Immunobiol., vol. 164, pp. 118-126 (1983)

Department of Pathology, University of Maryland School of Medicine, Baltimore, Maryland, U.S.A.

Lysis of Paroxysmal Nocturnal Hemoglobinuria Erythrocytes by Acid-Activated Serum

GERTRUD HANS CHI , C. HAMMER, R. JI]I, URSULA ROTHER, and MOON SHIN2

Received September 10,1982' Accepted November 19,1982

Abstract

Erythrocytes from paroxysmal nocturnal hemoglobinuria patients (PNH-E) are much more susceptible to lysis by acid-activated human serum than normal human erythrocytes. Acidifi­cation of normal human serum to pH 6.4 in the absence of erythrocytes generates this lytic activity independently of the alternative pathway of complement activation. A shift of pH of a mixture of purified human C5 and C6 to 6.4 at O°C generates a similar activity C(56)a that lyses PNH-E together with C7-C9 much more efficiently than normal erythrocytes. Since acid­activation of normal human serum occurs in the absence of C3, the acid-activated C56 appears to be the lytic principle in acidified human serum.

Introduction

Paroxysmal nocturnal hemoglobinuria (PNH) is a disease characterized by an abnormal sensitivity of erythrocytes (PNH-E) to complement­mediated lysis (1, 2). This enhanced susceptibility is observed when PNH-E are lysed by antibody and complement (classical pathway) (1), by serum activated with inulin or cobra venom factor (alternative pathway) (3, 4) or with the «reactive lysis» system, employing preformed C5b6 and C6-C9 (5).

In 1937, HAM described that PNH-E, but not normal human E (NHE) are lysed by acidified human serum (6, 7). The reason for the different susceptibility of PNH-E and NHE is poorly understood, but it has been shown that different mechanisms might be involved and that there exist different types of PNH-E with regard to complement sensitivity (5, 8, 9). Recently it has been shown that lowering the pH of normal human serum (NHS) to 6.4 generates an activity that lyses un sensitized chicken erythro­cytes (10). Further studies on the nature of acid-activation revealed that acid treatment of purified human CS and C6 results in a lytic complex desig­nated C(S6)" that lyses unsensitized erythrocytes together with C7, C8 and C9 (10, 11). Analysis of the chain structure of acid-activated CS demon-

1 Recipient of a grant of the Deutsche Forschungsgemeinschaft Ha 1129/1 2 The work was supported by USPHS grant 1 R01 NS15662-01 PTHA

Paroxysmal Nocturnal Hemoglobinuria . 119

strates a C6 dependent C5 a-chain cleavage (11). C(56)a generation is almost as efficient as C5b6 generated by C5-convertases. Both the activities react similarly with the target cells (11, 12).

In the present study, we examined the interaction of the acid-activated terminal complement components C(56)a-C9 with PNH-E and NHE.

Materials and Methods

Buffers

VBS: Veronal-buffered saline, pH 7.4, IJ, = 0.15, was prepared by diluting a 5X stock solution (13) 5-fold with water. EDTA-VBS was prepared by mixing 9 volumes ofVBS with 1 volume of 0.1 M EDT A.

Target cells

Erythrocytes from three PNH patients and from healthy donors (0 Rh negative) were collected, washed, and stored at 4°C in Alsever's solution. Control NHE were obtained at the same day as the PNH-E, because the degree of lysis of human erythrocytes by complement varies considerably with in vitro aging of cells. For use, the cells were adjusted to 1 X 109 Iml in EDTA-VBS. According to clinical data, the patients had about 7 to 10 % PNH-E cells. Lysis was also obtained with C5b6, C7-9 (reactive lysis); therefore the cells of all three patients could be classified as type III according to (9).

Complement

For NHS, blood was drawn from healthy donors (Ab positive), and the serum was obtained by clotting 30 minutes at room temperature and 60 minutes at 4°e. The serum was stored frozen at -70°e. Rabbit serum (NRS) or C6-deficient rabbit serum (C6d-RS) (14) was obtained from normal or C6-deficient rabbits and treated in the same way as human serum. Human C5 and C6 were prepared according to (15). Guinea pig C7 (C7gP ) and C8 (C8gP) were purified according to (16) and guinea pig C9 (C9gp) by methods described in (17). For the hemolytic assay, 100 CH50 units of each C8gp and C9gp titrated on sheep EA carrying complement components Cl through C7 were used. NHS or C6d-RS diluted 1 :100 with EDTA-VBS was used as the source of C8 and C9. NHS depleted of C3 (C3d-HS) was prepared by immunoadsorption with anti-C3 coated glass beads according to JUNGFER (18). Depletion of C3 was tested by C3 titration on sheep EAC142 and excess of isolated C5-C9.

