Persistent Inflammatory Activity in Blood Cells and Artery Tissue from Patients with Previous Bare Metal Stent.

BACKGROUND: Studies have pointed out a higher mortality after coronary artery bypass surgery (CABG) in patients with stent.
OBJECTIVE: To evaluate inflammatory markers in peripheral blood cells and in coronary artery tissue samples obtained during CABG in patients with stent compared to controls.
METHODS: The case series consisted of two groups, one with previous stent implantation (n = 41) and one control (n = 26). The expression of the LIGHT, IL-6, ICAM, VCAM, CD40, NFKB, TNF, IFNG genes was analyzed in peripheral blood cells collected preoperatively. The coronary artery was evaluated for: interleukin-6, ICAM, VCAM, CD40, NFKB, TNF-alpha and IFN-gamma by immunohistochemistry. A total of 176 tissue samples were grouped for analysis in: A1- arteries with stent (n = 38); A2- native arteries from patients with stent in another artery (n = 68); and A3- arteries without stent from controls undergoing routinely CABG surgery (n = 70). A significance level of 0.05 was adopted.
RESULTS: Patients with stent showed higher TNF (p = 0.03) and lower CD40 gene expression (p = 0.01) in peripheral blood cells than controls without stent. In coronary artery samples, the TNF-alpha protein staining was higher in the group A1, not only in the intima-media layer (5.16 ± 5.05 vs 1.90 ± 2.27; p = 0.02), but also in the adipose tissue (6.69 ± 3.87 vs 2.27 ± 4.00; p < 0.001). Furthermore, group A1 had a higher interleukin-6 protein staining in adipose tissue than group A3 (p = 0.04).
CONCLUSION: We observed a persistently higher systemic TNF expression associated with exacerbated TNF-alpha and interleukin-6 local production in patients with stents. This finding may contribute to a worse clinical outcome.

Introduction

Retrospective studies have suggested that coronary artery bypass grafting (CABG) surgery after percutaneous coronary intervention (PCI) can impair short and long-term outcomes.[1]-[7] Previous studies have demonstrated that PCI is associated with higher in-hospital mortality, despite the lower risk profile of PCI patients,[3] but there is no consensus in the literature.[8] An analysis of the MASS study[9] has shown that patients who underwent PCI treatment were more likely to develop progression in native coronary arteries, than those undergoing CABG or medical treatment.

During PCI, a focal inflammatory reaction occurs with plaque rupture caused by stent implantation, but there are still controversies if this reaction persists in the long term. There is scarce information about a persistent systemic inflammatory reaction or tissue mediators in coronary artery after stent implantation, as well as about the comparison between coronary arteries with stent, coronary arteries without stent but with stent in another artery and controls patient.

The CABG allows a unique opportunity to collect coronary artery specimen to evaluate local inflammatory reaction long after stent implantation. This study aims to evaluate inflammatory genes expression in peripheral blood cells and inflammatory protein localization in coronary artery tissue obtained during CABG from patients with and without previous stent implantation. It is worth mentioning that CABG represents a unique opportunity to obtain a tiny coronary artery tissue sample to evaluate local inflammatory in humans. Nowadays, patients who receive previous bare metal stent (BMS) implantation and need CABG surgery later represent a significant number of cardiology hospital patients, especially in developing countries. Our results can contribute to clarify the involvement of persistent local and systemic inflammation in the later phase of stent restenosis.

Methods

Case series

The patients admitted to elective CABG with previous stent implantation were consecutively included in this study after signing informed consent. The protocol was approved by the Dante Pazzanese Institute of Cardiology ethics committee (Protocol 4059-2011).

This study included 67 patients as follows: 41 patients with previous intracoronary BMS implantation and 26 patients without stent implantation submitted to elective CABG. All patients had stable angina and more than 6 months of the stent implantation, aiming to exclude ongoing restenosis. The exclusion criteria consisted of emergency surgeries, acute coronary syndromes and chronic renal failure in dialysis because of chronic inflammatory reaction.

Peripheral blood sample was collected preoperatively from the antecubital vein, using PAXgene tubes (PreAnalytiX®, BD Company, UK) for systemic gene expression analysis.

