Density Ultracentrifugation of Sickle Cells During and After Pain Crisis: Increased Dense Echinocytes in Crisis

Transcription

1 Density Ultracentrifugation of Sickle Cells During and After Pain Crisis: creased Dense Echinocytes in Crisis By James A. Warth and Donald L. Rucknagel An increase in the number of irreversibly sickled cells (ISCs) in pain crisis has been found by some investigators but not others. We have used the technique of discontinuous arabinogalactan density-gradient ultracentrifugation of whole blood to study ISCs from patients with sickle cell anemia (SCA) during pain crisis and again when pain free (5-1 days after crisis). Nine patients have been studied through ten episodes of pain crisis. Five layers with densities from 1.18 g/ml to g/ml have been used. Careful classification of the cells using Nomarsky optics demonstrated highly significant changes occurring in the layers of the gradient. The changes involve the appearance of an increased percentage of echinocytic ISCs and echinocytic cells that were not lscs. especially in the denser gradient layers. during crisis. and their replacement by T HE IRREVERSIBLY SICKLED CELL (ISC) is being increasingly recognized as a pathogenic cell in sickle cell disease, effecting hemolytic rate and having been correlated with splenic infarction, conjunctival vessel abnormalities, and hyposthenunia in childrent However, its role in pain crisis, a vasoocclusive event, is uncertain. Diggs7 stated that the percentage of ISCs does not change significantly at the time of crisis. Banreras and Diggs,8 however, noted a downward trend in the percentage of sickle cells in venous blood as the pain of crisis subsided. Reversible sickle cells were not distinguished from ISCs in that study. Rieber et a19 noted that this downward trend at the time of crisis applied to ISCs alone. Three of their patients, observed during the pain-free state, showed lower percentages of ISCs than during crisis. They did not separate different classes of sickle erythrocytes by shape or density. Chien et al #{176} have shown that the viscosity of ISCs exceeds that of normal erythnocytes regardless of shear rate or hematocnit. They stated that the abnormal viscosity of oxygenated sickle cell blood can be largely attributed to the altered rheology of the ISC. A possible pathogenic role for this cell in erythrostasis and vasoocclusive episodes was implied. A method has been developed for recovering the relatively heavy ISCs that employs arabinogalactan (Stractan) density-gradient ultracentnifugation. It has been used to separate normal erythrocytes by density and to obtain enriched preparations of ISCs for study of their cation transport system. We have utilized the technique of arabinogalactan density-gradient ultracentrifugation, plus careful classification of the cells we obtained, for a study of nine adult patients with normal-appearing discocytes in the pain-free state. There was no change in ISCs that were not echinocytic. Data collected previously demonstrated that reduced glutsthione activity correlated with increased echinocytosis in our gradient layers. This indicates that the echinocytic change may occur as a result of oxidant stress. Hemoglobin F levels and the percentage of hemoglobin F from reticulocytes showed no consistent change to coincide with the rise in normal-appearing discocytes in the lightest layers after crisis. Our data indicate that pain crisis occurs in association with an echinocytic change, which may be induced by oxidant injury. The rise in normal-appearing cells after crisis may reflect increased hemoglobin F production in some patients but mainly relates to the disappearance of these echinocytic erythrocytes. classical sickle cell anemia during and after ten episodes of pain crisis. MATERIALS AND METHODS All patients in this study, unless otherwise noted, had sickle cell anemia with only S. A, and F hemoglobins. Whole blood HBA levels were less than.5%, while HbF levels ranged from 1.5% to 15.7%. The blood of nine patients, five men and four women, aged I 9 to 6 years, was studied in and out of pain crisis. To be considered in crisis, the patient had to have otherwise unexplained pain and tenderness for at least 8 hours involving two or more separate areas and requiring parenteral analgesics. Patients in pain crisis were studied between two and four days after the start of crisis, following at least 6 hours of hydration in the hospital. The out of crisis study was done 5 to 1 days after the end of crisis, generally after five pain-free days. One patient had mild pain remaining after only a five-day interval. Patients were not studied if they had been transfused in the prior months, if they had bacterial infection, or if there was doubt about their pain status. One patient was studied twice. At the time of study the patients had no major complication of their disease, eg, nephrotic syndrome or stroke. The patients were