Flow dynamics of human sickle erythrocytes in the mesenteric microcirculation of the exchange-transfused rat

MICROVASCULAR

RESEARCH

34,

152-167 (1987)

Flow Dynamics of Human Sickle Erythrocytes Mesenteric Microcirculation of the Exchange-Transfused Rat JOSEPH KURANTSIN-MILLS,’

Departments Washington,

HELENA LAWRENCE

M. JACOBS, PANPIT P. KLUG,* S. LESSIN

of Medicine and Physiology, The George D.C. 20037, and *Department of Internal Medical School, Amarillo,

Received

in the

April

AND

Washington University Medical Center, Medicine, Texas Technical University Texas 79106

3, 1986

To analyze the microvascular rheology of sickle cells in an intact animal model, rats were isovolemically exchange transfused with human normal (hemoglobin AA) or sickle (hemoglobin SS) erythrocytes (blood group 0) or autologous red cells under ambient conditions, and the effects of the heterologous or autologous cells on (a) hemodynamics and respiration, (b) blood gases, and (c) acid-base status of the recipients were determined. Exchange transfusion of rats with autologous red cells or hemoglobin AA or hemoglobin SS erythrocytes was associated with stable mean arterial blood pressure, central venous pressure, respiration rate, blood pH, pC02, and p0, during the experimental period, except for tachycardia among the group of rats that received HbSS cells. Arteriovenous oxygen content varied among the three groups of animals, but, nonetheless, suggested adequate tissue oxygen supply under the conditions of the study. Acid-base status also was similar in the three groups of rats, The exchange-transfused rats were utilized to investigate the flow dynamics of red cells in the mesenteric microcirculation by applying intravital microscopy. Time-averaged velocities of the autologous red cells in 16- to 30-pm (id) vessels ranged from 1.07 to 1.25 mm/set in single unbranched arterioles with varying flux and wall shear rates. Time-averaged velocities of the HbAA cells in single 15 to 35pm arterioles ranged from 1.16 to 1.24 mm/set with wall shear rates similar to the estimates for the autologous cells. For both rat and human HbAA RBCs, the flow dynamics were indicative of normal shear-dependent and deformability characteristics of the cells under the flow conditions. Sickle cells exhibited time-averaged velocities of 0.384 to 0.452 mm/set, lower wall shear rates in lo- to 35-pm single unbranched arterioles, and three times less volumetric flux. In some arterioles, sickle cells with high axial ratio and low deformability showed definite adhesion to the endothelial surface, residing at such sites for several seconds until dislodged by the force of flow. Within single unbranched vessels or at microvascular bifurcations, sickle elliptocytes and sickle echinocytes with low deformability and high axial ratio obstructed flow and exhibited residence times of 2 to 88 set, thereby causing stasis. These data illustrate the microvascular flow behavior of sickle cells and demonstrate the rheological disequilibrium state that can result as sickle ceils course through successive segments of the microcircu~ation. 0 1987 Academic PKSS, IIIC.

’ To whom correspondence should be addressed, at Department of Physiology, The George Washington University Medical Center, 2300 Eye Street NW, Washington, DC 20037. 152 00X-2862/87 $3.00 Copyright Q 1987 by Academic Press, Inc. All rights of reproduction in any form reserved. Printed in U.S.A.

MICROVASCULAR

FLOW

OF SICKLE

CELLS

153

INTRODUCTION Sickle cell anemia is clinically characterized with vasoocclusive events that probably result from the abnormal microvascular rheology of the red cells. The intracellular polymerization of deoxygenated hemoglobin S results in the exponential increase in the internal viscosity of the sickle cells (Chien, 1977), concurrent with membrane changes (Lessin er al., 1978) and discocyte-sickle transformation. As a result, there is a heterogeneous subpopulation of red cells in sickle blood (Weems and Lessin, 1984) that alters the rheological features of the blood in the microcirculation. Experimental models for the study of flow characteristics of sickle cells in the microcirculation are relatively few (LaCelle, 1977; Klug and Lessin, 1977; Kaul et al., 1983; Lipowsky et al., 1982). The rat, exchange transfused with normal or sickle erythrocytes, may serve as a useful model for the short-term intravital microscopic study of the microvascular rheology of these cells and also for evaluating potential antisickling drugs. To explore this possibility, we exchange transfuse rats with autologous red cells or human normal (HbAA) or human sickle (HbSS) erythrocytes, and measure the hemodynamic and respiratory responses, acid-base status, and oxygen content of the blood. We also utilize this exchange-transfused rat model and employ intravital videomicroscopy to quantify the flow dynamics of autologous blood or normal or sickle erythrocytes in the mesenteric microcirculation. To this end, we measure red cell flow velocity in vessels of varying luminal diameter, and estimate the shear rate, microvascular hematocrit, red cell flux, and vasoocclusion by sickled cells.

MATERIALS

AND METHODS

Preparation of animals for exchange-transfusion. Male Sprague-Dawley rats weighing about 175-200 g were fasted overnight (14-16 hr) and then anesthetized with intraperitoneal pentobarbital sodium (50 mg/kg). Polyethylene catheters were inserted into the left common carotid artery for measurement of systolic and diastolic blood pressures, and via the jugular vein to the right atrium for central venous pressures, using Gould P50 pressure transducers. The femoral vein was also cannulated for the exchange transfusion, and for sampling blood for measurement of partial pressure of oxygen (PO,), partial pressure of carbon dioxide (PCO,), and pH. Heart rate was determined from the electrocardiogram which was monitored by external leads (AVL or AVR). These hemodynamic parameters were recorded continuously on a Beckman multichannel polygraph (Beckman Model 411). Body temperature was maintained at 37-38” by a heating plate. Serological cross-matching of rat and human erythrocytes. Prior to the exchange transfusion, the donor’s (human) and recipient’s (rat) erythrocytes were crossmatched for ABO and Rh antigens using direct anti-globulin (Coombs) test to detect serologic incompatibility to unexpected antibodies that might result in in vivo sensitization and hence agglutination of the red cells (Widmann, 1985). In further serological testing experiments, affinity-purified goat anti-globulin (2 mg/ml in 10 mM phosphate-buffered saline, pH 8.0) for rat immunoglobulin (heavy and light IgG, Calbiochem, San Diego, CA) was utilized to determine specific rat anti-IgG sensitization of human red cells. The cross-reactivity of this anti-globulin

