of atrial myocardial cells. Endothelial cells, adjacent macrophages and lymphocytes emerged exempt of HSP70. Patients with postoperative AFIB Atrial fibrillation (AFIB) was seen in 24 out of 101 patients (22.3%) (Table 2). The average cytoplasmic HSP70 score in these patients was 1.857±0.169 and the average nuclear score was 0.714±0.176. HSP70 seem to spare predominantly the nucleus in AFIB patients (Figure 2). In 4/24 of these patients (16.6%) HSP70 was not detected in their cytoplasm and in 14/24 (58.3%) of this group HSP70 was not Figure 2 Histological section of a right atrial biopsy in a patient with postoperative atrial fibrillation. Nuclei are spared from staining and a mild granular HSP70 presence is observed in the cytoplasm (X40)
detected in the nucleus (Figure 3). In both of the above scores (cytoplasmic and nuclear) the difference of the HSP70 scores between the patients with and without postoperative AFIB was statistically sigTable 2 Medication and medication combinations used for atrial fibrillation treatment in our study Medications for atrial fibrillation in 24 patients
Figure 1 Representative immunohistochemical microscopy of right atrial section of a patient with no atrial fibrillation after CABG. Nuclei and nucleoli are intensely stained. The cytoplasm shows intense granular HSP70 staining (X40)
230
Digoxin (1) alone Amiodarone (2) alone Diltiazem (3) alone Propranolol (4) alone (1) and (2) (1) and (3)
CARDIOVASCULAR SURGERY
1 2 1 – 18 9
JUNE 2002 VOL 10 NO 3
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al.
meters and the patient’s HSP70 nuclear or cytoplasmic scores.
Discussion
Figure 3 Right atrial microscopy of a patient with postoperative atrial fibrillation. No HSP70 was detected on either the nuclei or the cytoplasm (X40)
nificant (p⬍0.001 — 98 degrees of freedom, t = 7.746 and ⫺4.944, respectively) (Figure 4). Patients with prolonged (⬎48 h) AFIB Prolonged — as defined above — atrial fibrillation was encountered in 4/24 AFIB patients (16.6%). Statistical comparison of these four patients with the rest of the AFIB group revealed that patients with prolonged AFIB had a significantly lower (p⬍0.01) nuclear HSP70 expression. However, the difference in the cytoplasmic HSP70 expression between these four patients and the rest of the AFIB group was not significant (p⬍0.891). Medications for atrial fibrillation Table 2 shows the incidence, kind and number of medications used for the treatment of atrial fibrillation in the AFIB group. No statistically significant associations were observed between these para-
Figure 4 Graphic presentation of cytoplasmic (white) and nuclear (black) HSP70 staining scores between non-AFIB, AFIB and prolonged (⬎48 h) AFIB patients
CARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3
Several studies in the literature have pointed to the cytoprotective role of HSP70 from ischemia and reperfusion myocardial injury after actual or experimental ischemia or open heart surgery [8–10, 14, 15]. The HSP70 family of proteins contains at least four members; HSP72, HSC73, BiP and GRP 76. From these, myocardial cells contain for certain the first two; HSP72 is inducible following myocardial injury. HSC73 appears to be constitutively expressed and its levels seem to be dependent on previous injury, use of medications or are genetically determined. Our study is the first report that demonstrated a repeated low nuclear and cytoplasmic atrial myocardial expression of HSP72/HSC73 in preoperative specimens of patients that had AFIB after open heart surgery. Low HSP70 levels were significantly correlated with the occurrence of postoperative AFIB. HSP70 presence, however, did not correlate with the time of AFIB onset, nor the duration or the resistance of AFIB reversal to different administered medications. Although our study patient sample (n = 101) is of adequate size to value any statistically significant results, two issues in our methodology deserve further elucidation. First the performance of immunohistochemistry. Immunohistochemical techniques do not allow for accurate quantitative measurements, thus limiting the precision of the detection of HSP70 levels. Despite this, we feel that immunohistochemistry is a timely, simple and easily reproducible method that allows for a widely accepted semiquantitative detection of HSP70. At the same time, it preserves tissue architecture and in this sense provides valuable biochemical and pathological data regarding the intracellular distribution of HSP70. Second, we chose to determine and correlate the occurrence of postoperative AFIB, only with the preoperative and not the postoperative levels of HSP70. We did this because there appears to be considerable controversy in the literature regarding the time frame of the post stress induction of HSP70 synthesis, In one study, nuclear transcription assays have demonstrated increased synthesis of HSP72mRNA as early as 30 min after an ischemic insult [16] and in others [17, 18] this time period was 2–4 h and highly dependent on the degree of hypothermia/hypoxia. Also we chose to obtain atrial myocardial tissue for HSP70 determination because it has been clearly demonstrated that in cardial surgical patients, intramyocardial HSP70 levels do not correlate with either lymphocyte or serum HSP70 levels [15]. The polyclonal antibody used in the present study detects the inducible (HSP72) as well as the consti231
HSP70 atrial myocardial levels correlate with incidence of postoperative atrial fibrillation: K. St. Rammos et al.
