|Home | About | Journals | Submit | Contact Us | Français|
Several inflammatory mediators such as vascular endothelial growth factor (VEGF) and hepatocyte growth factor (HGF) are known to play a critical role in the regulation of vascular permeability and angiogenesis. We studied the serum levels of growth factors and gene expression profiles of genes involved in the growth factor signaling in the peripheral blood of diabetic (DM) and non-diabetic (ND) patients following cardiopulmonary bypass (CPB) and cardioplegic-arrest (C).
Serum and total RNA were obtained from the blood samples collected from DM and matched ND patients (n=7 patients each) who underwent coronary artery bypass graft (CABG); before (pre), 6-hours (6h), and 4-days (4d) after CPB/C. The cytokine panel consisting of growth factors such as VEGF, HGF, FGF, and EGF was quantified in DM and ND patients; pre, 6h, and 4d post-CPB/C using multiplex cytokine quantification system. cDNA microarray analysis was performed and fold-change was calculated. Length of hospitalization (10 vs. 6 days; p=0.04) and weight gain (5 vs. 2.5 kg; p=0.001) were significantly greater for DM compared to ND patients. The serum levels of VEGF and HGF were significantly elevated in DM patients when compared to ND patients; pre vs. 6h post-CPB/C. In addition, significantly elevated mRNA expression of HIF1α, CREB, and EP300 (> 2 fold) was observed, 4d post-CPB/C exclusively in DM patients.
The differential profile of gene and protein expression of growth factors and their related genes in DM vs. ND patients could be associated with increased edema and weight gain in DM patients after CPB/C.
Cardiovascular disease is the leading cause of death among individuals with diabetes mellitus (DM) and ~ 34% of the patients undergoing coronary artery bypass grafting (CABG) have been reported to possess diabetes (1). Several adverse surgical outcomes such as higher perioperative mortality, sternal wound infections, postoperative strokes, and longer hospitalization post-cardiac surgery have been reported in DM patients compared to patients without diabetes mellitus (ND) (2, 3). Cardiopulmonary bypass during cardiac surgery (CPB) is associated with extensive inflammatory triggers including the blood-bypass circuit interface, ischemia-reperfusion injury of heart and lungs, and operative trauma. These inflammatory signals induce the activation of microcirculation throughout the body, thus leading to multi-organ dysfunction in the immediate post-operative period (4). The other major contributory factor for organ dysfunction after CPB is post-operative hypoxia.
Hypoxic conditions can induce a variety of genes encoding proteins including growth factors such as VEGF, via hypoxia-inducible factor (HIF-1α), which regulates vascular permeability, and cell proliferation (5). Growth factors such as insulin-like growth factor (IGF-1) (6), hepatocyte growth factor (HGF) (7), keratocyte growth factor (KGF) (8), transforming growth factor (TGF- β1) (9) and also the inflammatory mediators including interleukin-1 (IL-1) and interleukin-6 (IL-6) (10) are known to regulate VEGF synthesis. VEGF is a multi-functional cytokine, over-expressed in hypoxia, and is known to be involved in wound healing (11), formation of collateral vessels in ischemic tissue (12), chronic inflammatory disorders (psoriasis, rheumatoid arthritis) (13), diabetic retinopathy (14), and tumor growth (15). The biological properties of VEGF include endothelial cell proliferation (16), chemotaxis of macrophages and vascular smooth muscle cells (17), induction of adhesion molecules (VCAM, and ICAM) (18), and metalloproteinase synthesis (19). On the other-hand, C-Reative Protein (CRP), a potential marker for acute inflammation and a potent independent predictor for cardiovascular diseases including arteriosclerosis (20), has been reported to down-regulate VEGF receptors on endothelial cells (21) and thus may lead to decreased angiogenesis in DM patients. Acute systemic hypoxia can induce leukocyte extravasation into tissues via VEGF signaling followed by fluid accumulation and thus induces overall weight gain in the patients after cardiac surgery.
The mechanism/s contributing to the differences in the post-surgical complications observed between DM and ND patients are not well established. In this study we evaluated serum levels of several growth factors including VEGF and HGF as well as gene expression profiles of genes involved in growth factor signaling in the peripheral blood of DM and ND patients following CPB/C. The observed differential quantitative and qualitative protein and gene expression between DM and ND patients could provide insight into the observed differences in post-operative inflammatory state between the two patient populations.
This prospective cohort study was initiated after being approved by the Beth Israel Deaconess Medical Center Institutional Review Board/Committee on Clinical Investigations. Before enrollment, informed written consent was obtained from all patients interested in participating in the study. The enrolled 14 patients were scheduled either for elective or urgent primary CABG, or valvular surgery (aortic; mitral; or both). Since all diabetic patients considered for the study had elevated preoperative blood sugar levels ≥ 240mg/dL, they underwent preoperative treatment using either an insulin infusion or subcutaneous treatment. Normoglycemic non-diabetic patients received no further glucose treatment pre-surgery.