154

KURANTSIN-MILLS

ET

AL.

for human IgG was less than 2%. Only human group 0 erythrocytes did not show sensitization and agglutination for the rat plasma, or with the purified antiglobulin for rat IgG. Hence, all subsequent exchange transfusions were done with only group 0 erythrocytes suspended in rat plasma. Experimental protocol and the exchange transfusion. The experimental protocol consisted of measurement of the hemodynamics, respiration, blood gases, acidbase status, and hematological profiles of rats under pentobarbital anesthesia. The exchange-transfusion protocol consisted of three groups of rats. Among the group I rats (n = 4), autologous exchange transfusion was carried out to obtain 80-100% exchange. Group II rats (n = 6) were exchange transfused with human (group 0) normal hemoglobin erythrocytes resuspended in rat plasma. Group III rats (n = 5) were exchange transfused with human (group 0) sickle erythrocytes resuspended in rat plasma. The donor’s blood from normal healthy subjects or sickle cell patients was centrifuged (300g x 15 min, 24”) to separate the plasma from the red cells. The red cells were washed three times by suspension in phosphate-buffered saline (PBS) of the following composition (in mM): NaCl, 145.5; KCl, 4.0; Na,HPO,, 1.65; KH2P04, 0.16; n-glucose, 11.1; bovine serum albumin 0.5%; pH 7.4; 295300 mOsm/kg. After the final wash, the human erythrocytes were resuspended in rat plasma to a final hematocrit of 40%. To ensure that the sickle hemoglobin was fully oxygenated or in the liganded conformation prior to the exchange transfusion, the cells were treated briefly with humidified 95% 0,/5% CO, mixture. The control cells were similarly treated before using for the exchange transfusion. The rat plasma was obtained from a similar strain of rat bled previously under light pentobarbital anesthesia for that purpose. The exchange transfusion was carried out by withdrawing 1 ml of the rat blood and replacing it with 1 ml of the donor’s erythrocytes in rat plasma. The transfusion was done slowly over a 2-min period, and repeated at 2- to 3-min intervals until by the end of the process 80-100% of donor cells were infused into the recipient rats. A sample of carotid artery blood examined by sodium metabisulfite test showed that about 80-100% of the cells were of sickle type in the cases where animals were exchange transfused with sickle cells. Blood gases and acid-base. Blood samples were drawn anerobically from catheters inserted in the carotid artery and jugular vein using al-ml tuberculin syringe before, during, and after the exchange transfusion at 15- to 20-min intervals for measurement of blood gases and acid-base status. Blood PO,, pCO,, and pH of the arterial and venous samples were analyzed as soon as possible (usually within 20 min) using a Radiometer ABL30 Acid-Base Analyzer at 37” (Radiometer, Copenhagen) which provides a profile of the blood gases, oxygen saturation of hemoglobin, oxygen content of blood, HC03- concentration, and total CO,. Prior to the analysis, the blood samples were sealed tightly and kept on ice. The instrument was calibrated with ultrapure gas standards and appropriate buffer standards (Radiometer) before use. Blood hematocrit was determined by centrifuging aliquots in a microhematocrit centrifuge (Clay-Adams, Parsippany, NJ) and measuring the proportion of packed red cells. Preparation of the mesenteric microvasculature. After each exchange transfusion of the rat, a mesenteric loop was exteriorized through a midline incision and suitably draped over a hollow plastic chamber attached to a stage support on

MICROVASCULAR

FLOW

OF SICKLE

CELLS

15s

which the rat lies on a heating plate. Care was exercised to avoid any stretch of the mesenteric preparation according to the recommendations of Zweifach (1973). The plastic chamber was positioned between the substage condenser and objective of a Zeiss compound microscope adapted for intravital use in our laboratory. The outer corridor of the chamber is circulated with warm water that maintains the inner chamber supporting the mesenteric loop at 37.5”. The surrounding tissue was supported with cotton gauze and, together with the exposed area, suffused with a modified Hepes-physiological salt solution (composition in mM: NaCl, 118; KCl, 5.9; CaCl,.H,O, 2.5; MgS04.7Hz0, 0.5; glucose, 10; Hepes, 12) and 1% gelatin so as to maintain tissue suffusate at osmotic equilibrium. The pH of the buffer was 7.45 and the osmolality about 300 mOsm/kg. Three experimental protocols were used for the study of the mesenteric microcirculation. The first was the observation, study, and video recording, by intravital microscopy, of the mesenteric microcirculation of the rat exchange transfused with autologous blood. The second protocol involved similar observations, study of video recordings, and measurements of microvascular flow dynamics in the rat exchange transfused with normal human erythrocytes suspended in rat plasma. The third protocol involved similar study as the second except that the rats were exchange transfused with human sickle erythrocytes. A total of 15 rats were studied successfully; 4 received autologous blood, 6 received HbAA cells, and 5 received HbSS cells. The exposed mesentery was transilluminated from a low-voltage illuminator and observed with a Zeiss Plan-Neofluar immersion objective (40: 1, 0.9 nA) and a 10 x eyepiece. The microcirculatory network was scanned and studied to isolate microvascular units that were not crowded with adipocytes. The microvascular flow of red cells in vessels with various luminar diameters was tape recorded via a Sony video camera (Model AVC 3200) connected to the binocular Zeiss microscope. These data were then analyzed off line to obtain information on microvessel luminar diameters; microvessel length; red cell type, namely discocytes, echinocytes, and irreversibly sickled cells (ISCs); flow velocity of the cells: red cell flux; and sickled cell residence time during vasoocclusion. Determination of rheologicalparameters. Intravascular red cell velocities were measured by a variation of the dual-slit photometric technique of Wayland and Johnson (1967) as applied on line with the self-tracking cross-correlator (Tompkins et a/., 1974) (Instrumentation for Physiology and Medicine, San Diego, CA). This device was calibrated to provide a direct output of red cell velocity along the vessel center line (Vrbc). A motorized rotating plastic disc (IPM, San Diego, CA) with a monolayer of red cells as velocity target was placed in the focal plane of the microscope and used to calibrate the velocity cross-correlator. The velocity of the disc was held at a steady velocity after each change in the output of the voltage motor tachometer. The error associated with these velocity measurements was about 5%. The output voltage of the cross-correlator (V,) was recorded as a function of the steady velocity of the rotating disc (V,) from 0.1 to 2.0 mm/set. Analysis of a total of 20 separate multiple measurements by least squares gave a linear regression equation V, = 0.004 + 1.019 V,, with a correlation coefficient of 0.993. The slope of this line served as a calibration factor for the correlator output voltage