tutively expressed forms (HSC73) of the HSP70 proteins. Assessing the preoperative intracellular level might, theoretically, reflect only the variations in the constitutively expressed form (HSC73). This preliminary study was not designed to detect intraoperative variations in the levels of various HSP70 members during cardiac surgery. This has been demonstrated by several well designed studies [8–10, 15], which prove in an experimental and in a clinical setting that HSP70 preconditioned hearts demonstrate improved post-ischemic ventricular performance and reduction in infarct size [19]. Our study intended to provide an answer to a specific question: do patients that undergo cardiac surgery with cardiopulmonary bypass and who have postoperative AFIB have a different HSP70 profile than patients that do not suffer any postoperative rhythm disturbances? Our data indicate that patients with low preoperative HSP70 atrial myocardial content have a significantly greater incidence of postoperative AFIB. Furthermore, patients with no postoperative rhythm disturbances have, as a group, statistically significantly higher HSP70 expression than the AFIB patients. Myocardial preoperative preconditioning in humans is still far from clinical application. Data derived from this study demonstrate that in the future it is not unreasonable to speculate that induction of intracellular intramyocardial HSP70 complex as a preoperative measure of cardiac preconditioning might have a potential preventive and therapeutic beneficial effect in the decrease of postoperative atrial fibrillation.
Acknowledgements We would like to thank Ms S. Roeder RN and Ms D. Crowther for their valuable help in the preparation and editing and typing of the manuscript.
References 1. Lauer, M. S., Eagle, K. A., Buckley, M. J. et al., Atrial fibrillation following coronary artery bypass surgery. Prog. Cardiovasc. Dis., 1989, 31, 367–378. 2. Frost, I., Molgaard, H., Christiansen, E. H. et al., Atrial fibrillation and flutter after coronary artery bypass surgery: epidemiology, risk factors, and preventive trials. Int. J. Cardiol., 1992, 36, 253–261.
232
3. Creswell, L. L., Schuessler, R. B., Rosenbloom, M. et al., Hazards of postoperative atrial arrhythmias. Ann. Thorac. Surg., 1993, 56, 539–549. 4. Malhew, J. P., Parks, R., Savino, J. S. et al., Atrial fibrillation following coronary artery bypass graft surgery: predictors, outcomes, and resource utilization. JAMA, 1996, 276, 300–306. 5. Yellon, D. M. and Marber, M. S., HSP 70 in myocardial ischemia. Experientia, 1994, 50, 1075–1084. 6. Mestril, R., Giordano, F. J., Conde, A. G. et al., Adenovirusmediated gene transfer of a heat shock protein 70 (HSP 70) protects against simulated ischemia. J. Mol. Cell Cardiol., 1996, 28, 2351–2358. 7. Plumier, J. C., Ross, B. M., Currie, R. W. et al., Transgenic mice, expressing the human heat shock protein 70 have improved post-ischemic myocardial recovery. J. Clin. Invest., 1995, 95, 1854–1860. 8. Currie, R. W., Karmazyn, M., Kloe, M. et al., Heat shock response is associated with enhanced postischemic ventricular recovery. Circ. Res., 1988, 63, 543–549. 9. Donnelly, T. J., Sievers, R. E., Vissern, R. L. et al., Heat shock protein induction in the rabbit after brief cardiac ischemia. J. Clin. Invest., 1991, 87, 139–147. 10. Amrani, M., Corbett, J., Allen, N. J. et al., Induction of heatshock proteins enhances myocardial and endothelial functional recovery after prolonged cardioplegic arrest. Ann. Thorac. Surg., 1994, 57, 157–160. 11. Meng, X., Banerjee, A., Ao, L. et al., Inhibition of myocardial TNF-alpha production by heat-shock. A potential mechanism of stress-induced cardioprotection against postischemic dysfunction. Ann. NY Acad. Sci., 1999, 874, 69–82. 12. Loncar, R., Flesche, C. W. and Deussen, A., Regional myocardial heat-shock protein (HSP 70) concentrations under different blood flow conditions. Eur. J. Pathol., 1998, 437(1), 98–103. 13. O’Brien, P. J., Bo, L. I., Locke, M. et al., Compensatory opregulation of cardiac SRCa+2-pump by heat-shock counteracts SRCa+2-channel activation by ischemia/reperfusion. Mol. Cell Biochem., 1997, 173(1-2), 135–143. 14. Taggart, D. P., Bakkenist, C. J., Riddolph, S. C. et al., Induction of myocardial heat-shock protein 70 during cardiac surgery. J. Pathol., 1997, 182, 362–366. 15. Demidov, O. N., Vadim, V. T., Svistov, A. S. et al., Heat-shock proteins in cardiosurgery patients. Eur. J. Card.-Thor. Surg., 1999, 16, 444–448. 16. Sharma, H. S., Snoecks, L. H. E. H., Saseen, L. M. A. et al., Immunohistochemical localization of heat-shock protein-72 in the procine heart: enhanced transcription during ischemic preconditioning. Circulation, 1993, 88, 156. 17. Benjamin, I. J., Kroger, B. and Williams, R. S., Activation of the heat-shock transcription factor by hypoxia in mammalian cells. Proc. Natl Acad. Sci. USA, 1990, 87, 6263–6267. 18. Iwaki, U., Chi, S., Dillman, W. H. et al., Induction of HSP70 in cultured rat neonatal cardiomyocytes by hypoxia and metabolic stress. Circulation, 1993, 87, 2023–2032. 19. Iliodromitis, E. K., Karavolias, G. K., Bofilis, E. et al., Enhanced protection of heat-shock in myocardial infarction: inhibition. Cardiovasc. Drugs Ther., 1999, 13(3), 223–236. Paper accepted 30 November 2001
CARDIOVASCULAR SURGERY
JUNE 2002 VOL 10 NO 3