Under general anesthesia, midline sternotomy was performed, and systemic heparin (100 U/kg) was administered to attain an ACT > 400. The right atrium and ascending aorta were cannulated, and CPB was initiated. CPB circuit consisted of a standard roller pump, a membrane oxygenator, and a 40-μm arterial filter. The circuit was primed with crystalloid solution. CPB strategy included the use of mild hypothermia (30°C - 34°C), maintenance of serum glucose levels ≤ 130 mg/dL with intravenous insulin injections, and α-stat pH monitoring. During CPB, insulin infusion was utilized to maintain normoglycemia, and this strategy was continued in the intensive care unit for 24 hrs. Subsequently, subcutaneous insulin was utilized for postoperative glucose control. Pump blood flow was maintained between 2 to 2.4 L/min/m2 of body surface area, and mean arterial pressures were maintained between 50 mm Hg and 90 mm Hg with the use of conventional vasoactive medications. Cell-Saver and cardiotomy suction were utilized. The decision to administer anti-fibrinolytic agents (either Aprotinin or Amicar) was left to the discretion of the operating team. Cardiac arrest was induced and maintained with antegrade administration of cold-blood hyperkalemic (25 mmol/L) cardioplegia solution.
Blood samples were collected in the sterile PAX tubes (Qiagen Inc, Valencia, CA) from the central venous line after induction of anesthesia but before skin incision pre-operatively (pre), and 6-hours (6h), and 4-days (4d) post-CPB/C. Blood samples were centrifuged immediately at 15,000g for 15 minutes, and serum/plasma samples were frozen at −80°C until the time of the assay. RNA extraction was performed using PAX gene kit (Qiagen Inc, Valencia, CA). For microarray analysis, after the quantitative and qualitative assessment of extracted total RNA, single-stranded followed by double-stranded cDNA synthesis was performed. Biotin labeled cRNA was obtained by in-vitro transcription of double stranded cDNA using AFFI kit (Affymetrix, Santa Clara, CA). cRNA was further purified, fragmented, hybridized overnight onto Affymetrix gene chips, washed in streptavidin followed by array scanning as previously described (22).
Transcriptional profiling was performed on HG-U133 plus 2.0 Affymetrix chips containing > 47,000 transcripts. The quality of the chips was assessed and the outliers calculated using dChip algorithm was eliminated from the analysis. Matrix of gene expression values was generated using Robust Multichip average (RMA) normalization statistics (John's Hopkins University, Baltimore, MD). The presence calls of transcripts obtained from Microarray system 5 (MAS5) in conjunction with gene expression values from RMA was used to generate a fusion model to compare the differences in gene regulation post-surgery (6h or 4d) compared to pre- cardiac surgery (pre) in DM vs. ND patients (n=7 patients each for DM and ND). Fold-change of > 2-fold along with enhanced expression of that transcript in 50% or more patients post- cardiac surgery for up-regulated genes or pre-surgery for down-regulated genes were the major criterions to determine significant differential gene regulation between the subsets (p<0.001).
For real-time PCR analysis, one step RT-PCR reactions were performed using SYBR Green detection in the final volume of 25μl. The specific primers for the genes of interest were designed using OligoPerfect™ software (Invitrogen Inc., Carlsbad, CA). SYBR green reactions were performed with SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) with added reverse transcriptase and RNAase inhibitor using ABI Prism™ 7000 Sequence detector system (Applied Biosystems). The following conditions were used to run the reaction: 30 min at 48°C, 10 min at 95°C for 1 cycle, and 15 sec at 95°C and 1 min at 60°C for 40 cycles. Relative expression levels of the genes were analyzed using 2-ddCT method in comparison to the expression of the endogenous control gene GAPDH (Invitrogen). Each sample was analyzed in duplicates.
A multiplex florescent-bead immunoassay was performed to detect cytokine concentrations from the serum samples (n=7 patients each, DM and ND) diluted 1:1 in buffer provided with the kit (Invitrogen). Cytokines were analyzed in duplicates. Standard curves were generated for each cytokine by using the reference cytokine concentrations supplied by the manufacturer. Raw data (mean fluorescent intensity) were analyzed by Beadview Software (Upstate Inc., Charlottesville,VA) to obtain concentration values.
RMA statistical software was used as a measure to calculate fold-change for analyzing differential gene expression by microarray. Values of p<0.001 were considered significant. For cytokine comparison analysis, one-way ANOVA in combination with standard T-test (Sigma plot software, St. Louis, MI) was used to calculate significance. P-values < 0.05 were considered significant.