156

KURANTSIN-MILLS

ET

AL.

with the units, millimeter per second per volt, and was established for the Zeiss microscope setup and the magnification used in these measurements. Intravascular volumetric flow (Q) rates based on the average velocity of cells plus plasma (V,,,,) or bulk velocity of blood were calculated from the correction factor, V, = 1.6 V,,,,, (Baker and Wayland, 1974) and vessel diameter, D (Q of this factor has been demonstrated by = vma” rr 02/4). The applicability Lipowsky and Zweifach (1978). The measured velocity was recorded at least three times and averaged for a minimum of 60 set after a stable correlation of photometric signals was established. The data recorded on a Beckman multichannel polygraph (Beckman 411) were digitized and mean values computed as a function of time using an IBM-XT computer. Arteriolar wall shear rate was estimated based on a Newtonian definition, (8V,,,,,/D) set- ‘, and served as a useful rheological index to compare the different experimental preparations for each of the animals. Viscosity of O-45% hematocrit of the erythrocyte suspension was measured by means of a cone and plate viscometer (Model LVTDCP, Brookfield Engineering Corp., Stoughton, MA) at shear rates between 2.30 and 230 see-’ as described previously (Kurantsin-Mills and Lessin, 1982). Measurement of vascular diameters and lengths. The video image was displayed on a television monitor at an effective magnification of 380 x . The diameters of vessels were measured by means of a videomicrometer (Colorado Video, Inc., Boulder, CO) noting a voltage output that was utilized as a reference for the separation of two video lines positioned over the TV image of the inner walls of the vessels. Such measurements were made by three independent observers focusing on the inner walls of the vessels. The average of the three interpretations of the observers was taken as mean diameter and the averaged error of these different measurements ranged from 5 to 10%. The recording system was calibrated against a stage micrometer device and had an accuracy of +0.5 to 1 pm, as described previously by Gorczynski et al. (1978). The length of vessels was measured using the same system. The length (L) and width (w) of sickled cells were also measured using the videomicrometer for the calculation of axial ratio L/W. Measurement of intravascular hematocrit. Microvascular hematocrit was determined by the method suggested by Barbee and Cokelet (1971) and applied to vessels in situ by Klitzman and Duling (1979). Briefly described, the method assumes that microvessels are cylindrical and that the mean erythrocytes volume is known and remains constant throughout the peripheral circulation. Furthermore, it is assumed that there is free movement of plasma and cells throughout the vessel. The latter assumption may be incorrect for sickled cells that exhibit adhesion to the endothelium (Hebbel et al., 1982) or block flow for several seconds. Notwithstanding the latter exception, it follows from these assumptions that hematocrit

(%) = (nV x 100)/[7~ x (O/2)2 x Ll

where n = number of erythrocytes in the vessel segment, V = mean cell volume (pm3), D = diameter of the vessel, and L = length of the vessel segment with IZerythrocytes. This method of estimating microvessel hematocrit has been shown to correlate with the optical density method with a correlation coefficient of 0.95 (Kanzow et al., 1982). Red cell flux in the microvessel was calculated as the

MICROVASCULAR

FLOW

OF SICKLE

CELLS

157

product of the blood flow and the microvascular hematocrit (Lipowsky et al., 1980). Erythrocyte residence time during vasoocculsion by sickle cells at microvascular bifurcations was determined from the video time code on line with the system and during the slow-motion video playback. This parameter was measured on individual red cells from the time of occlusion until dislodged by the force of flow resulting from some segment of the network. Data analysis. All data are presented as means + SD. Student’s t test for unpaired data was used to compare differences between the groups. One-way analysis of variance with the Student-Newman-Keuls multiple range test was used for comparison within groups (Snedecor and Cochran, 1967). Differences were considered to be statistically significant at P < 0.05 or 0.1 depending on the data as shown in the text. RESULTS Hemodynamic

