Abstract
In static or low-flow conditions erythrocytes form linear or three-dimensional aggregates with characteristic face-to-face morphology, similar to a stack of coins, often called rouleaux formation. This aggregation is reversible and shear dependent (i.e. dispersed at high shear and reformed at low shear or stasis) and caused by a variety of macromolecules present in the blood plasma. The plasma protein fibrinogen is the major plasma component promoting red blood cell (RBC) aggregation in blood, with an almost linear relationship between aggregate size and plasma fibrinogen concentration. However, other plasma proteins are also reported to increase RBC aggregation, e.g. α2-macroglobulin, immunoglobulin M or G. In addition, there is evidence, that plasma lipids like cholesterol or triglyceride may influence the aggregation of erythrocytes.
In this study we evaluated whether there is an independent influence of proteins and lipids on the RBC aggregation. Using a regression analysis, we analyzed the correlation between the fibrinogen-, α2-macrogobulin-, immunoglobulin M-, Antithrombin III-, Protein C-, Factor VIII-, total cholesterol- and triglyceride concentration with RBC aggregation in blood samples from 2717 apparently healthy subjects or patients.
An univariate analysis showed, that the only variable which correlates on a biologically relevant level is fibrinogen (r = 0.46). The multiple correlation coefficient corresponded to rmult = 0.589 what indicated that nearly 59% of the variation of the erythrocyte aggregation can be explained by the influencing factors used in this model. This clearly showed that there are additional factors which are involved in the process of erythrocyte aggregation and still are under discussion.
Introduction
Human erythrocytes are highly specialized anucleate cells that are involved in oxygen transport to tissues. Under high shear conditions they are monodisperse but tend to aggregate in the presence of certain macromolecules at low shear or during stasis. In static or low-flow conditions erythrocytes form loose linear or three-dimensional aggregates with characteristic face-to-face morphology, similar to a stack of coins, often called rouleaux formation. This aggregation is reversible and shear dependent (i.e. dispersed at high shear and reformed at low shear or stasis) and caused by a variety of macromolecules present in the blood plasma.
Red blood cell (RBC) aggregation affects low shear blood viscosity and can adversely affect cell distribution and flow dynamics in the microcirculation (see Fig. 1, [1–8]). The extent of RBC aggregation is reported to be the main determinant of low shear blood viscosity, thus predicting an inverse relationship between aggregation and in vivo blood flow. However, the effects of aggregation on hemodynamic mechanisms (e.g., plasma skimming, Fåhraeus Effect, microvascular hematocrit) are very complex and may promote rather than impede vascular blood flow especially in venules and veins [9].
Erythrocyte aggregates (black arrow) in venules of around 35 μm diameter in the conjunctiva bulbi of a patient with diabetes mellitus.
The level of aggregation can rise enormously in association with a wide variety of clinical conditions such as sepsis [10], diabetes [11, 12], myocardial ischemia [13] and renal failure [14]. However, in spite of the huge literature dealing with RBC aggregation, there are still many areas where additional information is needed.
The concentrations of high-molecular weight, fibrous biomolecules in the suspending medium are important determinants of aggregation. The protein fibrinogen is the major plasma component promoting RBC aggregation in blood, with an almost linear relationship between aggregate size and plasma fibrinogen concentration [6]. Also, other plasma proteins are described to increase the RBC aggregation, e.g. α2-macroglobulin, immunoglobulin M or G [15, 16]. In addition, there is evidence, that plasma lipids like cholesterol or triglyceride may influence the aggregation of erythrocytes.
While there are different in vitro studies using single proteins or lipid solutions [16–20], an analysis of the influence of plasma proteins and lipids on the RBC aggregation from a huge population is lacking. To clarify whether there is an independent influence of proteins and lipids on RBC aggregation we analyzed the correlation between the fibrinogen-, α2-macrogobulin-, immunoglobulin M-, Antithrombin III-, Protein C-, Factor VIII-, total cholesterol- and triglyceride concentration with RBC aggregation in 2,717 subjects.
Subjects
Between 1991 and 2008 2,717 subjects (aged: 16–85 years, 475 females and 452 males) were examined in a two-center study. 2,125 subjects were included by the Department of Clinical Hemostasiology and Transfusion Medicine and 592 in the Praxisklinik Herz und Gefäße Dresden. All subjects were Caucasians. Cardiovascular risk factors were recorded, including obesity (body mass index[BMI] >30 kg/m2), current tobacco use (>1 cigarette/day), hypertension (systolic blood pressure >130 mmHg, diastolic blood pressure >90 mmHg), hyperlipidemia (total cholesterol >220 mg/dl and/or triglycerides >175 mg/dl) or fasting glucose (>126 mg/dl).