There were no other significant differences between DM and ND patients in terms of demographics and perioperative data except for diabetes, hyperglycemia, and hemoglobin A1C (HbA1C) (*p<0.001) (Table 1). All patients in the study underwent successful cardiac-surgery and recovery (no re-operations or deaths). Clinical outcomes showed significant differences for length of hospital stay post-CPB/C (*p=0.04) and weight gain (*p=0.001) but not for atrial fibrillation and sternal wound infections between the two patient populations (Table 1). No statistically significant difference in postoperative diuretic, blood product, or intravenous fluid administration was detected between the groups. However, the average fluid balance on the day of surgery was 3562 ± 725.1 and 3582 ± 1042 (p=0.99) and on POD4 was +600 ± 1027.8 and -108 ± 903.2 (p=0.6) respectively for DM vs. ND patients. The fluid balance is a crude estimate of fluid homeostasis as the exact measurement of fluid gain and loss through bowel movements, food intake, and insensible losses is difficult, and weight gain more accurately represents the cumulative fluid retention.
VEGF a multi-functional cytokine, known to be involved in inflammatory responses and weight gain was significantly increased following CPB when measured at 6h post-CPB/C only in DM patients (p<0.05). The expression of VEGF tends to be higher even after 4d post-surgery when compared to pre-surgery only in DM patients, but was not significant. In the case of ND patients the variation in VEGF expression at 6h or 4d post-CPB/C compared to pre-surgery were minimal (figure 1A). The relative difference in VEGF expression (pre vs. 6h post-CPB/C) was significantly higher in DM patients when compared to ND patients (p<0.02). There was a trend towards higher relative rise of VEGF expression even at 4d post-CPB/C in DM patients compared to ND patients, but this trend was not statistically significant (figure 1B).
HGF another multi-functional cytokine, known to be involved in inflammation and hyper-vascular permeability was enhanced significantly 6h post-CPB/C in both DM and ND patients (p<0.05). Also the expression levels were significantly higher (p<0.01) in DM patients as compared to ND patients 6h post-CPB/C. The expression of HGF tends to be higher than pre-surgery even after 4d post-CPB/C in both the patient populations but with no statistical significance (figure 2A). Similar to VEGF expression, the relative difference in HGF expression (pre vs. 6h post-CPB/C) was significantly higher in DM patients (p<0.02) when compared to ND patients (figure 2B). The relative rise between DM and ND patients was transient and returned to baseline levels when assessed at 4d post-CPB/C (figure 2B). No significant differences in expression of other growth factors such as EGF and FGF was observed between DM and ND patients at all time-points tested pre-surgery compared to post-CPB (data not shown).
In correlation with VEGF and HGF expression, HIF1α, a crucial regulator of hypoxia signaling, and other HIF1α related transcription factors such as EP300, and CREB were significantly up-regulated greater than two-fold only in DM patients (p<0.001) 4d post-CPB/C (figure 3). Validation of these exclusively up-regulated genes in DM patients involved in hypoxia signaling pathway was confirmed by quantitative real-time PCR analysis (figure 4). In contrast, the above hypoxia related genes were not up/down-regulated in ND patients. The genes involved in downstream VEGF signaling such as KRAS, and SOS were also significantly up-regulated only in DM patients (data not shown). SHC, an adaptor protein known to mediate HIF1α induction under hypoxic conditions was down-regulated only in ND patients (data not shown). IGF1R was significantly expressed exclusively in DM patients (*p<0.001); 4d post-CPB surgery (data not shown).
CRP, a marker of acute inflammation and a contributor to inflammatory diseases including arteriosclerosis as well as other cardiovascular diseases was elevated to different levels post-CPB surgery in DM and ND patients. CRP expression was significantly increased following CPB when measured at 6h post-CPB/C only in DM patients (p<0.05) (figure 5A). The relative expression of CRP was significantly higher in DM patients when compared to ND patients at 6h post-CPB/C (figure 5B). CRP serum levels persistently increased post-CPB/C and were significantly enhanced at 4d in both DM and ND patients (figure 5). No positive correlation was observed between VEGF and CRP at any tested time points in both the patient populations.
In this study, the serum VEGF levels increased significantly in DM compared to ND patients at 6h post-CPB/C. Furthermore, the relative increase in VEGF and HGF levels pre vs 6h post-CPB/C was significantly higher in DM patients when compared to ND patients. The results of the current study are primarily a reflection of poorly controlled diabetic patients, and further investigation is required to extend the results of this study to patients with better preoperative glucose control.
Although the levels of VEGF were not significantly different between pre and 4d post-cardiac surgery, a trend towards higher VEGF levels was observed at least until 4d post-CPB/C in DM patients. Several studies reported up-regulation of VEGF and HGF expression under hypoxic conditions in many disease settings including diabetic retinopathy (7). HGF is also known to function as a potent inducer of VEGF expression by human epidermal keratinocytes (23) and its receptor flk1 in endothelial cells (24). VEGF and HGF growth factors were extensively studied in relation to diabetic retinopathy and are reported as major contributors to retinal edema.