Responses

and Acid-Base

Balance

Exchange transfusion of rats under pentobarbital anesthesia with autologous erythrocytes did not significantly alter the systolic and diastolic pressures and the heart rate at the end of the procedure. During the exchange transfusion, withdrawal of 1 ml of blood resulted in a brief and transient hypotension of about lo-15% below the normal mean arterial pressure, but there was no significant change in heart rate. Norma1 blood pressure was restored immediately during the reinfusion of 1 ml of autologous blood. Right atria1 pressures reflecting central venous pressure averaged 5.4 + 1.4 mm Hg. The rate of respiration of the anesthetized rats was 72 t 4 cycles/min before the exchange transfusion and 74 + 4 cycles/min after the autologous exchange transfusion. These cardiopulmonary parameters were not significantly different from the basal values measured in nontransfused rats (P > 0.1; Table 1). When human normal or sickle erythrocytes (group 0) suspended in rat plasma were exchange transfused into the rats, the hemodynamic responses were stable among the individual animals. The animals that received HbAA cells responded to the exchange transfusion with no significant change in the heart rate, and also had stable systolic and diastolic arterial pressures. The rate of respiration remained stable at about 70-75 cycles/min. However, there was tachycardia among the group of rats that received HbSS cells compared with those that received HbAA cells or autologous blood. The difference in heart rates was statistically significant (P < 0.05). The details of the results of the animals successfully exchange transfused with the heterologous erythrocytes are summarized in Table 1 over approximately 25 cardiac and respiratory cycles. The data shown in this table are the average values ( + SD) of hemodynamic responses, blood gases, and acidbase status of rats that were successfully exchange transfused 80-100% with the human red cells. The standard deviation reflects the variation in responses for each parameter during the experiment among individual animals. Successful exchange transfusion of rats with autologous or heterologous erythrocytes did not alter the acid-base status of the blood. The arterial and venous p0, values were higher for these spontaneously breathing rats because the blood samples were treated with 95% 02/5% CO;, to ensure complete oxygenation of

DATA,

ACID-BASE

ml)

ml)

STATUS,

AND

a a v v

Pre Post Pre Post

a ”

V

a

V

a

V

a

V

a

V

a

V

a

BLOOD

GASES

OF RATS

TABLE

1

103.3 330 5.4 14 7.35 7.30 38.5 41.1 20.9 22.5 22.2 23.8 119.8 47.2 98.9 56.8 21.0 10.9 11.7 38.7 28.2 37.3 24.2

AUTOLOGOUS

-r- 3.5 + 30 f 1.4 f 4 * 0.05 -r- 0.04 + 2.8 t 4.2 + 1.9 -c 1.0 + 2.5 k 1.0 2 5.50 2 2.5 zt 2.0 -r- 5.8 2 1.0 k 2.1 k 2.3 f 1.46 2 0.20 2 0.31 f 0.19

Rat RBC (n = 4)

FOLLOWING

AND

ISOVOLEMIC

116 329 5.1 71 7.36 7.34 40.5 43.8 23.4 23.7 24.6 25.0 100.5 42.1 96.8 69.8 20.5 14.6 5.9 39.3 32.6 35.2 30.4

27 k I? k3 f 2 e 2 2 k k iz k f f f f f f k 2 2 k 0.04 0.04 3.8 5.0 3.4 2.6 3.5 2.1 7.5 3.7 1.6 9.2 0.4 1.9 1.8** 2.5 4.2 4.5 2.1

24 1.2

HbAA red cells (n = 6)

HETEROLOCXUS

Note. These data are averaged results of measurements done at 15- to 20-min intervals during the exchange transfusion. a, arterial; v, venous. * Groups I and II vs group III, P < 0.05. ** Group I vs groups II and III, P < 0.05.

CaO, - CvO, (ml/l00 Hematocrit (%)

O2 content (ml/l00

Sat Hb02 (%)

~0~ (mmHg)

HCO-J

NO2 (mmI-M

PH

Mean arterial BP (mm Hg) Heart rate (beats/min) CVP (mm Hg) Respiration rate (min-‘)

HEMODYNAMIC

+ 10 + 54* 2 1.1 24 f 0.04 t 0.03 f 2.8 -c 4.0 k 1.4 k 1.6 f 1.5 k 1.6 2 1.9 + 6.2 f 0.7 k 5.1 k 0.2 f 1.3 f 1.2** k 5.8 k 3.4 + 5.1 -+ 4.5

HbSS red cells (n = 5)

TRANSFUSION

112 382 5.3 71 1.36 1.35 39.4 41.4 21.4 24.3 23.3 25.2 110 41.4 98.2 67.5 20.8 13.8 1.0 36.9 29.8 38.5 31.9

EXCHANGE

MICROVASCULAR

FLOW

OF SICKLE

CELLS

159

the sickle cells (Table 1). In some rats, the hematocrit of arterial and venous blood decreased by about 8-10% at the end of the exchange as a result of infusion of heparinized PBS to prevent clogging of blood in the femoral catheter. Under the conditions used in these experiments, oxygen saturation of both normal HbAA and HbSS were within a similar range. However, the arteriovenous oxygen content difference of the rats exchange transfused with the heterologous red cells was significantly lower than that of the control group (P < 0.05). In preliminary experiments in which the human cells were washed, resuspended in PBS containing 0.5% albumin, and used for the exchange transfusion, we consistently observed that the systolic and diastolic blood pressures, ~0, and pH, declined rapidly (within 20-30 min) accompanied by a rise in the pC0, (data not shown). Microvascular Flow of Erythrocytes Measurement of the microvascular flow of the red cells was performed under standardized conditions to determine and compare the flow dynamics of autologous rat cells, normal human erythrocytes, and sickle cells. Time-averaged red cell velocities obtained for arterioles of the rat mesenteric microcirculation through which rat cells, human normal, or human sickle red cells are flowing are shown in Fig. 1. The luminal diameters of these vessels varied from 10 to 35 pm. The average red cell velocities in these arterioles for the three types of cells studied shows the expected fluctuations of microvascular flow for the 60 set of recorded time. The velocities of control rat red cells flowing through lo- to 30-pm arterioles ranged from 1.07 -t 0.07 to 1.35 + 0.09 mm/set. Velocities of human HbAA erythrocytes flowing through arterioles of similar diameter ranged from 1.16 + 0.06 to 1.24 2 0.07 mm/set. These velocity values were not significantly different (P > 0.05). The flow velocities of human sickle cells averaged 0.384 + 0.028 to 0.452 + 0.041 mm/set in 12- to 35pm arterioles and were significantly different from those of rat and HbAA cells (P < 0.1). The hematocrit in arterioles (15-30 pm) ranged from 12 to 15% in the normal flow state for rats exchange transfused with autologous red cells. In the presence of normal human red cells, the hematocrit in 16- to 35-pm arterioles was 7-19%, and in the sickle-cell exchange-transfused rats, the hematocrit in arterioles (16 32 ,um) varied from 3 to 16% (Table 2). All estimated microvessel hematrocrits were lower than the systemic hematocrits, as also reported by Kanzow et al. (1982).