Antisera

Rabbit anti-human factor B and anti-human C3 were purchased from Behringwerke, Marburg, West Germany.

Immunoelectrophoresis

Immunoelectrophoresis was carried out in a Tris-Veronal buffer pH 8.6 in 1 % agarose (Litex, Glostrup, Denmark) containing 10 mM EDTA. To detect factor B, 20 IJ,I serum was applied, and anti-B was used in a 1:2 dilution. C3 cleavage was tested in crossed immunoelec­trophoresis according to LAURELL (19) with 1 % anti-C3 in the second dimension. NHS incubated with inulin (5 mg/ml) for 20 minutes at 30°C was used as positive control for testing conversion of factor Band C3.

Acid-activation

The pH of NHS, diluted 1/4 in EDTA-VBS, was adjusted to 6.4 with 0.1 N HCI at O°C, then neutralized immediately with 0.1 N NaOH. The final dilution of serum was adjusted to

120 . GERTRUD HANSCH, C. HAMMER, R. JIJI, URSULA ROTHER, and MOON SHIN

1/5. This concentration is referred to as relative concentration 1 in figures. At O°C, the lytic activity was stable for several hours. NRS was acidified in the same way. To generate C (56)a activity, human C5 (120 units) and human C6 (60 units) were mixed and brought up to a volume of 1.5 ml with VBS at o°e. The pH of this mixture was adjusted to 6.4 with 0.1 N HCl at O°e. Following neutralization with O.lN NaOH, the final volume was adjusted to 3 ml. This C (56)a concentration is referred to as relative concentration 1 in the figures.

Hemolytic assays

Lysis of E with acidified serum: 0.1 ml of E (1 X 109/ml) and 0.2 ml of acidified serum in various dilutions were incubated at 37"C for 60 min. 2.2 ml of ice-cold saline was added, and the released hemoglobin was measured at 412 nm. For 100 % lysis, E were treated with 0.1 % Triton-X-100. Measurement at 412 nm was performed with samples diluted 10-fold with saline.

Lysis of E with C(56}"-9: 0.1 ml of E were incubated with 0.2 ml of C (56)" in various dilutions and 0.1 ml of excess C7KP for 30 minutes at 27°e. The cells were washed in EDTA­VBS and incubated further with C8 and C9 of different species for 60 minutes at 37°e. Lysis was calculated as «lytic site per cell» (2) according to (20) (2 = 1 equals to about 60 % lysis).

Results

Lysis of PNH-E and NHE by acid-activated serum

The pH of NHS was adjusted to 6.4 at O°C, and immediately titrated back to 7.4 as described in Materials and Methods. Various dilutions of this acid-activated serum were incubated with PNH-E from patient I and from patient II or NHE for 60 min at 37°C. As shown in Figure la and lb, PNH-E were found to be much more sensitive to lysis by acid-activated NHS than NHE. Non-acidified serum did not lyse PNH-E or NHE in significant degree (results not shown). Acidified NRS was 2-fold more efficient than acidified NHS in lysing both NHE and PNH-E as demon­strated with patient II (Fig. lb, lc). The differential susceptibility, however, remained unchanged (Fig. lc).

~ .05

... .025

patient / a .2 patient II b .2

. , /., ~~ ~~/ -

0.25 0.5 0.25 0.5

,.e/alive cone.of acidified NHS

~-" /" 0

/ ~~

0.25 0.5

,.e/alive cone. of acidified NRS

Fig. 1. Lysis of PNH-E or NHE by acid-activated NHS or NRS: Various dilutions of acid­activated serum were incubated with PNH-E (0) from two patients. Lysis was compared with NHE from two donors (e) (a, b). A differential lytic susceptibility was also observed when acid-activated rabbit serum was used (c).