During CABG surgery, a tiny fragment of the bypassed coronary artery was obtained at the arteriotomy site, usually 10 mm after the stent implantation site, to evaluate local inflammation markers. All tissue samples were immediately immersed in buffered formalin for further paraffin-embedded block preparation. Some arterial fragments obtained were not adequate for histological analysis, thus, 176 artery samples were included and grouped as follows: A1- arteries with stent (n = 38); A2- native arteries from a patient with a stent in another artery (n = 68); and A3- arteries from patients without previous stent placement (n = 70).

RNA isolation, reverse transcription and real-time polymerase chain reaction (PCR)

Total RNA was extracted from peripheral blood collected in PAXgene tubes (PreAnalytiX®, BD Company, UK) using PAXgene blood RNA kit (QIAGEN GmbH, Hilden, Germany), being then quantified using Qubit® 2.0 fluorometer (Life Technologies, Inc., Grand Island, NY). The RNA integrity was performed using Tapestation® 2200 and R6K Screen Tape (Agilent Technologies, Inc. UK). The cDNA was transcribed from 200 ng of total RNA using High Capacity RNA-to-cDNA Master Mix (Applied Biosystems, Foster City, USA). Real-time qPCR was carried out in Rotor-Gene® detection system using the TaqMan® Fast Multiplex PCR kit (QIAGEN GmbH, Hilden, Germany) and primers from Applied Biosystem commercially designed for TaqMan® qPCR (Applied Biosystems, Foster, CA, USA).

The expression of the following genes was evaluated: LIGHT (Hs00542477_m1); IL-6 (Hs00985639_m1); ICAM (Hs00164932_m1); VCAM (Hs00174239_m1); CD40 (Hs01002913_g1); NFKB (Hs00231653_m1); TNF (Hs01113624_g1); IFNG (Hs00989291_m1) and GAPDH (Hs00266705_g1). For all genes, we constructed standard curves and determined the slope to calculate the PCR efficiency. Almost equal efficiency for all primer/probe systems was observed. All samples were tested in duplicate using GAPDH as a reference gene, which was previously chosen between the six most common endogenous myocardium genes using geNorm algorithm.[10] The samples amplified after 40 cycles of PCR were considered negative and excluded from further statistical analysis. The expression of the reference GAPDH gene was applied for data normalization, and the relative expression of each mRNA was calculated using the 2 -∆∆CT method.[11]

Immunohistochemistry staining

Firstly, 4-µm-thick formalin-fixed-paraffin-embedded artery samples were sectioned and fixed in silanized slides, followed by dewaxing at 70°C, in the oven, for 1 hour, and immersion in three xylene baths for 10 minutes. They were then rehydrated in decreasing concentrations (100%, 90%, 75%) of ethyl alcohol. Antigen retrieval was performed using the Trilogy® buffer (Cell Marque, California, USA) at a Decloaker equipment at 90°C for 40 minutes (Biocare Medical, CA, USA). The specific blocking reagents (Erviegas EasyPath, DuraEdge, USA) were applied for endogenous peroxidase and protein blocking. In the next step, the slides were incubated with respective primary antibodies previously titrated and diluted in universal thinner (Erviegas EasyPath, DuraEdge, USA). Primary antibodies against interleukin-6 (IL-6) (ab6672), ICAM (ab2213), VCAM (ab106777), TNF-alpha (ab1793), IFN-gamma (ab9657), CD40 (ab58612), and NFKB (ab16502) from Abcam (Cambridge, MA, UK) were used. The immunoperoxidase reaction was detected using Mach4 Kit Universal HRP Polymer + DAB (Biocare Medical, California, USA), and, finally, the slides were stained with Harris hematoxylin (Erviegas EasyPath, DuraEdge, USA) and assembled by synthetic resin Erv-Mount (Erviegas EasyPath, DuraEdge, USA). The positive control of immunohistochemistry reaction was performed using tissues that have the same constructive antigens as the antigen of interest. After immunohistochemistry processing, the slides were scanned through a Scanscope CS System unit (Aperio Technologies, Inc., CA, USA), with an objective 20x Olympus UPlanSApo with specifications 20x/0.75 attached to the scanner, generating image files in svs format. Scanned images were analyzed using the Aperio ImageScope viewing software (Aperio Technologies, Inc., CA, USA) that reports the percentage of positively stained area in relation to the total tissue area.