studied consecutively. However, the data from three patients were not analyzed. One had a borderline elevated hemoglobin A level and a low mean corpuscular volume (MCV), suggestive of probable Sfl#{176}- thalassemia. One was found to be taking vitamin E and another oral From the Harper-Grace Hospitals and the Department of Medicine, Wayne State University. Detroit. and the Departments of ternal Medicine and Human Genetics, University of Michigan, Ann Arbor. Supported by Sickle Cell Center Grant USPHS-HL and 5 MO! RR- and a grant from the Medical Staff Trust Fund of Harper Hospital. Submitted July /. /98; accepted March Address reprint requests to Dr James Warzh, Hematology Dlvision, New England Medical Center Hospital. I 7! Harrison Ave. Proger Building. Eighth Floor, Boston. MA 0! by Grune & Stratton, c !/8/ $0.OO/O Blood, Vol 6, No (August), 198: pp

2 508 WARTH AND RUCKNAGEL zinc in nonstandard formulations. Both substances have been reported to suppress ISC formation. Four patients were on fixed doses of oral zinc during our studies. Serial plasma zinc levels were stable. HbF and F Reticulocyte Studies Five additional patients were studied in and out of pain crisis. Reticulocyte and hemoglobin F stains were performed on the density-separated cells of one patient. The four others, two women and two men, were to 0 years old. Hemoglobin F was quantified by radial immunodiffusion assayt5 in the unseparated but carefully washed cells and in the density-separated layer 1. The percentage of hemoglobin F from reticulocytes was measured by pulse labeling these same layers with H-leucine. 6 The HbS and HbF were separated from dialyzed hemolysates using a high performance liquid chromatograph (Beckman I loa, Beckman strument Corp. Palo Alto, Calif) and SynChropak AX 00 column (SynChrom, Linden, lnd). 7 The relative proportions of HbS or HbF were determined by liquid scintillation spectroscopy (with internal correction for quenching); these proportions reflect hemoglobin synthesis in reticulocytes. Measurements in the layers other than layer 1 were not reliable because of the low number of reticulocytes and F cells in those layers. Controls One patient was studied while pain-free during an upper respiratory viral illness and again while pain-free after the illness to determine if the changes in erythrocytes were related to viral infection. Two patients were studied while pain-free, one on two consecutive days. Two hematologically normal people, one black and one white and two individuals with sickle cell anemia trait were also studied as controls. Medications During pain crisis patients were on medications not used when pain-free. Drug records obtained during crisis indicated no common medication among our nine patients. However, we incubated normal erythrocytes for one hour at #{176}Cwith the three most common medications used during crisis: hydromorphone hydrochloride, meperidine hydrochloride, and hydroxyzine pamoate. The respective concentrations used were 0.08 mg, I.088 mg, 0.5 mg per milliliter of whole blood. Each was ten times the estimated in vivo concentration. The cells were then washed twice with HEPES buffer, fixed, and counted blind under Nomarsky optics. Similarly, for a patient with sickle cell anemia, an incubation was carried out using the same drug concentrations and separate aliquots of sickle cells for each drug incubation. An arabinogalactan density-gradient ultracentrifugation was then performed for each incubated sample and for a control incubated with HEPES buffer. Hydration During pain crisis patients are hydrated. To determine the effect of hydration, three patients with sickle cell anemia, two men and one woman, aged to 9 years were admitted while pain-free to the Clinical Research Unit at the University of Michigan and hydrated for 1 hours at an average rateof 169 ml/h. Ofthis, 100 ml/h was intravenous 0.5% NaCI. The remainder was ad libitum oral fluids. A density gradient separation was performed at the end of the period of hydration and again I days later. Density Separation of Cells Density separation of cells was carried out as previously described.t8 brief, discontinuous arabinogalactan (Sigma Chemical Co. St Louis) density gradients were prepared using Rexyn I 00 (Fisher Scientific Co. Fair Lawn, NJ) and 10 mmol/l HEPES buffer. The densities (g/ml) were 1.158, 1.150, 1.1, 1.15, and The specific gravity, p1-i (7.), and milliosmolarity (95-00 mosm) were checked for each layer. Layer 1 was designated as the lightest layer (1.18). The degree and duration of pain crisis in SCA subjects was assessed by the authors using the above criteria. After informed consent was obtained, I 5 ml of venous blood was drawn, mixed with sodium heparin (15 U/mL), and oxygenated (humidified 95% 0 and 5% CO) at 7 #{176}C for 15 minutes. A complete blood count (CBC), indices (Coulter Counter, Model ZBI), and reticulocyte count were performed. The blood was immediately