Measuring methods
Venous blood was drawn with minimum stasis in a sitting position from the cubital vein in the morning after 12 hours fasting. In accordance with the new guidelines for hemorheological laboratory techniques [21], the hemorheological parameters were determined within two hours of sampling to avoid deterioration of the rheological properties.
Erythrocyte aggregation was measured using an erythrocyte aggregometer (Myrenne GmbH, Germany) at stasis after adjusting the hematocrit to 45% with autologous plasma [22]. Details of the measurement method and reference range were described earlier [23]. The variation coefficient is 5.2% in sequence and 7.1% from day to day.
Proteins and lipids were measured according to standard laboratory tests [24].
Statistical analysis
All samples are described with arithmetic mean and standard deviation; classified data with percentage frequency. First, a linear regression analysis for all variables (fibrinogen-, α2-macrogobulin-, immunoglobulin M-, Antithrombin III-, Protein C-, Factor VIII-, total cholesterol- and triglyceride concentration was performed; the fitted regression line with 95% confidence bands is drawn in the dotted cloud (scattergram). In addition, a multivariate regression analysis showed the impact of the different proteins and lipoproteins in the same model. Probabilities smaller than 0.05 were evaluated as significant.
Results and discussion
Data from 2,717 subjects (f: 1,378; m: 1,339) were included; among them 282 apparently healthy subjects and 342 patients with peripheral arterial occlusive disease. The other patients suffered from venous disease, coronary artery disease, arrhythmias, kidney disease, cancer, diabetes mellitus, lipid metabolism disorders, hypertension, etc. The mean age was 51.8±7.2 years.
Table 1 shows the results of a univariate regression analysis for proteins (fibrinogen, α2-macrogobulin, immunoglobulin M, FactorVIII, Antithrombin III, and protein C) and lipoproteins (total cholesterol and triglycerides).
Univariate regression analysis between the erythrocyte aggregation index and different proteins (fibrinogen, α2-macrogobulin, immunoglobulin M, FactorVIII, Antithrombin III, and protein C) and lipoproteins (total cholesterol and triglycerides)
The only variable which correlated on a biologically relevant level was fibrinogen with a correlation coefficient of r = 0.46. This is in line with earlier studies in pure fibrinogen solutions [25, 26], which showed an increase of erythrocyte aggregation with increasing fibrinogen concentration. Also, an association between the fibrinogen concentration in blood plasma and the erythrocyte concentration was reported in different studies; especially in patients with vascular disease and with metabolic or inflammatory disorders [27–31]).
Figure 2 shows the correlation between the erythrocyte aggregation and the fibrinogen concentration for the whole group (open blue dots) and additionally for apparently healthy subjects (red squares) and patients with peripheral arterial occlusive disease (filled blue dots).
Correlation between erythrocyte aggregation SEA [–] and the fibrinogen concentration FIB [g/l] for the whole group (open blue dots) and additionally for apparently healthy subjects (red squares) and patients with peripheral arterial occlusive disease (filled blue dots).
To compare the influence of the different variables (proteins and lipoproteins) on the erythrocyte aggregation, a multivariate regression analysis of data from 2,400 patients was performed (see Table 2).
Partial correlation coefficients for the different variables included in the regression model
The multiple correlation coefficient corresponded to rmult = 0.589 what indicated that nearly 59% of the variation of the erythrocyte aggregation can be explained by the influencing factors used in this model. This clearly showed that there are additional factors which are involved in the process of erythrocyte aggregation.
The bridging model proposed by Shu Chien describes that RBC aggregation occurs when bridging forces due to the adsorption of macromolecules onto adjacent cell surfaces exceed disaggregating forces due to electrostatic repulsion, membrane strain and mechanical shearing [32–35]. In addition, there is now increasing experimental evidence indicating that RBC cellular properties can also markedly affect aggregation, with the term “RBC aggregability” coined to describe the cell’s intrinsic tendency to aggregate [35]. RBC membrane surface properties and structure, such as surface charge and the ability of macromolecules to penetrate the membrane glycocalyx (as recently shown by Franke et al. [35]), can greatly affect aggregation for cells suspended in a defined medium [37, 38]. Additionally, an increased intracellular free Ca2+ concentration in RBCs has been proposed as a trigger for erythrocyte aggregation [32]. Erythrocyte aggregation could be increased twice, when doubling the intracellular Ca2+-ion concentration.
Such cellular factors might explain the remaining 41% variability of the erythrocyte aggregation which are missing in this bridging model.