Paroxysmal Nocturnal Hemoglobinuria' 121

Lysis of PNH-E and NHE with C(56J and C7-C9

Next we tested if the enhanced lytic susceptibility of PNH-E is also found with C(S6Y-C9. A mixture of purified human CS and C6 was adjusted to pH 6.4 as described above. Various amounts of C(S6Y and excess C7gp were incubated with PNH-E or NHE from three patients or three donors, respectively. After incubation for 30 min at 27°C, excess human CS plus C9 was added, and the cells were incubated further for 60 min at 37°C. In all three cases, PNH-E lysed to a higher extent than NHE (Fig. 2a-c). From this result, it is apparent that the pH shift of CS and C6 results in an activity that lyses PNH -E as well as NHE together with C7-C9. The lysis of normal E in each experimental setting, however, was much less efficient.

Lysis of PNH-EC(56J7 intermediates with C8 and C9 from different speCles

PNH-E from patient II were incubated with C(S6)a in various concen­trations and excess C7gp to form EC(S6)a7 intermediates. The cells were washed and lysed with NHS in EDTA-VBS as a source of excess human CS, C9 or with excess CSgp, C9gp

, or with C6d-RS as a source of excess

Fig. 2. Lysis of PNH-E or NHE with acid-activated C (56)" and C7-C9: Vari­ous amounts of C (56)" were incubated with PNH-E from three patients (0) or three donors (.), respectively, and excess C7-C9 was added.

patient I a

. ,

o~~~==~========~~ patient /I b

0 patient //I c

.3

o~ .2 /

0 . , 0

Q25 as relative cone.of C(5S)a

122 . GERTRUD HANSCH, C. HAMMER, R. JUI, URSULA ROTHER, and MOON SHIN

.3 a b

PNH-£ NH£

.2

as

Fig. 3. Lysis of PNH-EC (56)" 7 intermediates (a) or NHEC (56)" 7 intermediates (b) with CS and C9 from various species: PNH-EC (56)' 7 or NHEC (56)' 7 intermediates made with various amounts of C (56)' were lysed with C6d-RS (.&) or with NHS in EDTA-VBS (e) as the source of CS and C9 or with purified CSgp, C9gp (\7).

rabbit CS, C9. As shown in Figure 3a, the degree of lysis was much higher with CS and C9 from rabbit than from human source. With csgp and C9gp, no lysis was observed. These results show that the different lytic suscepti­bility of PNH-E to lysis by acid-activated NHS or NRS (see Fig. lb, c) is, at least in part, attributable to a species incompatibility between human E and the source of C8 and C9. Rabbit C8 and C9 lysed NHEC(56)? intermediates also more efficiently than human C8, C9 (Fig. 3b).

Independence of acid activation of the classical or the alternative pathway of complement activation

A shift of pH to 6.4 activates purified C5 and C6 in the absence of a classical or alternative pathway convertase. Since it was suggested earlier that acidification of serum triggers the alternative pathway activation (3), we examined if lowering the pH of serum to 6.4 indeed affects the proteins of the alternative pathway. The pH of NHS was adjusted to 6.4 at a °C and neutralized. Factor B conversion was tested in immunoelectrophoresis against anti-B. Serum incubated with inulin was used as positive control. Factor B was not converted (i.e., not cleaved) by acid treatment (Fig. 4b), whereas factor B conversion was seen in the inulin-activated serum (Fig. 4a). A similar experiment showed that C3 was not cleaved by acid treatment (Fig.4c). C3 cleavage in inulin-activated serum is shown in Figure 4d. Furthermore, in serum depleted in C3 by immunoadsorption, the lytic activity could be generated by acidification to the same extent as in normal serum (Fig. 5).

Fig. 4. Electrophoresis of acidified NHS and NHS activated by the alternative pathway against anti-B or anti-C3: While NHS acti­vated via the alternative pathway shows con­version of factor B (a) and C3 (d), acidifica­tion did not cleave factor B (b) or C3 (c). (Kathode is to the right, or the top, resp.).

Fig. 5. Generation of the lytic activity in C3-depleted serum: NBS depleted in C3 (_) was acidified to pH 6.4 and tested for lytic activity on PNH-E (patient

.8

III). Lysis was found to the same 0 extent as in a serum containing C3 (\7).