centrifuged at 65 g at #{176}C for ten minutes. The plasma was aspirated and the packed cells and buffy coat resuspended in isotonic 10 mmol/l HEPES buffer, to a hematocrit of 0%. dices, red blood cell count, and hematocrit were determined. A cell count was performed after using 1% gluteraldehyde-eagle s medium to fix the cells. One thousand cells were counted (the HEPES count). The above prepared gradients were layered into an ultracentrifuge tube, and 1.6 ml of the erythrocytes suspended in HEPES buffer to a hematocrit of 0% was placed on the top. Six such tubes were prepared. The tubes were spun in a Beckman L-50 ultracentrifuge at 1.5 x 105g forone hourat #{176}C in a ltrotor. Eachofthelayers was then removed and suspended in isotonic 10 mmol/l HEPES buffer and then washed twice using this buffer. The total volume of each layer was measured. One aliquot of each cell layer was prepared for Coulter Counter determination of erythrocyte indices, red blood cell count, and hematocrit. The percentage of recovery for the five layers was determined and the percentage in each layer was corrected for 100% recovery. One 0.0-mL aliquot of each layer was fixed in 1% gluteraldehyde-eagle s medium. Two hundred cells were counted from this aliquot. It took approximately seven hours from phlebotomy to fixing the cells. To avoid any artifact of fixation, we used a dilute solution (1 %) of gluteraldehyde-eagle s medium as recommended by Bessis and Weed. 9 The milliosmolarity (00 mosm) and ph (7.) were carefully maintained for this solution. We avoided any glass contact before fixation, since such contact can cause echinocytosis. 9 Each patient also served as their own control. Cell Classification All counting was done in a blind fashion by one observer using Nomarsky optics. The coefficient of variation for blind counting of a species of cell constituting 1% or more of the total cell count averaged 1% (1 ± 1%). Cells were classified (Fig 1) as normalappearing discocytes (D), echinocytes I, II, III, spherocytes, and spheroechinocytes (SE), following the criteria of Bessis.#{176}Irreversibly sickled cells were defined as cells at least twice as long as they were wide. An eyepiece micrometer was used to measure doubtful cells. Those ISCs with a smooth border and a central disc were called D Those with a ruffled border (at least three coves) or loss of the central disc or both were called Ia,. Those ISCs with at least two spicules not located at the ends were called SE Groups of cells analyzed were group I : total ISCs = D E SE; group : spiculated ISCs = I, 1SIE; g#{176}up: spiculated nonlscs = E11, E111, SE; and group : echinocytic forms = E SE E1, E,,, E,11,SE. Reticulocytes were stained with new methylene blue, and 1,000 cells counted. Hemoglobin F erythrocytes were stained with the Fetaldex (Ortho Diagnostics, Raritan, NJ) modification of the

3 ERYTHROCVTES IN PAIN CRISIS 509 Betke-Kleihauer technique and 500 cells counted. This latter stain provided the best fetal hemoglobin stain of cells we isolated on the density gradient. All counts were done blind. Generally, we loaded between 8 x l0 and x io erythrocytes per run. By determining the total percentage recovered and correcting both the in-crisis and out-of-crisis runs to 100% recovery and to an equal number of cells loaded, we were able to determine changes in the absolute number of a given cell type. RESULTS The Gradient Fig 1. Classification of cells.#{176}nomarsky optics x 500. (A) Normal appearing discocyte (with pit)-(d); (B) Echinocyte I- irregular disc (E,); (C) Echinocyte Il-flat with spicules (Es); (D) Echinocyte Ill-round to ovoid spicules (E5); (E) Spheroechinocyte-round. fine spicules (SE); (F) Spherocyte-round. smooth; (G) ISC-smooth, central disc preserved (la); (H) ISCruffled border (at least three coves) and/or loss of disc (li); (I) ISC-with two or more spicules (la). Axial ratio.0. Table 1 lists the pertinent clinical data. For a hematobogically normal person 97% to 99% of the cells loaded were recovered in layer 1 with 1% to % in layers and. For those patients with sickle cell anemia listed in Table 1, 9.6 ± 6.6% of the cells were recovered in layer 1 and the remainder in layers through 5 (Fig ). Layer 5 tended to have fewer cells than layer, but this was not invariable. The percentage of total cells loaded reaching a given layer did not change characteristically in crisis but was similar for the same person under varying clinical conditions. a normal person, the top layer contained 95% normal Table 1. Summary of Clinical Data for Nine Patients Studied Through Ten Episodes of Pain Crisis terval Patient (dl Sex Age (yr) A (%) F 1%) Hb (g/dl) Retic (x 10/L) WBC (x 1 O /L) 1 F crisis ofcrisis M crisis ofcrisis M.9 1. crisis of crisis #{19} 9 M lncrisis ofcrisis F crisis ofcrisis M 6..5 crisis ofcrisis F crisis ofcrisis M crisis of crisis F.1 6. crisis ofcrisis M crisis ofcrisis Paired t Test (P) <.01 NS NS Same person.