The volumetric flux (a function of both the erythrocyte velocity and the microvascular hematocrit) of the autologous and heterologous erythrocytes in arterioles of different luminal diameters was estimated for the three types of cells and these data are summarized along with other flow statistics in Table 2. There is a relative invariance in the volumetric flux of both the normal human cells and the rat red cells. On the other hand, sickle cells showed about 50-60% lower flux for arterioles of approximately similar diameter. This apparent invariance in bulk volumetric flow rate in small vessels (D s 30 ,um) has also been reported by others (Lipowsky and Zweifach, 1977; Lipowsky et al., 1980). Because of this relative invariance of red cell volumetric flux, estimation of the vessel wall shear rates based on the Newtonian definition, (8V,,,,/D) set-‘, showed vessels containing sickle cells (D 2 10 G 35 pm) as having significantly lower wall shear rates. As an illustrative case, the flow statistics in 25-pm arterioles in which the

160

KURANTSIN-MILLS

ET AL.

0.0 0.4 1 1.4

*---* -

1 O-15pm 1%20pm

*--0

253Opm arlerbbs

HbM

1

arterbbs wterbks

Erythrocytes

&--+16-l

7pm arterioles

-25pmterbles 0.6

8c

1

0.20

HbSS

Erythrocytes

6---d

12-l Bpm tieMe

*--a

20-24/msdriole8 26-3Spm~~ 10

20

30 rime. 8eamda

40

50

3 60

FIG. 1. Time-averaged velocities of normal autologous rat RBCs and HbAA and HbSS RBCs flowing through arterioles of the rat mesenteric microcirculation. Each point is the average of three to four measurements digitized every 5 set from correlator signal output. SD is 5-8’26, omitted for clarity. There are IO arterioles in each group of vessels shown.

three types of cells were flowing are compared in Table 2. The apparent viscosities of rat and human blood were also determined in vitro for hematocrits of O-45% over the range of shear rates estimated from the in vivo measurements. The apparent viscosity estimated from the in vitro measurements, using the hematocrit and shear rate data of the present report, showed that in the normal flow state

MICROVASCULAR

FLOW

OF SICKLE

TABLE 2 FLOW STATISTICS AND SOME RHEOLOGICAL PARAMETERS FOR RAT HUMAN NORMAL (HbSS) ERYTHR~CYTES IN 25-pm PRECAPILLARY ARTERIOLES OF THE MESENTERIC

Parameter Number of arterioles Diameter (pm) RBC velocity (V,, mm/set) Bulk mean velocity

Rat RBCs 19 2s t2 1.05 2 0.09** 0.66 2

HbAA

161

CELLS

RBCs

27 25 k-2 0.98 Itr 0.07

(HbAA) AND SICKLE MICROCIRCULATION

HbSS RBC’s 31 24 + 2 0.39 -c 0.02'

0.05**

0.61 k 0.04

0.24 "

O.Ol*

3.24 x 10m4**

3.00 x 10 J

1.09 x IO '*

8.4 -c 2.9**

10.1 -t 2.7

7.x

30.30 x 1om4

8.51 x IO '*

W,,,, mm/set) Bulk volumetric flow rate (Q = Vmea.d2/4 mm’/sec) Microvessel hematocrit Wmcro) Flux Qffmcromm3/sec) Fraction of systemic kmm-it W,,,,,IH,,,) Wall shear rate0 C3V,,,,lD set ‘) Apparent viscosity” (cP)

27.21 x lo-0.27 5

0.05

t

16**

211

1.9 k

0.27**

0.30 2 0.04 195

k 13

1.8 -' 0.24

-t

0.25 5 80 2.8

3.x

0.06

kS* 2

0.45

* HbAA RBCs vs HbSS RBCs, P < 0. I. ** Rat RBC vs HbAA, P > 0.1. ” Assumed a Newtonian definition (SV,,,,/D) as suggested by Lipowsky and Zweifach (1978). ’ Estimated from in vitro measurement using cone-plate viscometer, and based on the wall shear rates estimated in this study. Assumed 10% hematocrit (Lipowsky er al., 1980).

the viscosity of sickle blood at the estimated microvascular hematocrit is as high as 40% greater than that of normal blood (Table 2). The microvascular flow of sickle cells in arterioles of luminal diameter 15-35 pm was characterized by periodic interruptions due to the low deformability of cells with high axial (L/w) ratios and perhaps high hemoglobin concentrations, sickle-elliptical or sickle-echinocytic morphology (Fig. 2). To quantitate the effect of these cells on the flow dynamics, we have applied the concept of “residence time.” Residence time defines the time interval that a deformed sickle erythrocyte resides at a microvascular site (or is entrapped at a microvascular site), be it a single unbranched vessel or a vascular bifurcation, to interrupt the normal flow state to zero velocity and induce vasoocclusion and significant stasis. Based on this notion, we have applied the video time code electronically imprinted on the videotapes to determine the residence time of a number of sickle cells encountered in a variety of network flow patterns in the rat mesenteric microvasculature. Figure 2 illustrates an example of flow interruption by sickle cells with low deformability, high L/W ratios, and presumably high intracellular hemoglobin S polymer fraction. This diagram is an artist’s representation of actual flow events in the vessel shown at stasis. The diameters of the vessels, the geometry of the network at the site of the vasoocclusion, and the radial distribution of the cells are also shown. These cells, with abnormal morphology, usually occluded vessels with luminal diameters of 10 ,um or less although, depending on their spatial orientation in flow, they also occluded 12- to 15pm arterioles. The residence

162

KURANTSIN-MILLS

ET

AL.