VEGF is a multi-functional cytokine known to be involved in endothelial cell proliferation (16), chemotaxis of macrophages and vascular smooth muscle cells (17), and induction of adhesion molecules (VCAM, and ICAM) (18) and metalloproteinase synthesis (19). Acute systemic hypoxia can induce VEGF expression, which regulates leukocyte extravasation into tissues followed by fluid accumulation and thus induces overall weight gain in patients after cardiac surgery. In this study, clinical outcomes such as length of hospitalization (10 vs. 6 days post-surgery; p=0.04) and weight gain (5 kg vs. 2.5 kg 4d post-surgery; p=0.001) were significantly greater for DM patients compared to ND patients which may be a consequence of VEGF higher and sustained serum levels in the DM population. Pre-operative glucose control and/or restoration to a normal HbA1C may ameliorate the effects of cardiopulmonary bypass upon postoperative recovery. In conjunction to VEGF protein expression levels, CRP, a well studied marker for systemic inflammation, was also expressed at higher levels in DM patients compared to ND patients. CRP could serve as another potential marker for weight gain in DM patients. Consideration must be given to strategies that reduce the inflammatory response in diabetic patients undergoing cardiac surgery. Such strategies include pharmacological manipulation and performance of off pump coronary bypass in appropriate patients.
In correlation with VEGF cytokine levels, HIF1α, a crucial regulator of hypoxia signaling pathway, and thus a potent stimulator of VEGF, was significantly increased only in DM patients 4d post-CPB/C. CREB and EP300 that act as co-activators of HIF-1α, were also significantly expressed only in DM patients 4d post-CPB/C. UBE2, a ubiquitin protein known to catalyze hydroxylated human HIF1-α protein and other ubiquitin proteins (Ub) was up-regulated only in DM patients (25). Some of the crucial genes that are involved in the downstream signaling of VEGF such as KRAS and its activator gene, SOS were also up-regulated only in DM patients 4d post-CPB/C. VEGF signaling induces several genes including VEGF that acts in an autocrine manner (26). In contrast, the genes regulated in ND patients (such as SHC, and PKHR) were down-regulated, thus negatively impacting the VEGF signaling pathway. Taken together, DM patients had an overall increase in regulation of the HIF-1α pathway at multiple levels.
In addition, IGF1R that is known to induce VEGF expression by positively regulating HIF1-α expression was also significantly expressed only in DM patients 4d post-CPB/C (27). Therefore, the differential expression of growth factors and their signaling related genes between DM vs. ND patients could contribute to the difference in observed clinical outcomes in the two patient populations.
There are some limitations to our study. Firstly, the study was performed with few patient samples from each group (DM and ND patients; n=7 each). Based upon an effect size of 1.2 and power of 0.8, we estimated a sample size of 12 in each group. However, given the large differences noted between the diabetic vs. non diabetic groups, the differences between the two groups were significant at a sample size of 7 in each group. This study is underpowered to detect an effect size less than 1.5. Secondly, more time points are needed to be analyzed to get a better understanding of growth factor functions post-cardiac surgery in DM and ND patients. Previous studies suggest a peak regulation of several inflammatory and apoptotic genes within 6h post-CPB. Therefore 6h post-surgery was considered as a time point of interest. Since the inflammatory response is expected to abate by 4d post-surgery, gene regulation was assessed at 4d post-CPB. Another limitation of the study is that while there was no significant difference in the duration of CPB between DM and ND patients, there was a trend towards longer bypass time in the DM patients.
Although our study implies the possibility that VEGF may have some role in developing edema and fluid retention in DM patients, it could be difficult to predict if increases in plasma VEGF adversely influence weight gain in DM patients or could be useful for therapeutic angiogenesis for ischemic artery disease, which is often seen in type 2 diabetic patients (28). However, the systemic edema and weight gain may partially be attributed to VEGF function along with the expression of other inflammatory markers such as CRP. Further studies may be required to clarify these speculations. In conclusion, we hypothesize that the differential expression of VEGF and HGF and up-regulation of the hypoxia regulated master gene HIF-1α in DM vs. ND patients may reflect more severe post-surgical complications in DM patients.
We would like to acknowledge the genomics core facility at the Beth Israel Deaconess Medical Center www.bidmcgenomics.org for their assistance and expertise in the gene array analysis. Dr. Manoj Bhasin has performed the statistical analysis for the microarray data.
Funding sources: This project was funded by NIH HL46716 grant and Irving Bard Memorial Fellowship. Dr. Emani was supported by NIH grant T32 HL07917-07 to Dr. Bruce Furie.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.