4 510 WARTH AND RUCKNAGEL episodes of pain crisis there was an increase in the hemoglobin level (P <.01) out ofcnisis, but no significant changes in the reticulocyte (percentage or absolute) or leukocyte counts were observed (Table 1 ). No patient was transfused. Changes in Cell Population Fig. Density gradients. Arabinogalactan density gradient ultracentrifugation of normal blood (left) and sickle cell anemia blood (right). erythrocytes, while the remaining two layers contained mostly spherocytes and echinocytic forms with less than 5% normal cells. The method concentrated ISCs in dense layers so that in SCA patients the top layer contained approximately % ISCs. Layers through 5 contained 0% to 0% ISCs. Seventy percent of the ISCs were spiculated. Discocytes and echinocytic forms made up the remainder. Total cell recovery averaged 98%. The percentage of recovery of all ISCs averaged 99% and of speculated ISCs 91% of those loaded onto the tube, ensuring that we were not generating ISC or spiculated ISC forms. Furthermore, each of the nine cell types was seen in fresh sickle cell blood, which was drawn and immediately oxygenated before fixing. The P0 at the bottom of our gradient after ultracentnifugation was approximately 1 50 torn. The two patients with sickle cell trait had cell-layer distributions similar to the two hematologically normal individuals. Less than % of the cells in layer or met the criteria for lscs in normal persons or individuals with sickle cell trait. dices For 1 8 runs the average MCV and mean corpuscular hemoglobin concentration (MCHC) in whole blood was ±.7 fl and.7 ±. g/dl, respectively. There was a significant, stepwise decrease in the MCV (P <.001) and an increase in the MCHC (P <.001) between the whole blood and layer 5 values, 76. ± 6.8 fl and 9. 1 ± 5.6 g/dl, respectively. These results were characteristic of the changes seen in size and MCHC when erythrocytes lost water and became more dense. There was no significant difference in the percentage changes in crisis v out ofcrisis for the MCV or MCHC, and no change between whole blood and the washed cells suspended in HEPES buffer. Only 18 runs were analyzed because for two runs not enough cells reached layer 5 to allow a determination of indices. For the nine patients investigated in and out of ten The changes found in the density-gradient separated layers as patients went from crisis to out of crisis are shown in Table. The changes were a fall in spiculated ISCs and echinocytic forms (groups and ), and a rise in normal appearing discocytes, which was most apparent in the mean for layers through 5, and in layers and S individually. Discocytes were significantly increased in all layers. No other single cell type demonstrated statistically significant changes in layer 1. On the other hand, a change in spheroechinocytes only occurred in layer 5 (a fall, P <.0). Group 1 (IDe E 1SE) demonstrated a mean change, a fall, for the five layers that was statistically significant (P <.05) even though no single layer reached significance (Table ). On closer inspection the changes in smooth ISC (ID) were not significant in any layer, while the spiculated ISC forms (group ) showed significant decreases in layers and 5. creased spiculation occurred in crisis in the ISC fraction, and, as seen in the change in group (E,,, E111, SE), in the non-isc fraction as well. No changes of significance were seen in the erythrocytes suspended in HEPES buffer that were loaded onto the ultracentrifuge tubes (unsepanated erythrocytes), although the same trends were apparent. Of interest, layer reflected many of the most significant changes observed (Table ), while in layer 5, discocytes, which made up 1.6 ± 9.0% of the cells in crisis and. ± I.0% of the cells out of crisis, showed a consistent rise of 16% or more of the in-crisis value. seven cases, the rise exceeded 50%. Figure depicts the changes in mean percentage for the four densest layers ( through 5) for each person in crisis and out of crisis. The consistency of the echinocytic change in crisis and the rise in normal appearing discocytes out of crisis can be seen. Patients and 10 are the same individual. The rise in the mean percentage of discocytes in layers 1 through 5 correlated directly with the number of days between the two studies (r =.66; P <.05) but this was not significant for layer I on 5 individually. Paired data analysis of the absolute number of discocytes out of crisis compared to in crisis showed a rise (P <.05) for all five layers. The discocytes that appear normal are, at least in the dense layers, abnormal cells as their increased density proves. Using the cells from eight density separations, stains for F erythrocytes and reticulocytes revealed less than

5 ERVTHROCVTES IN PAIN CRISIS 511 Table. Morphologic Changes in the Erythrocytes of Nine Patients Studied During Ten Episodes of Pain Crisis: Mean Percentage ± 1 SD and of Crisis Layer Cell Type Discocytes Group 1 Group Group Group 11 HEPES 79.5 ± ± 5.9. ±.8 5. ± ± ± ±.6.6 ±.0.6 ±. 1. ± 5.8 P. 1< <. j<. <.5 < ± ±.. ±.5 7. ± ± ± ±.1 1. ± ± ± 8. P 1 <.0 NS NS NS < ± ±.6.5 ± ± ± ± ± ± ± 1.. ± 19.1 P 1 <.0 NS NS NS I < ± ± ± ± ± ± ± ± ± ± 1. P 1 <.00 NS I <.0 1 <.0 < ± ± ± ± ± ± ± ± ± ± 0.8 P 1 <.006 NS NS NS I < ± ± ± ± ± ± ± ± ± ±.8 P 1 <.00 NS <.01 <.0 1 <.0 Mean -5.1 ± ± ± ± ± ± ± ± ± ± 1. P 1 < <.06 1 <.01 1 <.0 1 <.00 Arrow indicates direction of change in the cell type specified going from the in-crisis to the out-of-crisis state. P values shown ( t test, paired data). NS - Not significant. ttotal lscs: ID, I, 1a $$piculated lscs: 1. l. Spiculated non-iscs: E,1, E111,SE. IlEchinocvtic forms: E l, E,, E, E111.SE. % F cells and less than % reticulocytes in layers through 5. Layer 1 contained % to 1 % F cells and 10% to 6% reticulocytes. Controls We extended the ultracentnifuge spin to two hours without any significant effect on the percentage of ISCs, spiculated ISCs, discocytes, echinocytic forms, on SE. Similarly, adding glucose to all buffers and to the arabinogalactan was without effect on these cell populations. Three patients were studied when pain-free to determine if there were significant changes in the percentage of ISCs, spiculated ISCs, discocytes, echinocytic forms, or SE. The first was studied during an upper respiratory tract infection and again 68 days later, well after the infection had ended. There was no significant change in the above cell species. Two other patients were studied twice. One was studied hours after the initial study. The other was studied 1 5 days after the initial study. both cases there was no significant difference in any of these cell types. Medications The medications studied did not generate echinocytes, total ISCs, spiculated ISCs, or discocytes in a normal person. Mepenidine hydrochloride and hydroxyzine pamoate produced a cupping effect in cells compared with controls. a patient with sickle cell anemia whose erythrocytes were incubated, ultracentrifuged, and fixed, these three drugs produced no statistically significant change in any of the four cell types compared with controls (HEPES buffer). Since our cells were washed at least four times before fixation, alteration of morphology by a drug or other plasma component was unlikely. The four patients on zinc displayed the same changes in echinocytic cells and discocytes in layers through 5 and layer as the other patients. Suppression of ISCs by zinc was not evaluated.

