American Journal of Hematology 5: 323-333 (1978)

The Lipids of the Erythrocyte in Paroxysmal Nocturnal Hemoglobinuria Jon P. Gockerman, MD Department of Medicine, University of Kentucky Medical Center, Lexington

Paroxysmal nocturnal hemoglobinuria (PNH) is a disorder characterized by the production of abnormal erythrocytes (PNH-E). The PNH-E undergoes early intravascular destruction which appears t o be related t o an intrinsic membrane defect that results in extreme sensitivity t o lysis by complement. The lipids of the PNH-E have been evaluated because of the variable results reported in the past. This study shows a normal content of cholesterol and lipid phosphorus with a normal distribution of the phosopholipids except for a slight increase in lysophosphatidyl choline (LPC) when measured using the complete lipid extract of the PNH-E. Fractionation of the lipid extract over a silicic acid column produced an alteration in the distribution of the phospholipids, with LPC and sphingomyelin increasing, and the other phospholipds decreasing. These alterations were seen only when the phospholipids were fractioned over a silicic acid column and were clearly an in vitro phenomenon. The levels of plasmalogen and malonaldehyde were normal, unlike those seen with oxidant stress. Fatty acid analysis of t h e phospholipids showed small but consistent changes: increased 16:O and decreased 18:2. The total glycosphingolipids were slightly decreased in the patients with severe PNH. The results support and extend previous lipid data indicating abnormalities in the phospholipids of the PNHE. The alterations of the phospholipids with fractionation correlate with previous whole cell data that suggest lipid instability in the PNH-E. Key words: paroxysmal nocturnal hemoglobinuria, erythrocytes Abbreviations: PNH) paroxysmal nocturnal hemoglobinuria; PNH-E) paroxysmal nocturnal hemoglobinuria erythrocyte; IIa) total chloroform phase lipid extract; F1) sterol containing fraction after silicic acid column chromatography; F2) glycosphingolipid containing fraction after silicic acid column chromatography; F3) phospholipid containing fraction after silicic column chromatography; LPC) lysophosphatidyl choline; SM) sphingomyelin;PC) phosphatidyl choline; PS/PI) phosphatidyl serine and phosphatidyl inositol; PE) phosphatidyl ethanolamine; PA) phosphatidic acid; TLC) thin-layer chromatography; GLC) gas-liquid chromatography; PCV) packed cell volume; RBC) red blood cells; CLS) complement lysis sensitivity. Received for publication June 12, 1978; accepted September 27, 1978. This work was presented in part at the Annual Meeting of the American Society of Hematology, Dallas, Texas, 1975 (Blood 46,1018,1975). Address reprint requests to Dr. J.P. Gockerman, Division of Hematology and Oncology, University of Kentucky Medical Center, Lexington, K Y 40506.

0361-8609/78/0SO4-0323$02.300 1978 Alan R. Liss, Inc

324

Gockerman

I NT ROD UCTl ON

Paroxysmal nocturnal hemoglobinuria (PNH) is a disease characterized by the increased sensitivity of the PNH erythrocyte (PNH-E) to lysis by the action of complement [ 11 . This presumably results in the clinical situation of intravascular hemolysis with subsequent hemoglobinuria and anemia. The underlying defect resulting in the increased sensitivity of the PNH-E to complement lysis is unknown. Rosse and co-workers have shown that for a given amount of C1 activated, larger quantities of the later complement component C3 are added to the PNH-E than to normal erythrocyte membranes and yet less total membranebound complement components are required for PNH-E lysis [2]. Gotze and associates have shown that activation of the alternate complement pathway results in greater binding of C5 to the PNH-E than to normal erythrocytes [3]. This occurred using any source of human complement, therefore implicating a basic defect in the membrane of the PNH-E. A membrane defect in the PNH-E is further supported by the following observations: PNH-E have decreased acetylcholinesterase in their membranes and increased reactivity of the iI antigen complex; and incubation of normal erythrocytes with an oxidizing system, sulphydryl compound, or nonsulphydryl reducing agents produces a PNH-like erythrocyte [4-61. Biochemical evaluation of the membrane of the PNH-E has shown some protein banding abnormalities on polyacrylamide gel electrophoresis and conflicting results on membrane lipid analysis [7-91. Owing to the evidence of membrane abnormality in the PNH-E and because of the inconsistent lipid abnormalities reported, a reevaluation of the lipids in the PNH-E membrane was undertaken. MATERIALS AND METHODS

