RBC aggregation in dextran solutions can be measured by flow cytometry

1 Introduction

Red blood cells (RBCs) tend to stick face-to-face with each other and form the so-called “rouleaux” under low shear stress. This is known as RBC aggregation [14]. RBC aggregation plays important role in determining blood flow in vivo, though there are still extensive debates about its effects on hemodynamics. On the one hand, since RBC aggregation is one of the determinants for blood viscosity in low shear rate, increased flow resistance in microcirculation was found to follow enhanced RBC aggregation [5]. On the other hand, studies on organ blood flow revealed that moderate RBC aggregation induce decrease in flow resistance [11]. Similarly, with a simple theoretical model, RBC aggregation was found to reduce blood flow resistance [18]. Furthermore, changes in RBC aggregation have been linked to diseases (e.g., diabetes mellitus, unstable angina, acute myocardial infarction) [1, 19]. Thus, RBC aggregation continues to be of interest in hemorheology and hemodynamics.

Among many investigations, the impact of macromolecules on RBC aggregation have been recognized early and received considerable attention. As endogeneous plasma protein, fibrinogen was found to be the dominant contributor to RBC aggregation [2, 25]. The association of plasma proteins and RBC aggregation have witnessed applications in diagnostic and treatment [21, 28]. At the same time, correlations between artificial polymers and RBC aggregation have been thoroughly studied for the following objectives:

First, understanding effects of plasma expanders and related mechanisms. Plasma expanders are often administered in large volumes during fluid therapy. Alterations in hemorheology caused by plasma expanders are linked to changes in microcirculation and tissue oxygenation. Dextran, a widely used plasma expander in the past, has been used in many studies discussing models of “Bridging” or “Depletion” for RBC aggregation [10, 24]. Rheologic effects of hydroxyethyl starch (HES) and gelatin, which are still in widespread use nowadays, have been studied in hemorrhagic shock and hemodilution. Gelatin was found to facilitate RBC aggregation [3, 12].

Second, improving the biocompatibility of biomaterials. For example, star-shaped cholic acid based PEG polymers with molecular weight between 10 and 16 kDa were found to reduce RBC aggregation [15].

Third, evaluation of components in preservation solutions for organ transplants. High molecular weight HES, as one of the components of the University of Wisconsin (UW) solution, showed a potent hyperaggregating effect. This may accounts for stasis of blood and inadequate washout of organs for transplantation [20].

Although RBC aggregation has been studied since early 1900’s, present measurement methods still have limitations and need improvement. Microscopic observation provides direct assessment of RBC aggregation, distinguishing single cells, doublets, triplets and bigger aggregates (linear or branched rouleaux). However, hematocrit (Hct), lower than 0.6% is necessary to avoid unrecognizable rouleaux formation at high Hct [26]. This accounts for the discrepancy between in vitro and in vivo study, because Hct in vivo is usually higher than 35%. Erythrocyte sedimentation rate (ESR) reflects suspension stability of the blood and correlates with RBC aggregation [13], while ESR data are affected by viscosity of suspension, and high viscosity retards erythrocyte sedimentation. Therefore, samples with similar viscosity are needed in ESR detection. Light-transmission and light-reflectance are conventional photometric methods to measure RBC aggregation. The measurement principle is based on recording changes of optical signal immediately after dispersion of pre-existing aggregates. Amplitude (AMP), half-time (t_1/2), and aggregation index (AI) are indices to describe the process of RBC aggregation [8]. RBC aggregation in a horizontal glass tube was studied using the time course of light transmittance, impedance, capacitance and electrical conductance after blood flow was abruptly stopped. The impedance data failed to agree with aggregation index, whereas the capacitance data in general correlate with aggregation index [7]. Electrical conductance was found to be sensitive to changes other than RBC aggregation [6]. These photometric and electrical methods provide precise time course of RBC aggregation but lack size distribution of aggregates as obtained by microscopic observation [1]. Usually, 0.5–1.5 mL sample volumes are needed in ESR and photometric methods. When the impact of macromolecules on RBC aggregation are investigated as mentioned above, different concentrations or molecular weight are included, and hence 10 mL or more blood are requiredin all.

In order to overcome the limitations of above-mentioned methods, a new method with flow cytometry was used to measure RBC aggregation in the present study. RBC fixed by glutaraldehyde exhibit strong auto fluorescence, which can be detected by flow cytometry. Single cells, doublets, triplets and larger aggregates can be distinguished quantitatively and quickly due to the correlation between auto fluorescence intensity and number of RBC per measured event. This method allows Hct of 35% or higher in the sample preparation. Less than 0.2 mL blood volume is needed for a multiple measurement of a certain sample.

2 Materials and methods

3 Results and discussion

Since solutions of Dx500 are approximate Newtonian fluids, similar viscosities were obtained when shear rates changed from 0.01 s^–1 to 1000 s^–1. Viscosities at the shear rate of 1000 s^–1 are shown in Table 1.

Since samples were fixed by GA, their auto fluorescence intensity was used to distinguish between single cells, doublets, triplets and bigger aggregates. As shown in Fig. 2A and B, colors of pink, blue, green and grey were used to represent single cells, doublets, triplets and bigger aggregates, respectively. The FSC-A to SSC-A dot plots (Fig. 2A) show the distribution of 10,000 analyzed events, including single RBC and aggregates, according to their size and morphology. Both single and aggregated RBCs are included in the gate. Figure 2A shows examples of FSC-A to SSC-A dot plots in the absence (PBS) and presence (4% Dx500) of RBC aggregates. Large aggregates were induced with the addition of Dx500 at a concentration of 4%.