6 51 WARTH AND z 61 5o 61 a. icui*iw lsc EHlNOCYTIC FOHt OISCOCYTES A IN OUT IN OUT 1 OUT 100 p<.oi p<.ol pool ioo 60 o) a.-.-. P D C P61166T P*T16NT o---o P6T1EJdT a.-. PATiENTS To 60 5O 0 I- z Table. Percent Recovery Corrected to 100% for Three Pain-Free Sickle Cell Anemia Patients (X. V. Z.) When Hydrated (H) and 1 Days Later (NH) Patient x Y Z Layer H NH H NH H NH 1 St Hb(WB) i _ 0 0.P < 0.01 by paired data analysis. tp < 0.01 for 1 pairs (three people, layers through 5). Hb(WB) = hemoglobin (g/dl) in whole blood. $PIO)LATED ec ECIeNOCYTiC FORMS DISCOCYTES B OUT IN OUT IN OUT 100 r < <.01 p < PAT#{18}NT6 P P6TIET8 0-#{19}--0 -a P P l reached significance ( P <.0 1 ) for layer (1.15 g/ml). No other layer demonstrated a significant change, but by comparing all 1 dense layers a statistically significant result was seen (Table ). Our patients in pain crisis were studied only after hydration. For the ten studies in and out of pain crisis there was no similar trend at the time of crisis towards greater cell recovery in layer or in the dense layers (Table 5) Sc a. \ 0 C a % \ 0 D C. \...,.-:i a \ a a #{19} Fig. Changes in mean percentage (vertical axis) for layers through S of the density gradient for each person going from in pain crisis to out of pain crisis. (A) Patients 1 through 5. (B) Patients 6 through 10. Patients and 10 are the same person. Spiculated lscs - l. l; Echinocytic forms l. 1R. E. SE. P values for paired data analysis. ten pairs. Hydration When three patients who were pain-free were hydrated, their whole blood hemoglobin levels were lower than when studied without hydration two weeks later (Table ). Morphologically, there was no significant change or even consistent trend in the percentage of discocytes, total or spiculated ISCs, or echinocytic forms in the various density layers during hydration compared with the nonhydnated result (Table ). The percentage of cells recovered (corrected to 100%) tended, surprisingly, to be higher in the dense layers ( through 5) and lower in layer 1 during hydration. This 8 I z HbF and F Reticulocyte Studies For one patient proceeding from in crisis to out of pain crisis, we demonstrated microscopically an increase in discocytes, F cells, and reticulocytes in the unseparated fraction (9%, 10%, and 0%, respectively) and to a greater extent in layer 1 (1 %, 1 88%, 175%, respectively). A rise of 0% in discocytes without a change in the percentage of F cells or reticulocytes was noted in layer 5. The percentage change was calculated as (out - in)/(in) x 100. Table 6 summarizes the results of the hemoglobin F studies on four patients in and out of pain crisis. Only one patient, A, showed a rise in total hemoglobin F and hemoglobin F from reticulocytes out of crisis. The other patients (B,C,D) showed an average fall of I5.% in total HbF and of 0.7% in HbF from reticulocytes in unseparated cells and layer 1. Patients C and D did demonstrate a fall in enrichment ratio in layer I, indicating that although they had less total hemoglobin F, a greater proportion of it came from neticulocytes. DISCUSSION The results demonstrate two alterations in erythrocyte morphology in association with pain crisis in sickle cell anemia. The first is an echinocytic change. While metabolic depletion as indicated by decreased ATP levels is a cause of echinocytosis, no ATP deficit has been demonstrated in ISC-nich dense fractions. Using the same density gradient technique described

7 ERVTHROCVTES IN PAIN CRISIS 51 Table. Percentage of Cell Type in Layers 1-5 for Three Pain-Free Sickle Cell Anemia Patients (X. V. Z): Hydrated (H) v Nonhydrated State (NH) Patient x V Z Cell Type Layer H NH H NH H NH Discocytes HEPES 1 5 P. NS Group i HEPES P. NS Group* HEPES P. NS P value: paired d ata analysis: 1 pairs (layers through 5, three patients). ftotal ISC: l, 1, lie. Echinocytic forms: 16, ISP. E1. E11.E61, SE. and blind Nomarsky counts, we have previously shown, however, that reduced glutathione (GSH) levels comelated inversely (r = -.6; P <.0) with the percentage ofechinocytic cells (IE SE E1111, SE) present in the IS layers (S x ) of three patients with sickle cell anemia. 