Paroxysmal Nocturnal Hemoglobinuria . 123

,.e/ative conc.of the ,.espective se"a

124 . GERTRUD HANSCH, C. HAMMER, R. JIJI, URSULA ROTHER, and MOON SHIN

Discussion

We studied the interaction of PNH-E or NHE with acid-activated NHS or C(S6Y and C7-C9. Erythrocytes from three patients and control donors were used as target cells. In all three cases, PNH -E were much more susceptible to lysis by acid-activated NHS or C(S6)'-9 than NHE (Fig. 1 and Fig. 2 resp.). The lysis of PNH-E as well as NHE by acid-activated rabbit serum was much higher than by acid-activated human serum (Figs. Ib, c). This could be attributed to the species of C8 and/or C9, since the lysis of PNH-E or NHE C (56)" 7 intermediates carrying a limited number of C7 sites was also much higher with rabbit C8, C9 than with human C8, C9 (Fig. 3). The failure of guinea pig C8 and C9 to lyse PNH-E has shown to be localized to guinea pig C9 (21). The sensitivity of PNH-E to lysis by acidified serum (Fig. 1) or C (56)"-9 (Fig. 2) was different in the three patients studied. The differences in lytic sensitivity of PNH-E to CSb-9 have been described earlier and are used to classify PNH-E (1, 5, 8, 9).

The lysis of PNH-E by acid-activated NHS is often used as a convenient laboratory test for the diagnosis of PNH. The test system employed in the present studies differs from the Ham test in so far as the serum is briefly acidified in the absence of the target cells at a °C and is neutralized before the reaction with the erythrocytes. Under those conditions we found that the generation of the lytic activity in whole serum by pH shift did not require C3, since a lytic activity can be obtained by acidification of C3-depleted serum to the same extent as in normal human serum. Furthermore, no factor B or C3 cleavage was observed following acid treatment of serum. In this particular experiment, the PNH-E (patient III) were much more susceptible to complement mediated lysis compared to the other experi­ments, because they were aged for two days at 4°C.

These findings are not in agreement with a previous report (3) in which the authors described that the lowering of serum pH to 6.5 activates the alternative pathway of complement activation. In their experiments, con­version of factor B or C3 was obtained after prolonged incubation of serum at pH 6.5 for 60 min at 37°C, while control serum was kept at pH 8.0. In our experiments, C3 and B conversion was tested immediately after acidifi­cation at 0 °C, since under those conditions a lytic activity could be demonstrated.

Even though acid activation of NHS does not require alternative pathway activation, it has been shown by others that large amounts of C3 are bound to PNH-E when the serum is acidified together with the cells (22). Since cell-bound C3 enhances the lysis by C (56)" and C7-C9 (12) in acid­activated serum, it is possible that different mechanisms lead to cell lysis: C (56)" generated by acid-activation can bind directly to the cells and cause lysis in the presence of C7-C9. This pathway would be enhanced greatly by cell-bound C3b. In addition, cell-bound C3 also would trigger the alterna-

Paroxysmal Nocturnal Hemoglobinuria . 125

tive pathway, generating C5b6 that similar to C(56)a would lyse the cells, together with C7-C9.

Since PNH-E bind much higher amounts of C3 than NHE (22), this could at least partly explain the higher lytic susceptibility of PNH-E.

On the other hand, with isolated C(56)" and C7-C9, PNH-E were still more susceptible to lysis than NHE. This would suggest that additional factors such as membrane properties inherent only to PNH -E but not to NHE are responsible for the enhanced lytic activity.

References

1. ROSSE, W. F., and J. V. DACIE. 1966. Immune lysis of normal human and paroxysmal nocturnal hemoglobinuria (PNH) red blood cells. 1. The sensitivity of PNH red cells to lysis by complement and specific antibody. J. Clin. Invest. 45: 736.

2. ROSSE, W. F., and J. V. DACIE. 1966. Immune lysis of normal human paroxysmal nocturnal hemoglobinuria (PNH) red blood cells. II. The role of complement compo­nents in the increased sensitivity of PNH red blood cells to immune lysis. J. Clin. Invest. 45: 749.

3. GOTZE, 0., and H. J. MOLLER-EBERHARD. 1972. Paroxysmal nocturnal hemoglobinuria. Hemolysis initiated by the C3 activator system. N. Engl. J. Med. 286, 4: 180.

4. KABAKCI, T., W. F. ROSSE, and G. L. LOGUE. 1972. The lysis of paroxysmal nocturnal hemoglobinuria red cells by serum and cobra factor. Br. J. Haematol. 23: 693.

5. PACKMAN, C. H., S. E. ROSENFELD, P. E. JENKINS, Jr., P. A. THIEM, and J. P. LEDDY. 1979. Complement lysis of human erythrocytes. Differing susceptibility of two types of paroxysmal nocturnal hemoglobinuria cells to C5b-9. J. Clin. Invest. 64: 428.