Statistical analysis

Continuous variables are reported as mean and standard deviation or median and interquartile interval, depending on the assumption of normality. Categorical variables are reported as absolute and relative frequency. Values between groups were compared by unpaired Student t test after testing for normal distribution by KS test; otherwise, nonparametric Mann-Whitney U tests were used. Fisher exact or chi-square test was used for categorical variables with nominal scales. For comparison of artery tissue markers, Kruskal-Wallis test (or ANOVA, assumption of normality) was used, and, for non-parametric multiple comparisons, Tukey's test. A p-value lower than 0.05 was considered statistically significant. The SPSS version 19 was used.

To detect 3 units with standard deviation of 4, 80% test power, and 5% alpha, the sample size calculation is 105 cases.

Results

Clinical characteristics of study group

The frequencies of clinical characteristics (Table 1), such as sex, diabetes, dyslipidemia, smoking, previous stroke and heart attack, are similar in both groups. However, the stent group was younger and had a higher prevalence of ventricular dysfunction, characterized by an ejection fraction lower than 50%. The control group had a higher blood platelet count (269,560 ± 74,461) than the stent group (237,355 ± 70,831), but with no statistical significance (p = 0.12). All patients were on statins and acetylsalicylic acid treatment.

Table 1

Biodemographic data of the studied groups: with previous stent implantation and controls

Variables	Stent Group	Control Group	p value
Cases	41	26
Age (years)	60.2 ± 7.1	6.3 ± 8.69	0.004
Female (%)	13 (31.3)	8 (30.8)	0.58
Blood platelet count	237.355 ± 70.831	269.560 ± 74.461	0.12
Hypertension (%)	35 (85.4)	22 (84.6)	0.60
Diabetes (%)	17 (42.5)	8 (36.4)	0.42
Current smoking (%)	4 (10.5)	3 (14.3)	0.48
Stroke (%)	2 (5.3)	0	0.43
CKD (%)	2 (5.3)	0	0.43
Dyslipidemia (%)	25 (62.5)	16 (64)	0.56
MI (%)	14 (35.9)	5 (22.7)	0.22
LVEF < 50% (%)	14 (41.2)	2 (9.1)	0.009
ASA (%)	41 (100)	26 (100)	1.00
Statins (%)	41 (100)	26 (100)	1.00

CKD: chronic kidney disease; MI: myocardial infarction; LVEF: left ventricular ejection fraction; ASA: acetylsalicylic acid. Age and blood platelet count were expressed as mean ± standard deviation. Sex, hypertension, diabetes, smoking, stroke, CKD, dyslipidemia, MI, LVEF < 50%, ASA use, and statin use are expressed in numbers and percentages of the patients studied. The statistical significance level adopted was p < 0.05.

The time between stent implantation and CABG was over 6 months. Nine patients (22%) had stent implanted within 6 and 12 months, 25 patients (61%) had only one stent, and 16 patients (39%) had two or more stents implanted.

Gene expression in peripheral blood cells by real-time PCR

Total RNA from peripheral blood cells was obtained and the expression of the following genes was evaluated: LIGHT, IL-6, ICAM, VCAM, CD40, NFKB, TNF, IFNG and GAPDH. Of eight genes, the expression of only two differed between the stent and control groups: the expression of TNF was significantly higher (p = 0.0308) in the stent group (Figure 1f-) and that of CD40 was higher in the control group (p = 0.0106) (Figure 1a-). No difference was detected in the expression of IL-6, IFNG, LIGHT, NFKB, ICAM and VCAM genes (Figure 1).

Figure 1

Expression of inflammatory genes in peripheral blood. The expression of the genes CD40 (a), IL-6 (b), INFG (c), LIGHT (d), NFKB (e), TNF (f), ICAM (g) and VCAM (h) was evaluated by real-time PCR using GAPDH gene as internal control to calculate relative expression (2-ΔΔCT). The patients with coronary stent implantation (n = 35) were compared with controls (n = 25) using nonparametric Mann-Whitney U test. The difference is considered significant for p-values < 0.05.