FIG. 2. Representation of vasoocclusion induced in a 12-pm precapillary vessel at a microvascular bifurcation by sickled cells. Network geometry and radial distribution of cells at stasis are depicted as seen in the videotapes.

times of sickled cells entrapped at various bifurcations are shown in Fig. 3. These data are pooled from all the vessels with luminal diameters ranging from 10 to 25 pm having vasoocclusion induced by the three types of sickle cells. At microvascular bifurcation, the junctional angles varied widely (e.g., 24 to 125”, measured relative to the parent vessel as shown in Fig. 2). On several occasions, cells with apparently rigid appearance have been observed to flow and adhere to the endothelium or there may be sudden arrest of flow, thereby inducing local rheological disequilibrium.

DISCUSSION These studies demonstrate that the rat exchange transfused with the appropriate blood group of human sickle erythrocytes maintains adequate cardiovascular and respiratory functions to serve as an experimental model for the short-term investigation of the microvascular flow dynamics of these abnormal red cells. Our intent was to utilize this model to quantify the flow characteristics of sickle cells in a variety of arterioles and precapillaries under conditions where perfusion of the cells is maintained by normal systemic hemodynamics. In several experiments conducted on rats exchange transfused with autologous red cells or human normal or sickle erythrocytes, systemic hemodynamics, respiratory rate, and hematocrit were essentially the same for all the experimental animals. The hematocrit of some animals declined by about 8-10% after the exchange transfusion due to the colloid replacement with heparinized PBS containing 0.5% albumin to clear the femoral vein catheter and avoid clogging the line. However, the blood acidbase state was within the normal reported values for rats under pentobarbital anesthesia (Brun-Pascand et al., 1982). Kerns et al. (1981) also noted that the

MICROVASCULAR

FLOW OF SICKLE CELLS

163

.i . i. ..

.. . .. .

-

.. . .. .

. .

10-L

Sickle Discocytes

Sickle Echinocytes

Irreversible Sickle Celk

FIG. 3. Estimated residence times of three types of sickle cells during stasis and vasoocclusion in single unbranched vessels or at microvascular bifurcations.

physiologically stable rat model which they described for microvascular studies showed 8-10% decrease in hematocrit at the end of the 4-hr procedure. We believe that gassing the sickle cells with O2 + CO, mixture prior to the exchange transfusion may have contributed to the stability of the model for the completion of the experiments, The comparably higher arterial and mixed venous oxygen tensions measured among the individual animals throughout the exchange indicate adequate oxygen supply to tissues. Teisseire and Soulard (1984) have shown that in the pentobarbital-anesthetized normoxic rats, hemodynamic, acidbase balance, and oxygen transport characteristics following exchange transfusion with human red cells (containing Hb Cretiel) did not change. The authors supposedly achieved complete isovolemic exchange tranfusion in their experiments over a relatively short period. Our results compare reasonably well with their data. Marked differences were noted between the human normal and sickle erythrocytes in their microvascular flow dynamics. Autologous rat erythrocytes and normal human red cells exhibited similar time-averaged flow velocities in arterioles of approximately similar diameters with values comparable with those published by other investigators for the rat (Gaehtgens et al., 1976; Henrich et al., 1978). Therefore, compared with the autologous red cells, the normal human erythrocytes

164

KURANTSIN-MILLS

ET

AL.

exhibited no additional resistance due to size or deformability in traversing the rat microcirculation. On the other hand, time-averaged flow velocities of sickle cells measured in arterioles or precapillaries of diameters 10-35 pm were significantly lower than the velocities measured for the control cells. Other studies by Kaul et al. (1983) on the isolated mesoappendix infused with a bolus of sickle cells have also indicated that HbSS cells have 50% higher resistance compared with HbAA cells. No comparisons were made with autologous rat cells by these authors. Within the mesenteric microcirculation, Lipowsky and Zweifach (1978) have determined that the arteriovenous (AV) distribution of intravascular pressures declines most rapidly in the precapillary vessels of 15-35 pm diameter. This sudden change in pressure at the arteriolar and precapillary levels of the mesenteric microcirculation, attendant with the relative invariance of the red cell volumetric flux in the precapillaries (D s 30 pm) (Lipowsky et al., 1980) and the concurrent desaturation of oxyhemoglobin S, is likely to promote gelation of hemoglobin S in the sickle cells. The decreased affinity of deoxyhemoglobin S, together with other metabolic factors, would also enhance the polymerization process and further influence the flow velocity of sickle cells. During flow through arterioles and precapillaries of a wide range of luminal diameters, sickle cells show significantly lower wall shear rates and volumetric flux. The relatively greater heterogeneity of sickle cells compared with that of normal red cells (Weems and Lessin, 1984; Rodgers et al., 1985) will also influence their microvascular flow dynamics and distribution. The apparent reduced flux of sickle cells may be the results of numerous factors such as red cell shunting and plasma skimming at the pre- to postcapillary vessels, deformability of the cells which is related to polymer fractions within the cells, as well as nonuniform radial and network distribution of the cells. In some of the exchange-transfused rats, low flow velocities of the cells were attendant with larger numbers of plasma gaps in between the cells, flow intermittences at microvascular branchpoints, and transient or prolonged obstruction of flow by elliptical sickle cells with obviously high axial ratios (L/W). This accumulation of sickle cells in flow and their hydrodynamic separation effects in single unbranched arterioles and at microvascular bifurcations will probably affect hematocrit distribution within the network of arterioles, precapillaries, and capillaries (Johnson et al., 1971; Kanzow et al., 1982; Gaehtgens, 1984). Sickle elliptocytes and sickle echinocytes on occasion adhered to sites of unbranched arterioles and at bifurcations and induced clogging, resulting in a rheological disequilibrium in that segment of the network. These sickled cells have been shown to have acquired irreversible membrane impairment (Lessin et al., 1978) that results in abnormal microrheological behavior of the cells as reflected in their flow characteristics (Evans et al., 1984). As an illustrative case, we have applied the term residence time to quantify the interruption of flow by these cells with abnormal morphology that is discernible on the videotapes (Fig. 2). The residence times of sickle elliptocytes at precapillary bifurcations can be as high as 500 times those of sickle discocytes. Although we have determined the apparent viscosity from only in vitro measurements, the data suggest that under the experimental flow states, apparent viscosity may be of the order of 2-3 CP in 25-pm arterioles for the same hematocrit and at the estimated in vivo