8 5 Evidence for the role of oxidant stress in sickle cell anemia includes (1 ) elevated levels of superoxide dismutase and decreased levels of glutathione peroxidase and catalase in sickle erythrocytes; excess HO and oxidative damage might ensue as a result of this combination;6 () sickle erythrocytes, including the dense fraction separately, generate about twice as much superoxide, peroxide, and hydroxyl radical as normal erythrocytes.7 Thus, the dense fractions that display the most echinocytosis produce increased amounts of oxidants and have decreased (compared to black, normal controls) GSH protection. Normal erythrocytes can be made echinocytic by exposure to H0.8 Hebbel et al have presented data to show that oxidative stress on normal erythrocytes may inhibit membrane calcium ATPase.9 This would be expected to contribute to the elevated calcium levels found in sickle erythrocyte membranes and ISCs.0 Calcium is a known echinocytic agent. Thus, it may be that sickle erythrocytes, both ISC and non-isc, are prone to oxidant stress that leads to increased echinocytic change in pain crisis. The echinocytic change was statistically significant for each gradient layer (group, Table ). This change is important because echinocytes, especially spheroechinocytes, have been shown to have an increased apparent viscosity that correlated with the degree of echinocytosis and did not relate to ATP depletion. The dense, ISC-rich fractions of sickle erythnocytes have been shown to be largely responsible for the increased viscosity of oxygenated, sickle cell anemia blood; 0 to be the fractions most adherent to human endothelium in vitro; and, via an altered lipid distnibution on the outer membrane surface, to accelerate the Russell viper venom time,5 implicating them in the increased clotting tendency demonstrated in pain cmisis.67 It is within our high density fractions, layers Table 5. Percentage of Recovery Corrected to 1 00% for Nine Patients and of Ten Episodes of Pain Crisis Patent l0 Layer I Same person. Patient Days Table 6. Summary of Hemoglobin F Studies in Four Patients and of Pain Crisis Unseparated CelIst Layer 1 A B C D / /.90 = / /.95 =.5.90/ / / /.0 =.8.0/1.55 = /7.60 = / / / /.90 =.9.00/ / Between in-crisis and out-of-crisis study. ttotal % hemoglobin F by radial immune diffusion/% hemoglobin F from reticulocytes. This ratio is called the enrichment ratio.

8 51 WARTH AND RUCKNAGEL through 5, that the echinocytic change, which might also expose more membrane to serve as procoagulant, is most significant (groups and, Table ). The absence of consistent changes in the percentage ofcells reaching the densest layers in crisis implies that erythnocyte water loss may not be a factor in the continuation of pain crisis. Yet the echinocytic change that occurred within the densest layers during pain crisis may be of importance because of the nature of these dense cells. The polymerization of hemoglobin in sickle erythrocytes, as occurs with deoxygenation, has been implicated in producing dense cells whose increased specific gravity appears to be not meyersible.8 These cells display abnormal rheology because of their increased mean corpuscular hemoglobin concentration.9 Their abnormal rheology has been recently demonstrated using the vasculature of an artificially perfused rat mesoappendix. It was shown that these dense cells when oxygenated demonstrated increased peripheral resistance that became even more pronounced during deoxygenation. This latter effect was proportional to cell density. #{176} The amount of polymer present in sickle cells has been previously noted to depend directly upon the hemoglobin concentration and hence the density. Thus, a group of potentially polymer-enriched cells, whose rheology is distinctly abnormal, have been shown in our studies to develop echinocytosis in pain crisis. This would be expected to only decrease their ability to traverse the microcinculation, as for example in the model provided by Noguchi and Schechten. creased stasis, deoxygenation, polymer formation, and tissue damage would result. Of interest are the results with respect to hydration. The three patients who were hydrated while pain-free displayed increased recovery percentages (connected to 100%) in the dense layers, through 5, of the gradient separation performed at the end of 1 hours of hydration versus 1 days later (Table ). No such change was seen in our crisis patients, all of whom were hydrated for at least 6 hours before being studied (Table 5). The lower hemoglobin value in crisis may relate in part to hydration, since the same effect was produced by hydration of patients who were pain-free. However, hydration also appears to mobilize a group of dense erythnocytes in the pain-free state that are not mobilized in pain crisis, perhaps because they are adherent to endothelium or are not as easily released from the bone marrow during crisis. Changes occurring before hydration would have been missed in this crisis study. The increased Hb level postcnisis without a consistent change in neticulocytes (Table I ) indicates that destruction of erythrocytes at the time of crisis may also occur. The second change in erythnocyte morphology observed was a consistent rise