Quantitative analysis by immunohistochemistry

The quantification of proteins staining by immunohistochemistry is presented in Table 2 and illustrated in Figures 2 and 3. The TNF-alpha staining in the arteries of the adipose tissue was higher in group A1 than in group A2 (6.69 ± 3.87 vs 2.27 ± 4.00; p < 0.001) (Figure 2-a). In addition, group A1 had higher TNF-alpha staining in the intima-media region than group A3 did (5.16 ± 5.05 vs 1.90 ± 2.27; p = 0.023) (Figure 2-b). A large amount of TNF-alpha was detected in the cytoplasm of inflammatory cells and around the lipid core (Figure 3-C and D).

Table 2

Quantification of CD40, ICAM, VCAM, MHC-II, TNF-alpha, NFKB, IL-6 and IFN-gamma proteins in arterial intima-media layer, adventitia and adipose tissue by immunohistochemistry

Protein	Artery layers	Group A1			Group A2			Group A3			p value (α = 0.05)
		n	mean	Std. dev	n	mean	Std. dev	n	mean	Std. dev	3 groups	1 vs 2	1 vs 3	2 vs 3
CD40	intima-media	16	1.37	2.02	27	1.51	1.73	23	1.11	1.56	0.55	ns	ns	ns
	Adventitia	13	0.70	0.77	26	0.82	0.71	19	1.27	1.66	0.58	ns	ns	ns
	Adipose tissue	3	0.73	0.12	6	0.62	0.58	9	0.58	0.74	0.45	ns	ns	ns
ICAM	intima-media	18	3.27	3.00	27	3.47	3.77	20	2.81	4.23	0.16	ns	ns	ns
	Adventitia	14	4.07	3.67	25	4.22	4.82	19	3.99	5.39	0.76	ns	ns	ns
	Adipose tissue	1	5.93	-	7	1.99	1.37	5	1.52	1.64	0.22	ns	ns	ns
VCAM	intima-media	14	11.88	16.01	24	10.33	9.64	23	7.88	6.12	0.76	ns	ns	ns
	Adventitia	11	4.69	8.39	19	4.09	4.74	21	2.56	2.26	0.68	ns	ns	ns
	Adipose tissue	2	4.31	5.28	4	1.76	1.22	4	0.32	0.36	0.10	ns	ns	ns
MHC II	total area	7	0.74	0.59	8	0.47	0.12	30	0.75	0.52	0.307	ns	ns	ns
TNF-alpha	intima-media	15	5.16	5.05	24	3.11	3.01	21	1.90	2.27	0.03	ns	0.023	ns
	Adventitia	14	4.05	2.82	21	2.28	2.24	20	3.57	5.95	0.10	ns	ns	ns
	Adipose tissue	4	6.69	3.88	4	1.27	0.84	9	2.27	4.00	0.05	0.001	ns	ns
NFKB	intima-media	14	1.11	1.14	24	0.93	0.88	20	0.92	1.07	0.96	ns	ns	ns
	Adventitia	14	0.64	0.61	21	0.83	0.78	19	0.76	0.55	0.65	ns	ns	ns
	Adipose tissue	2	1.18	0.87	6	0.52	0.44	6	1.63	2.58	0.66	ns	ns	ns
Interleukin-6	intima-media	16	1.15	1.12	23	1.27	1.61	21	0.66	0.88	0.17	ns	ns	ns
	Adventitia	15	1.65	2.04	20	1.65	2.02	21	0.91	0.74	0.69	ns	ns	ns
	Adipose tissue	4	1.12	1.03	3	2.29	1.96	11	0.28	0.33	0.01	ns	0.061	0.048
IFN-gamma	intima-media	14	0.67	0.70	25	0.66	0.69	22	0.56	0.84	0.36	ns	ns	ns
	Adventitia	12	0.52	0.41	20	0.59	0.54	21	0.40	0.46	0.36	ns	ns	ns
	Adipose tissue	3	0.54	0.53	4	0.14	0.10	7	0.10	0.11	0.13	ns	ns	ns

The comparison was performed among groups (Kruskal-Wallis test or ANOVA): A1 (arteries with stent), A2 (native arteries from patients with a stent in another artery), and A3 (control, patients without previous stent placement). The statistical significance level adopted was p < 0.05.