The diagnosis of PNH was made by the presence of a positive complement lysis sensitivity test as described by Rosse et a1 using a rabbit anti-human erythrocyte antibody [ l , 101. Laboratory personnel on no medications were the source of normal erythrocytes. Blood was drawn into acid-citrate-dextrose (NIH Formula A) anticoagulant and stored at 4°C. All blood donors gave informed written consent. All lipid extractions were done at 4°C under an atmosphere of nitrogen within 24 hours of bleeding. The methods used were those previously described [ l 11 . Red blood cell ghosts were prepared by the method of Dodge et a1 [12] and the lipids were extracted by the Bligh-Dyer procedure [ 131 , as modified by Kinsky et a1 [ 141 . The final lipid-containing chloroform phase, referred to as IIa, was fractionated on a silicic acid column according to the method of Vance and Sweely [ 151 . Three fractions were obtained: F1, containing the sterols; F2, containing the glycosphingolipids; and F3, containing the phospholipids including plasmalogens. The recovery of sterols and phospholipids from the column was greater than 90%.All fractions were collected at 4°C and stored under nitrogen at -20°C. Lipid phosphorus was determined by the method of Gerlach and Denticke [ 161 , cholesterol by the method of Zlatkis et a1 [ 171 , and plasmalogen and sphingomyelin by the method of Dittmer and Wells [18] . Individual phospholipids (LPC, SM, PC, PS/PI, PE, and PA) were separated and measured quantitatively by thin-layer chromatography as described previously using chloroform-methanol-acetic acid-water, 50:30:4:2 (v/v) [acid plate] or chloroform-methanolammonium hydroxide, 56:24:4 (v/v) [basic plate] [ l l ] . Fatty acid analysis of the phospholipids was performed on the F3 fraction and its individual phospholipids were separated by acid TLC. Fatty acid methyl esters were prepared by the boron fluoride-methanol technique of Morrison and Smith [ 191 . The material was analyzed as before with a Packard gas-liquid chromatograph apparatus, model 800 (Packard Instrument

Lipids of the PNH Er y t h r o cy t e

325

Co., Downers Grove, Illinois). The total glycosphingolipids were quantitated by GLC and a qualitative assessment of the individual glycosphingolipids was done by TLC as previously described [l 1] .Evidence of oxidative alterations in lipids was evaluated by the thiobarbituric acid assay for malonaldehyde and ultraviolet absorbency at 234 and 268 nm [ 2 0 , 2 1 ] . All lipid data are expressed as ymoles per 100 ml of packed cell volume and pmoles per 1010 red blood cells and presented as mean k 1 SD. The term “separate preparations” refers t o a complete lipid preparation starting from the erythrocyte. Each separate preparation was analyzed at least two times. RESULTS

All patients had quantitative complement lysis sensitivity assays performed except for patient E.P. These correlated with the severity of the anemia and the degree of reticulocytosis. The results of the lipid studies of the PNH-E are tabulated in decreasing order of severity on the basis of the CLS titer in the data tables. E.L. had a CLS titer of 200, representing 100%of the PNH-E; A.G. had a CLS titer of 45, representing 77% of the PNH-E; E.B. had a CLS titer of 50, representing 35% of the PNH-E; D.R. had a CLS titer of 39, representing 17% of the PNH-E and a second population with a CLS titer of 22; and D.T. had a CLS titer of 16, representing 65% of the PNH-E. The second population of the PNH-E had a normal CLS titer in all cases except as noted above. Phospholipids and Cholesterol

Table I shows the results of measurements of the total lipid phosphorus and cholesterol in five PNH patients analyzed 15 times. No significant difference between the means of the normal subjects and the PNH patients exists. One patient (E.L.) who had very severe PNH showed decreased lipid phosphorus when expressed as PCV but not when expressed on a 1010 RBC basis. Phospholipids

Table I1 shows the results of the phospholipid measurements of the total lipid extract (Ha) of the PNH-E using both acid and base solvent systems for TLC. No major differences

TABLE 1. Phospholipid and Cholesterol Analysis of PNH Erythrocyte Membranes*

Erythrocyte membranes Na Normal subjects 10 PNH patients E.L. 2 E.B 3 D.R. 5 D.T. 4 E.P. 1 Summary PNH preps 15

pmoles phosphate per: 100 ml PCV 1010 RBC

pmoles cholesterol per: 100 ml PCV 1010 RBC

372 f 23

3.0 t 0.4

366 r 37

2.9 r 0.3

253 f 83 346 t- 73 371 t- 35 352 t 43 375

2.4 t 0.7 4.2 t 0.8 3.5 f 0.8 2.5 f 0.2 2.7

320 * 60 338 t 49 357 f 38 390 r 47 328

3.1 f: 0.2 4.3 f 0.3 3.4 f 0.8 2.8 f 0.5 2.4

347 ?- 64

3.1 t 0.9

355 t 46

3.2 f 0.8

* Mean f 1SD. a Number of separate preparations analyzed.