MICROVASCULAR

FLOW

OF SICKLE

CELLS

165

shear rates for the three types of cells under study, as has been reported for the microcirculation of cat mesentery (Lipowsky et al., 1980) and rat cremaster muscle (Lipowsky et al., 1982). Sickle cells showed about 40% greater apparent viscosity when compared with the control cells. Under low flow states of sickle cells or in cases of microvascular occlusion by sickled or dense cells, the apparent viscosity could increase exponentially to 8-10 CP (Lipowsky et al., 1982) and thereby significantly increase the regional resistance. For smaller precapillaries and capillaries, theoretical analysis by Tozeren and Skalak (1978) suggests that apparent viscosity can increase substantially and that these increases are dependent on red cell deformabilty and hematocrit. The interruption of microvascular flow for extended periods during vasoocclusion could have several consequences on the rheological equilibrium and the homeostatic integrity of the microcirculation. These include (i) reduction in perfused capillaries, (ii) increase in regional resistance, and (iii) decrease in arteriolar blood flow, which will reduce oxygen supply and lead to an increased oxygen extraction and a lower venous oxygen saturation accompanied by, perhaps, enhanced hemoglobin S polymerization provided the critical kinetic requirements for gelation are satisfied (Fernone et al., 1980), and (iv) leukocyte adhesion and platelet aggregation. The rheological significance of leukocytes under low flow states has been emphasized in recent studies in myocardial ischemia in which leukocyte-induced arrest of microvascular flow correlated with increased leukocyte plugging of capillaries (Engler et al., 1983). Furthermore, recent studies from this laboratory using autologous indium-1 1 l-labeled platelets and radionuclide imaging have shown focal platelet accumulation at sites of pain in patients in sickle cell crises (Siegal et al., 1985). In conclusion, the sickle cell exchange-transfused rat provides a reasonably stable model for the analysis of the flow dynamics of these erythrocytes in the microcirculation. We fully recognize that more extensive experimentation is required to answer questions relating to the contributions of hemoglobin S polymer fraction in the cells, blood viscosity, and oxygen affinity of hemoglobin S to the AV distribution of pressure, flow, and hematocrit. The relationships of these and other factors to tissue oxygenation and hence the continuous flow or occlusion of sickle cells also requires further study. Within the scope of this study, it is clear that subpopulations of sickle cells can induce a rheological disequilibrium in the microcirculation of the rat mesentery such that the compensatory mechanisms may be overburdened. One of the main advantages of this model is its potential application to the study of the microvascular flow of sickle cells in different organs and microcirculatory beds as well as putative pharmacologic agents for the amelioration of microcirculatory disturbances in sickle cell disease.

ACKNOWLEDGMENTS This work was supported by NIH-NHLBI Grant 15160, a subcontract collaboration between The George Washington University and Howard University (J.K.-M. and L.S.L.). The authors thank Mariann Nelson, R.N., for assisting in obtaining blood from patients, and Linda Vaughan for typing the manuscript. The authors also appreciate the assistance of Dee Simons of the Blood Bank Laboratory, The George Washington University Hospital, in the serological assays. Finally, we thank the Biomedical Communications Department of GWUMC, especially Dewayne Coates for his expert assistance on the video engineering and Judy Gaunther for the artwork.

166

KURANTSIN-MILLS

ET

AL.