Figure 2

Comparison of the staining of TNF-alpha and IL-6 proteins in different layers of arterial tissue. The analysis was performed comparing three groups: A1 (arteries with stent), A2 (native arteries from patient with a stent in another artery) and A3 (control, patients without previous stent placement). TNF-alpha protein staining was higher in the adipose tissue of group A1 (6.69 ± 3.87 vs 2.27 ± 4.00; p < 0.001) (a), as well as in the intima-media layer (5.16 ± 5.05 vs 1.90 ± 2.27; p = 0.02) (b). The IL-6 protein staining was also higher in the adipose tissue from group A1 than that from group A3 (2.29 ± 1.96 vs 0.28 ± 0.33; p = 0.048) (c). The Kruskall-Wallis test and nonparametric Tukey’s multiple comparisons test were used for statistical analysis. The difference is considered significant for p-values < 0.05.

Figure 3

Panoramic (left side) and high-power view (right side) of immunostained arterial intima-media layer from individuals with previous stent implantation. Panels A and B show MHCII-positive cells, with morphology of macrophages (arrows), surrounding the lipid core (LC). Panels C and D show a large amount of TNF-alpha in the cytoplasm of inflammatory cells (arrows) and in the lipid core (LC). Panels E and F exhibit fewer inflammatory cells positive for IL-6 protein in similar sites (arrows).

Higher IL-6 was detected in arteries of the adipose tissue of group A1 than in those of group A3 (2.29 ± 1.96 vs 0.28 ± 0.33; p = 0.048) (Figure 2-c), also observed under the microscope (Figure 3-E and F). There was no difference between groups in the quantification of the CD40, ICAM, VCAM, NFKB and IFN-gamma staining.

On histological examination, we identified major histocompatibility complex class II (MHCII) positive cells surrounding the lipid core, probably macrophages (Figure 3-A and B). These infiltrate cells also stained for TNF-alpha, being detected in adipose tissue and in intima layer (Figure 3-C and D). Fewer IL-6-positive infiltrate cells were also observed in the same layer of artery tissue (Figure 3-E and F).

Discussion

This study analyzed both the gene expression in peripheral blood cells and the protein localization in coronary artery tissues, to evaluate simultaneous systemic and local inflammation. A persistently higher TNF-alpha systemic expression was observed in peripheral blood cells, in addition to a local exacerbated TNF-alpha and IL-6 production in coronary arteries.

Our study evaluated, in stable patients, local coronary and systemic inflammation after stent implantation as compared to controls. All patients included had undergone PCI more than 6 months before and had CABG indication. Regarding stent implantation time, 9 of 41 patients (22%) were within one year from PCI, and 32 patients (78%) had it more than one year before. A former study showed that the inflammatory reaction inherent to PCI no longer exists in this period.[12] No patient with drug-eluted stent (DES) was enrolled in the present study. On the basis of previous literature, a higher number of T lymphocytes and macrophages is observed in DES lesions than in BMS lesions, suggesting the mechanism of restenosis after DES implantation may be different from that observed after BMS implantation.[13]

In this study, the gene expression in peripheral blood cells and the protein localization in coronary artery tissues were carried out to evaluate systemic and local inflammation, respectively. It is worth mentioning that CABG surgery represents a rare opportunity to obtain coronary tissue samples for research without harm to patients, that is why few previous studies worked with this kind of biological sample. In most of them, samples were obtained from the atherosclerotic plaque by endarterectomy.[14],[15]

Blood Analysis

The analysis of mRNA expression from circulating blood cells pointed out a significant higher expression of the TNF gene in the group with previous stent implantation than in controls (Figure 1-f), suggesting a greater activation of this gene in leukocytes from stented patients. This gene encodes a pleiotropic cytokine involved in a broad range of biological activities, including inflammation, cell survival, cell proliferation, and, paradoxically, cell death.[16]

We also observed a significantly lower CD40 gene expression in blood cells from the stent group than from controls (Figure 1-a). The CD40 is the receptor for CD40L, being present in platelets. Gerdes et al.[17] have demonstrated in knockout mice for CD40 and ApoE that platelet plays a crucial role in inflammation by stimulating leukocyte and endothelial cells activation, thereby promoting atherosclerosis.