26.8 28.5 21.5

2 2

PE

f

f

f

+_

f

1.9

0.7 2.6

2.3

27.9 t 4.0 28.2 * 2.9 28.6 f 2.5 32.3 2.2

31.7

4

3.4 t 1.9 1.1 t 0.7 2.5 t 1.8 0.7 ? 0.2

0.4

2 3 4 4

f

0.9

4

PE f

2.6

12.9 f 1.3 12.7 f 1.7

13.4 ?- 2.2

30.3 f 0.5 29.3 f 0.7

32.1 * 1.3

15.0 k 3.8 26.0 f 3.3 13.3 f 2.2 31.8 f 4.5 13.6 f 1.4 27.3 f 2.0 11.9 t 1.7 26.8 f 1.7 IIa basic plate PS/PI PC

28.5

IIa acid plate PC

12.6 k 2.7

PS/PI

0.6 t 0.6 0.5 f 0.5

27.7 i: 0.6 30.1 i. 3.2

0.4

0.5

27.2 t 1.8

t

PA/LPC

2.5 0.8 1.3 0.3

1.3 f 1.3

SM

1.6

4.8 * 1.7 f 2.3 + 0.9 t

f

LPC

25.2 t 2.1 24.5 t 5.7 27.8 t 2.3 28.5 t 2.3

26.3

SM

*Values expressed as percentage o f total lipid phosphorus; each represents mean ?- ISD. a Number of separate preparations analyzed. PA) phosphatidic acid, PE) phosphatidyl ethanolamine, PS/PI) phosphatidyl serine/phosphatidyl inositol, PC) phosphatidyl choline, LPC) lysophosphatidyl choline, SM) sphingomyelin.

Normal subjects PNH patients D.R. D.T.

Normal subjects PNH patients E.L. E.B. D.R. D.T.

Erythrocyte membranes.. Na PA

TABLE 11. Percent of Individual Phosphatides in Total Lipid Extract (Ha) of PNH Erythrocyte Membranes"

N Q\

w

Lipids of the PNH Erythrocyte

327

appear in the individual phospholipids, though thk levels of LPC were increased in the erythrocyte preparations from some of the patients with severe to moderate PNH (E.L., D.R.). When the total lipid extract (Ha) of PNH-E was fractionated over a silicic acid column and the phospholipid-containing fraction (F3) was analyzed, a significant abnormality was found (Table 111). There was an increase in the phospholipid in the LPC area on TLC and a decrease in PE, PS/PI, and PC. These alterations were never seen in lipid preparations from normal erythrocytes handled in exactly the same manner. These alterations were not seen in every PNH F3 preparation from the same patient, but they were more frequent (SO-90% of the F3 preparations) in the clinically most severely affected PNH patients. Those samples without the alterations had a normal phospholipid distribution. The PNH F3 preparations not initially showing this abnormality tended to develop it with storage under nitrogen at -20°C. These abnormalities were never seen in normal F3 preparations even with storage under identical conditions as the F3 PNH preparations for as long as six months. To make an independent check for oxidation of vinyl-containing lipids the levels of SM and plasmalogen were determined by alternate methods on both the IIa and F3 preparations. The results presented in Table IV show normal values for these lipids. Levels of malonaldehyde as determined by the thiobarbituric acid assay and ultraviolet absorbency at 234 and 268 nm were performed on the IIa and F3 lipid preparations. These results were the same in the lipid preparations from normals as compared to the PNH preparations. The F3 PNH preparations showing the abnormal phospholipid distribution did not differ from normal F3 preparations, suggesting that peroxidation did not account for changes in phospholipid patterns owing to fractionation on silicic acid columns. Both normal and PNH preparations developed elevated values after storage for several months; but, as noted above, the F3 preparations from normal erythrocytes did not show abnormal phospholipid distribution. The abnormal F3 preparations did not show streaking on acid TLC after spraying with ninhydrin, which suggested a lack of peroxidation [21]. Fatty Acids of Phospholipids

Table V gives the results of the analysis of the major fatty acids of the phospholipids. No major differences (using the t test of the means of two independent samples) were found in the fatty acid composition of the PNH-E regardless of whether an abnormal phospholipid distribution was seen on TLC. A slight decrease in 18:2 fatty acids (P = 0.05) was detected when the fatty acids were analyzed as a whole from the phospholipids. Analysis of the fatty acids of individual phospholipids showed little variation from normals except with PS/PI, where the fatty acid 16:0 was increased (P < 0.01) and 22:0/21:0/20:3 were decreased (P