in normal-appearing discocytes (Table ). The rise in layer 1 was of special interest, since 80% to 90% of the cells in layer 1 were discocytes. F erythrocytes and meticulocytes were also concentrated theme. The rise in discocytes in this layer might have been due to an increase in HbF-nich, normal-appearing cells, ISCs having been shown to be F poor cells. While one patient did show a rise in discocytes, F erythnocytes, and neticulocytes, when four other patients were studied using more sophisticated techniques, only patient A (Table 6) demonstrated a rise in HbF and HbF from reticulocytes in unseparated cells or layer I. The enrichment ratio fell since the reticulocyte component nose to a greater extent than the total HbF level. It may be that a rise in new HbF synthesis and total HbF postcnisis plays a part in only a subgroup of sickle cell anemia patients responding to pain crisis and that other factors are important in producing the discocytes seen postcrisis. Dover et al have demonstrated that HbF levels are attained in unique ways by each patient in the steadystate and that multiple factors appear to modify F erythrocyte survival and production. Our more limited results would support that idea. The F neticulocytosis may have been missed in patients B, C, D because of variability in when it occurs. The rise in discocytes in layers and S (Table ), where F cells and reticulocytes are few, indicates that the rise in these cells relates to the fall in echinocytic forms, because of their destruction on the termination of the echinocytic event. When we counted 1,000 unseparated erythnocytes (HEPES) we were unable to demonstrate statistically significant changes although the same direction of change was seen. Only by counting ISC- and echinocyte-enniched fractions were significant results achieved. Rieber et al9 reported significant changes counting 500 cells, but five of their eight patients were studied within five days of the initial study and the cells were not classified using Bessis s criteria. Whether the increase in echinocytic forms in pain crisis is primary or secondary cannot be determined from these studies. Yet its occurrence as well as the rise in normal-appearing discocytes after crisis may provide clues to the origin and propagation of pain crisis. ACKNOWLEDGMENT We wish to acknowledge the help of the following people: George J. Brewer, MD, and Ananda S. Prasad, MD, PhD, for advice regarding the manuscript, Jiri Palek, MD and Aileen Liu, for their direction in setting up the gradient technique; Karen Near and Peter Morris for technical assistance; Allen H. Reed, PhD, for statistical advice; and Annette Phillips for secretarial assistance.

9 ERYTHROCVTES IN PAIN CRISIS 515 REFERENCES 1. McCurdy PR, Sherman AS: Irreversibly sickled cells and red cell survival in sickle cell anemia. Am J Med 6:5, 196. Serjeant GR, Serjeant BE, Milner PF: The irreversibly sickled cell; a determinant of haemolysis in sickle cell anemia. Br J Haematol 17:57, Serjeant GR: Irreversibly sickled cells and splenomegaly in sickle cell anemia. Br J Haematol 19:65, Serjeant GR, Serjeant BE, Condon P1: The conjunctival sign in sickle cell anemia: A relationship with irreversibly sickled cells. JAMA 19:18, Nagpal KC, Asdourian GK, Goldbaum MH, Raichand M, Goldberg MF: The conjunctival sickling sign, hemoglobin 5, and irreversibly sickled erythrocytes. Arch Ophthalmol 95:808, Warner R, Sarnaik S, Brewer GJ: Irreversibly sickled cells and hyposthenuria in sickle cell anemia. Blood 5(suppl 1 ):5A, 1978 (abstr) 7. Diggs LW: The crisis in sickle cell anemia. Am J Clin Pathol 6:1109, Barreras L, Diggs LW: Bicarbonates, ph and percentage of sickled cells in venous blood of patients in sickle cell crisis. Am J Med Sci 7:710, Rieber EE, Veliz G, Pollack 5: Red cells in sickle cell crisis: Observations on the pathophysiology of crisis. Blood 9:967, Chien 5, Usami 5, Bertles if: Abnormal rheology of oxygenated blood in sickle cell anemia. J Clin vest 9:6, I. Corash LM, Piomelli 5, Chen HC, Seaman C, Gross E: Separation of erythrocytes according to age on a simplified density gradient. J Lab Clin Med 8: 17, Clark MR. Unger RC, Shohet SB: Monovalent cation composition and ATP and lipid content of irreversibly sickled cells. Blood5l:1l69, Natta CL, Machlin Li, Brin M: A decrease in irreversibly sickled erythrocytes in sickle cell anemia patients given vitamin E. Am J Clin Nutr :968, Brewer Gi, Brewer LF, Prasad AS: Suppression of irreversibly sickled erythrocytes by zinc therapy in sickle cell anemia. J Lab Clin Med 90:59, Chudwin DS, Rucknagel DL: Immunological quantification of hemoglobins F and A. Clin Chem Acta 50:1, Adams JG III, Winter WP, Rucknagel