The histograms (Fig. 2B) show the fluorescence distribution of the analyzed events in samples with different concentration of Dx500 in the FITC-channel. The fluorescence distribution varies in the different samples in accordance to the different Dx500 concentrations. The histogram of the sample in PBS has a normal distribution with a single narrow peak. In contrast, the samples in Dx500 have broad histograms with at least three recognizable peaks. Moreover, the maximum fluorescence intensities of the second and the third peaks are proportional to the maximum intensity of the first peak by factors 2 and 3, respectively. Take the histogram of 4% Dx500 as an example, the fluorescence intensities of the first, the second and the third peaks are about 1300, 2600 and 3900, respectively. For this reason, we considered the first peak corresponding to single RBC, the second to doublets, and the third to triplets. All events with higher fluorescence intensities are considered as “bigger aggregates”.

Following this definition, a quantitative comparison of the RBC aggregation caused by different Dx500 concentration was performed and displayed in Fig. 2C. Percentages of all aggregates and bigger aggregates were two parameters used to represent aggregation index. As shown in Fig. 2C, with the increase in concentration of Dx500, both parameters increased significantly at concentration of 2%, 4% and 6% (p < 0.01), while decreased when the concentration of Dx500 reached 8% and 10% (p < 0.01). The percentage of bigger aggregates in concentration of 4% was higher than that in 2% and 6% (p < 0.05).

The data of FACS was confirmed by microscopic observation as shown in Fig. 3. Auto fluorescence due to GA-fixation existed both in single cells and in aggregates. The images of samples in PBS confirmed the absence of RBC aggregates. Consistent with the data obtained by flow cytometry (Fig. 2B-C), the samples with 4% Dx500 showed more aggregates than that in PBS and 10% Dx500. Moreover, 4% Dx500 induced large-size RBC aggregates with linear rouleaux and irregular geometry. Only small and less RBC aggregates were observed in samples with 10% Dx500.

The results are in good agreement with the study performed by Bäumler H et al. Their study was conducted in a Couette system with light back scattering technique and revealed that the strongest aggregation happened when the concentration of Dx500 was 3% to 4% [9]. In other study, dextran also exerted byphasic effect on RBC aggregation. RBC aggregation was induced when the concentration of Dx500 was 0.5% and 2%, and RBC began to disaggregate when the concentration of Dx500 became 5% [30]. Therefore, compared to the literature, the method in the present study detects variation of RBC aggregation in a reasonable way.

In addition to the method mentioned above, there are detailed explanations as follows: (1) RBC suspension with Hct of 70% and polymer solutions were mixed in eaqual volumes to obtain a Hct of 35% in the mixture. Varied Hct, concentration and volume are allowed according to different aims. (2) Vortex was utilized for an adequate mix between RBCs and polymers. At the same time, pre-existing RBC aggregates was dispersed by vortex, similar to the stirrer in conventional photometric methods [8], or shear stress provided in shearing microscopy technique [17]. (3) Since the process of RBC aggregation is time dependent, 15 min in this study, longer than 2 min recorded in photometric methods [22], allowed a steady state of aggregation. (4) Usually, it took more than 2 hours to fix cells by GA [16]. In this study, to obtain an adequate and quick fixation, a 24-well plate was used, which allowed a homogenous fixation. When eppendorf tubes were used, many red cells concentrated at the tip and lose the chance to react with fixative solution. (5) RBC aggregates induced by Dx500 are reversible and concentration dependent. To avoid disaggregation during fixation, Dx500 in the same concentration as the original aggregation-inducing solution were added in GA solution. In the meantime, avoidance of shaking the samples helped to maintain RBC aggregates. (6) Auto fluorescence of GA got similar histograms in FITC, PE and APC channels. Distribution in FITC histogram was quantitatively analyzed.

The size range of RBC aggregates measured in this study should also be mentioned. According to the technical specification, the flow cytometer utilized in the present study can measure particles with a maximum diameter of 50 μm. Since typical human erythrocyte has a disk diameter of approximately 6.2–8.2 μm and a thickness at the thickest point of 2–2.5 μm, RBC aggregates of linear rouleaux with cell number less than 20–25 could be detected in principle. Ami RB et al. set the size range of 2–4 RBC/aggregate, 5–32 RBC/aggregate, and 33 or more RBC/aggregate as small, medium and large aggregates, respectively [1]. Hence, the method in the present study is available to measure small and medium RBC aggregates, but not the large aggregates. Since the formation of small and medium aggregates reflect the tendency of RBC aggregation, the method in the present study will be useful to determine RBC aggregation in polymer solutions.

4 Conclusion

The method in the present study provide additional advantages to the conventional measurement of RBC aggregation. It gets the distribution of single cells, doublets, triplets and bigger aggregates as derived from the microscopic observation, but it can keep Hct as high as 40%, which is unavailable for microscopic observation. Moreover, it gets distribution of cell populations quantitively in an easier and quicker way compared to the image anlysis for microscopic observation. It uses sample volume as 1/5∼1/10 as needed in ESR and photometric methods. It also avoids infuences by the viscosity of suspension which affects the data in ESR detection.

Based on the advantages of this method, the possible applications are as follows: (1) In the investigations about the impact of macromolecules on RBC aggregation, this method helps to evaluate hemocompatibility of biomaterials and to understand rheological effect of plasma expanders or preservation solutions for organ transplantation. (2) To detect irreversible RBC aggregates in blood samples. Unlike RBC aggregation induced by polymers in an aggregation-dispersion reversible process, sometimes RBC aggregation is irreversible in pathological state. For example, irreversible RBC aggregates are induced by methemoglobin in malaria [4]. (3) This method utilizes GA to maintain RBC aggregates, and to obtain fluorescence for FACS detection as well. In a broad way, it can be used to detect cluster of other cells. For example, in the study of metastasis, the cluster of cancer cells in blood or lymph fluid of animal models can be detected by this method.