Coronary artery tissue samples

We analyzed the arterial tissue separated in three layers (adventitia, intima-media and adipose tissue) by hematoxylin-eosin and immunohistochemistry staining. It is worth pointing that, although only few samples collected contained adipose tissue, due to the difficulty to obtain all layers in such small fragments, we could discriminate a greater amount of TNF-alpha and IL-6proteins in the adipose tissue from groups A1 and A2 than in that from controls (Figure 2-a and 2-c). The white adipose tissue is considered to be an endocrine gland, and the main feature is insulin and leptin resistance, as well as the production of inflammatory cytokines (TNF-alpha and IL-6) and monocyte chemoattractant protein,[18],[19] which are involved in atherogenesis.[20],[21] Interestingly, on the histological study, we observed activated immune cells, with MHCII expressed in membrane (Figure 3), surrounding the lipid core. Furthermore, these cells were colocalized with TNF-alpha and IL-6 staining, suggesting a greater inflammatory response in the adipose tissue around the artery from individuals with previous stent placement.

The TNF-alpha protein was also expressed in higher quantity in intima-media layer from group A1 than from groups A2 and A3 (Figure 2-b). Probably immune cells migrated from circulation to that layer, mainly macrophages, which are responsible for the production of that cytokine. Those are also MHCII-positive cells, primarily responsible for presenting antigenic peptides to T cells of the immune system.

Interleukin-6 is a multifunctional cytokine playing a central role in inflammation and tissue injury.[22] Interleukin-6 activates platelet receptor GPIIb/IIIa and leukocyte-platelet interaction, thus favoring the prothrombotic and atherogenic formation. Previous studies have shown that the increased circulating IL-6 is associated with the risk of coronary restenosis and de novo coronary artery lesions,[23] as well as the severity of stenosis.[24] The increase in IL-6 mRNA and protein has been observed in human arterial atherosclerotic wall. In this study, our data showed a significant higher amount of IL-6 protein in the coronary tissue of patients with previous stent placement than in that of controls, suggesting that local arterial inflammation is intensified by stent placement.

Our finding indicated the presence of persistent systemic and local chronic inflammation in individuals with previous stent implantation and can probably contribute to the worst outcome described in a previous meta-analysis study.[7]

It is known that persistent inflammatory response may result in several complications, such as atherosclerotic plaque formation in arterial vessels. In a substudy of the MASS II Trial,[9] consecutive angiographic results were compared with the progression of coronary artery atherosclerosis in medical treatment (MT), CABG and angioplasty. The authors have observed a greater progression in at least one native vessel in angioplasty patients than in CABG and MT patients, concluding that angioplasty treatment has the worst progression in native coronary arteries, especially in the left anterior descending territories. Our result also showed inflammation process in native arteries from individuals with previous stent implantation.

Limitations

This study was limited to BMS. We had very few DES with restenosis and CABG indication. Because of the tiny size of the samples, only a small number of adipose tissue samples was obtained in group A2. Unfortunately, because of the tiny size, many samples collected were inadequate for analysis, and some had insufficient material for analysis. It is worth mentioning that the surgeon primarily ensures the patient’s safety.

The arterial sample was collected at the least affected segment, for best surgical results in graft implantation, distal to the stent, place of possible less inflammation and less affected by the stent.

Restenosis is associated with a local and systemic inflammatory reaction that could be related to obstructive lesions in stented arteries. However, only nine patients were operated upon in less than 365 days, and the arterial samples were taken at least 10 mm after the stent implantation, which reduces its influence on the results.

Statin, which has known anti-inflammatory activity, can influence partially on this result, but it was minimized because all patients were under this therapeutic procedure that has a class I indication.

Conclusion

In conclusion, the persistently higher systemic expression of TNF in association with the local exacerbated TNF-alpha and IL-6 production in coronary arteries with previous BMS implantation may contribute to worse clinical outcomes after CABG surgery.