6. HAM, T. H. 1937. Chronic hemolytic anemia with paroxysmal nocturnal hemoglobinuria: Study of the mechanism of hemolysis in relation to acid-base equilibrium. N. Eng!. J. Med. 217: 915.

7. HAM, T. H., and 1. H. DINGLE. 1939. Studies on destruction of red blood cells. II. Chronic hemolytic anemia with paroxysmal nocturnal hemoglobinuria. Certain immunological aspects of the hemolytic mechanism with special reference to complement. J. Clin. Invest. 18: 657.

8. PACKMAN, C. H., S. 1. ROSENFELD, D. E. JENKINS, and J. P. LEDDY. 1980. Complement lysis of human erythrocytes: II. A unique interaction of human C8 and C9 with paroxysmal nocturnal hemoglobinuria erythrocytes. J. Immuno!. 124: 2818.

9. ROSSE, W. F. 1973. Variations in the red cells in paroxysmal nocturnal hemoglobinuria. Br. J. Haematol. 24: 327.

10. ROTHER, D., G. HANSCH, E. W. RAUTERBERG, H. JUNGFER, and K. ROTHER. 1978. Deviated lysis: Lysis of unsensitized cells by complement. V. Generation of the activity by low pH or low ionic strength. Z. Immunitatsforsch. Immunobiol. 155: 118.

11. HAMMER, c., G. HANSCH, H. GRESHAM, and M. L. SHIN. 1982. Acid activation of C5 causes a-chain cleavage in the presence of C6. Fed. Proc. 41,485: 1258.

12. HANSCH, G., C. H. HAMMER, M. M. MAYER, and M. L. SHIN. 1981. Activation of the fifth and sixth component of the complement system: Similarities of C5b6 and C (56)' with respect to lytic enhancement by cell bound C3 or A2C, and species preferences of target cell. J. Immunol. 127: 999.

13. KABAT, E. A., and M. M. MAYER. 1961. Experimental Immunochemistry. 2nd ed. Charles C. Thomas, Springfield, Ill. p. 149.

14. ROTHER, D., and K. ROTHER. 1961. Uber einen angeborenen Komplementdefekt bei Kaninchen. Z. Immunitatsforsch. 121: 224.

126 . GERTRUD HANSCH, C. HAMMER, R. JI]I, URSULA ROTHER, and MOON SHIN

15. HAMMER, C. H., M. M. FRANK, G. WIRTZ, L. RENFER, H. GRESHAM, and B. TACK. 1981. Large scale isolation of functionally active components of the human complement system. J. BioI. Chern. 256: 3995.

16. HAMMER, C. H., A. NICHOLSON, and M. M. MAYER. 1975. On the mechanism of cytolysis by complement: Evidence on insertion of CSb and C7 subunits of the CSb67 complex into phospholipid bilayers of erythrocyte membranes. Proc. Nat!' Acad. Sci. 72: 5067.

17. TAMURA, N., and A. SHIMADA. 1971. The ninth component of guinea pig complement. Isolation and identification as an a-2 globulin. Immunology 20: 415.

18. JUNGFER, H. 1975. Immunoadsorption mit Protein-Glas-Derivaten. Arzt!' Lab. 21: 80. 19. LAURELL, C. B. 1965. Antigen-antibody crossed electrophoresis. Analyt. Biochem. 10:

358. 20. RApp, H. P., and T. BORSOS. 1970. Molecular basis of complement action. Meredith

Corporation, New York, p. 129. 21. ROSENFELD, S. I., C. H. PACKMANN, D. E. JENKINS, J. K. COUNTRYMAN, and G. P.

LEDDY. 1980. Complement lysis of human erythrocytes. III. Differing effectiveness of human and guinea pig C9 on normal and paroxysmal nocturnal hemoglobinuria cells. J. Immunol. 125: 2063.

22. LOGUE, G. L., W. F. ROSSE, and G. P. ADAMS. 1973. Mechanisms of immune lysis of red blood cells in vitro. Paroxysmal nocturnal hemoglobinuria cells. J. Clin. Invest. 52: 1129.

Dr. GERTRUD HANSCH, Institut fur Immunologie der Universitat Heidelberg, 1m Neuenheimer Feld 305, D-6900 Heidelberg, F.R.G.