DL, Spencer HH: Biosynthesis of hemoglobin Ann Arbor: Evidence for catabolic and feedback regulation. Science 176:17, Hanash SM, Shapiro DN: Separation ofhuman hemoglobins by ion exchange high performance liquid chromatography. Hemoglobin 5:165, Wetterstroem N, Brewer Gi, Warth ia, Mitchinson A, Near K: The relationship of glutathione levels and Heinz body 10:589, 198 irreversibly sickled cells in sickle cell anemia. i Lab Clin Med 10:589, Bessis M, Weed RI: Preparation ofred blood cells (r.b.c.) for SEM-A survey of various artifacts, in Johari 0, Corvin I (eds): Proceedings of the Fifth Annual Scanning Electron Microscope Symposium and Workshop on Biological Specimen Preparation for Scanning Electron Microscopy. Chicago, lit Research stitute, 197, p Bessis M: The erythrocyte series, in Living BlOOd Cells and Their Ultrastructure. (Transl Weed RI). New York, Springer- Verlag, 197, p Atwater J, Erslev AJ: Reticulocyte staining, in Williams Wi, Beutler E, Erslev Ai, Rundles RW (eds): Hematology (ed ). New York, McGraw-Hill Book Co., 1977, p Atwater J, Erslev Ai: Fetal hemoglobin-differential staining, in Williams Wi, Beutler E, Erslev AJ, Rundles RW (eds): Hematology (ed ). New York, McGraw-Hill Book Co, 1977, p 160. Nakao M, Nakao T, Yamazoe S: Adenosine triphosphate and maintenance of shape of the human red cells. Nature I 87:95, Weed RI, LaCelle PL, Merrill EW: Metabolic dependence of red cell deformability. i Clin vest 8:795, Wetterstroem N, Brewer Gi: Sickle cell oxidant damage and its treatment. Clin Res 9:7lA, 1981 (abstr) 6. Das 5K, Nair RC: Superoxide dismutase, glutathione peroxidase, catalase and lipid peroxidation of normal and sickled erythrocytes. Br i Haematol :87, Hebbel RP, Eaton iw, Balasingam M, Steinberg MH: Spontaneous oxygen radical generation by sickle erythrocytes. i Clin vest 70:15, Jam 5K, Shohet SB: Apparent role of cholesterol as an erythrocyte membrane anti-oxidant. Clin Res 9:6A, 1981 (abstr) 9. Hebbel RP, Eaton JW, Steinberg MH: Autoxidation and the membrane abnormalities of sickle RBC. Blood 60:5A, 198 (abstr) 0. Eaton JW, Skelton TD, Swofford HS, Kolpin CE, Jacob US: Elevated erythrocyte calcium in sickle cell disease. Nature 6: 105, Palek J, Thomae M, Ozog D: Red cell calcium content and transmembrane calcium movements in sickle cell anemia. J Lab Clin Med 89:165, Meiselman Hi, Baker RF: Flow behavior of ATP-depleted human erythrocytes. Biorheology 1:1 Il, Meiselman Hi: Rheology of shape-transformed human red cells. Biorheology 15:5, Hebbel RP, Yamada 0, Moldow CF, Jacob US, White ig, Eaton JW: Abnormal adherence of sickle erythrocytes to cultured endothelium. i Clin vest 65:15, Chiu D, Lubin B, Roelofsen B, Van Deenen LLM: Sickled erythrocytes accelerate clotting in vitro: An effect of abnormal membrane lipid asymmetry. Blood 58:98, Leichtman DA, Brewer GJ: Elevated plasma levels of fibrinopeptide A during sickle cell anemia pain crisis-evidence for intravascular coagulation. Am i Hematol 5: I 8, Alkjaersig N, Fletcher A, Joist U, Chaplin H ir: Hemostatic alterations accompanying sickle cell pain crisis. J Lab Clin Med 88:0, Fabry ME, Nagel RL: The effect ofdeoxygenation on red cell density: Significance for the pathophysiology of sickle cell anemia. Bbood6O:l70, Clark MR. Mohandas N, Shohet SB: Deformability of oxygenated irreversibly sickled cells. i Clin vest 65:189, Kaul DK, Fabry ME, Windisch P. Baez 5, Nagel RL: Erythrocytes in sickle cell anemia are heterogenous in their rheological and hemodynamic characteristics. J Clin vest 7:, Noguchi CT, Schechter AN: The intracellular polymerization of sickle hemoglobin and its relevance to sickle cell disease. Blood 58:1057, Bertles if, Milner PFA: Irreversibly sickled erythrocytes: A consequence of the heterogenous distribution of hemoglobin types in sickle cell anemia. i Clin vest 7: 17 1, I 968. Dover Gi, Boyer SH, Charache 5, Heintzelman K: dividual variation in the production and survival of F cells in sickle cell disease. N Engl i Med 99:18, 1978

10 198 6: Density ultracentrifugation of sickle cells during and after pain crisis: increased dense echinocytes in crisis JA Warth and DL Rucknagel Updated information and services can be found at: Articles on similar topics can be found in the following Blood collections formation about reproducing this article in parts or in its entirety may be found online at: formation about ordering reprints may be found online at: formation about subscriptions and ASH membership may be found online at: Blood (print ISSN , online ISSN ), is published weekly by the American Society of Hematology, 01 L St, NW, Suite 900, Washington DC 006. Copyright 011 by The American Society of Hematology; all rights reserved.