REFERENCES BAKER, M., AND WAYLAND, H. (1974). On-line volumetric flow rate and velocity profile measurement for blood in microvessels. Microvasc. Res. 7, 131-143. BARBEE, J. K. AND COKELET, G. R. (1971). Prediction of blood flow in tubes with diameters as small as 29 pm. Microvasc. Res. 3, 17-21. BRUN-PASCAND,M., GAUDEBOLJT,C., BLAYO, M. C., AND POCIDALO,J. J. (1982). Arterial blood gases and acid-base status in awake rats. Respir. Physiol. 48, 45-57. CHIEN, S. (1977). Rheology of sickle cells and erythrocyte content. Blood Cells 3, 283-303. DEAN, J. AND SCHECHTER,A. N. (1978). Sickle-cell anemia: Molecular and cellular bases of therapeutic approaches. New Engl. J. Med. 299, 752-763, 804-811, 863-879. ENGLER, R. I., SCHMIT-SCH~NBEIN, G. W., AND PAVELEC, R. S. (1983). Leukocyte capillary plugging in myocardial ischemia and reperfusion in the dog. Amer. J. Pathol. 111, 98-l 11. EVANS, E., MOHANDAS, N., AND LEUNG, A. (1984). Static and dynamic rigidities of normal and sickle erythrocytes. J. Clin. Invest. 73, 477-488. FERRONE,F. A., HOFTXHTER, J., SUNSHINE, H. R., AND EATON, W. A. (1980). Kinetic studies on photolysis-induced gelation of sickle hemoglobin suggest a new mechanism. Biophys. J. 32, 361380. GAEHTGENS, P. (1984). “Regulation” of capillary haematocrit. Intern. J. Microcirc. Clin. Exp. 3, 147-160. GAEHTGENS, P., BENNER, K. U., SCHICKENDANTZ,S., AND ALBRECHT, K. H. (1976). Method for simultaneous determination of red cell and plasma flow velocity in vitro and in vivo. Pjlugers Arch. 361, 191-195. GORCZYNSKI,R. J., KLITZMAN, B., AND DULING, B. R. (1978). Interrelations between contracting striated muscle and precapillary microvessels. Amer. J. Physiol. 235, H494-H504. HEBBEL, R. P., EATON, J. W., STEINBERG,M. W., AND WHITE, J. G. (1982). Erythrocyte/endothelial interactions in the pathogenesis of sickle-cell disease: A “real logical” assessment. Blood Cells 8, 163-173. HENRICH, H., HERTEL, R., AND ASSMANN, R. (1978). Structural differences in the mesentery microcirculation between normotensive and spontaneously hypertensive rats. Pflugers Arch. 375, 153159. JOHNSON,P. C., BLASCHKE, J., BURTON, K. S., AND DIAL, J. G. (1971). Influence of flow variations on capillary hematocrit in mesentery. Amer. J. Physiol. 221, 105-l 12. KANZOW, G., PRIES, A. R., AND GAEHTGENS, P. (1982). Analysis of the hematocrit distribution in the mesenteric microcirculation. Znt. J. Microcirc. Clin. Exp. 1, 67-79. KAUL, D. K., NAGEL, R. L., AND BAEZ, S. (1983). Pressure effects on the flow behavior of sickle (HbSS) red cells in isolated (ex vivo) microvascular system. Microvasc. Res. 26, 170-183. KERNS, L. L., SEIFERT, F. C., AND ESPOSITO,R. A. (1981). A physiologically stable rat model for in vivo studies. Microvasc. Res. 22, 228-231. KLITZMAN, B. M., AND DULING, B. R. (1979). Microvascular hematocrit and red cell flow in resting and contracting striated muscle. Amer. J. Physiol. 237, H481-H490. KLUG, P. P., AND LESSIN, L. S. (1977). Microvascular blood flow of sickled erythrocytes. Blood Cells

3, 273-282.

KURANTSIN-MILLS, J., AND LESSIN, L. S. (1982). Rheological assessment of the antisickling properties of pyridoxal and pyridoxine. Blood Cells 8, 315-328. LACELLE, P. L. (1977). Oxygen delivery to muscle cells during capillary occlusion by sickled erythrocytes. Blood

Cells

3, 273-282.

LESSIN, L. S., KURANTSIN-MILLS, J., WALLAS, C., AND WEEMS, H. (1978). Membrane alterations in irreversibly sickled cells: Hemoglobin-membrane interaction. J. Supramol. Struct. 9, 537-554. LIPOWSKY, H. H., USAMI, S., AND CHIEN, S. (1980). In vivo measurements of “apparent viscosity” and microvessel hematocrit in the mesentary of the cat. Microvasc. Res. 19, 297-319. LIFQWSKY, H. H., USAMI, S., AND CHIEN, S. (1982). Human SS red cell rheological behavior in the microcirculation of cremaster muscle. Blood Cells, 8,113-126. LIPOWSKY, H. H., AND ZWEIFACH, B. W. (1977). Methods for the simultaneous measurement of pressure differentials and flow in single unbranched vessels of the microcirculation for rheological studies. Microvasc. Res. 14, 345-361. LIPOWSKY, H. H., AND ZWEIFACH, B. W., (1978). Application of the two-slit photometric technique to the measurement of microvascular volumetric flow rates. Microvasc. Res. 15, 93-101.

MICROVASCULAR

FLOW

OF

SICKLE

CELLS

167

G. P.. SCHECHTER. A. N., AND NOGUCHI, C. T. (1985). Cell heterogeneity in sickle cell disease: Quantitation of the erythrocyte density profile. J. Lab. C/in. Med. 106, 30-37. SIEGEL, R. S., RAEL. J. L., KLEIN, D. L., RYAN, N. L., AND LESSIN, L. S. (1985). Platelet localization and kinetic characteristics in sickle cell crisis. BIood 66(5). 66a. SNEDECOR,G. W.. AND COCHRAN,W. G. (1967). “Statistical Methods.” The Iowa State Univ. Press. Ames. TEISSEIRE,B. P.. AND SOULARD, C. D. (1984). Pulmonary vasoconstrictor response to acute decrease in blood P<(,.J. Appl. Physiol. 56, 370-374. TOMPKINS W. R., MONTI, R., AND INTAGLIETTA. M. (1974). Velocity measurement of self-tracking correlator. Rev. Sci. Inst. 45, 647-649, TOZEREN. H., AND SKA~AK, R. (1978). The steady flow of closely-fitting incompressible elastic spheres in a tube. J. Fluid Mech. 87, l-16. WAYLAND, P. C.. AND JOHNSON, P. C. (1967). Erythrocytes velocity measurement in microvessels by a two-slit method. J. Appl. Physiol. 22, 333-337. WEEMS. H. B., AND LESSIN, L. S. (1984). Erythrocyte density distribution in sickle cell anemia. Ac,rcr RODGERS,

Huematol.

71, 361-370.

WIDMANN. F. K. (1985). The antiglobulin test. In “Technical Manual of the American Association of Blood Banks.” 9th ed., pp. 91-102. Arlington, VA. ZWFIFACH. B. W. (1973). Microcirculation in the intestinal mesentery. Mirrovusc. Res. 5, 363-367.