Complement-mediated neurotoxicity is regulated by homologous restriction

BRAIN RESEARCH ELSEVIER

Brain Research 671 (1995)282-292

Research report

Complement-mediated neurotoxicity is regulated by homologous restriction Yong Shen a,*, Jose A. Halperin

b

Chi-Ming Lee

a,t

a Neuroscience Department, APIO, D47U, Pharmaceutical Products" Ditqsion, Abbott Laboratories, lOOAbbott Park Road, Abbott Park, IL 60064-3500, USA b Laboratory for Membrane Transport and Department of Medicine, Brigham and Women ~ Hospital, Hart~ard Medical School, Boston, MA 02115, USA Accepted 18 October 1994

Abstract

The ability of /3-amyloid peptides to activate the classical complement cascade and the presence of various complement proteins including the membrane attack complex (C5b-9) on dystrophic neurites in Alzheimer's disease brains, raises the possibility that the complement system may contribute to this neurodegenerative disorder. To address this issue, we have studied the effect of complement activation on nerve growth factor (NGF)-differentiated rat pheochromocytoma PC12 cells, and on retinoic acid (RA)-differentiated human neuroblastoma SH-SY5Y cells. Although incubation of both cell types with human serum resulted in activation of complement, as indicated by iC3b formation, only PC12 but not SH-SY5Y cells were killed by human serum treatment. In contrast, heat-inactivated serum (56°C, 45 min) was not neurotoxic. On SH-SY5Y cells, both PCR amplification and immunocytochemistry demonstrated the presence of CD59, a gtycosylphosphatidylinositol-anchored protein that restricts homologous complement activation by inhibiting the formation of the membrane attack complex. The presence of CD59 probably accounts for the inability of human complement to lyse the human cell lines. Indeed, removal of glycosylphosphatidylinositol (GPI)-anchored proteins with phosphatidylinositol-specific phospholipase C (PI-PLC) rendered SH-SY5Y cells vulnerable to complement attack and eventually led to serum-mediated cell death. Reconstituted C5b-9 was also toxic to both PC12 and PI-PLC-pretreated SH-SY5Y cells. These observations suggest that complement activation can cause neuronal cell death and that this process is regulated by homologous restriction.

Keywords: Neurodegeneration; Cell death and membrane attack complex; CD59 complement inhibitor I. Introduction

Alzheimer's disease (AD) is characterized by neurofibrillary tangles, senile plaques which contain the /3-amyloid peptide (A/3) and neuronal loss [1,36]. Although the mechanism underlying neurodegeneration in A D is still unknown, recent evidence suggests that complement proteins may play a pivotal role in its pathogenesis and contribute to the neuropathological features observed in brains of A D patients [15,21,22, 25,32,33].

* Corresponding author. Fax: (1) (708) 937-9195. l Present address: Institute for Dementia Research, Miles Inc., 400 Morgan Lane, West Haven, CT 06516, USA. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 4 ) 0 1 2 6 4 - 4

The complement system is one of the major humoral defense mechanism against infection. Its activation via either the classical or the alternative pathway can increase inflammatory mediator production, recruit phagocytic cells and lyse target cells by the formation of membrane attack complex. In addition, the terminal complement complexes can induce transient changes in membrane permeability resulting in Na +, K + and Ca 2+ fluxes and the release of cytokines from the target cells, without leading to colloidosmotic lysis [10,11]. Numerous complement glycoproteins act in a cascade to mediate phagocytosis and cytolysis. Activation of the classical pathway can be initiated with binding of Clq to the Fc region of immunoglobulin. This triggers a cascade of proteolytic events resulting in the activation of C5 convertase which cleaves C5 into C5b and C5a. The C5b binds C6, C7, C8 and gives rise to a C5b-8 complex. Binding of C9 molecules to C5b-8 form C5b-9,

Y. Shen et al. / Brain Research 671 (1995) 282-292

the ' m e m b r a n e attack complex' ( M A C ) which pore size increases as the n u m b e r of C9 in the complex increases. If this m e m b r a n e lesion persists a n d results in u n c o n t r o l l e d ion fluxes, the ceils swell a n d e v e n t u a l l y lyse [12,17,29]. If the lesion is t r a n s i e n t , c h a n g e s in m e m b r a n e p o t e n t i a l a n d ionic c o m p o s i t i o n of the cells can activate cell signaling pathways [10,13,28]. E v i d e n c e for a n i n v o l v e m e n t of c o m p l e m e n t activation in the p a t h o g e n e s i s of A D includes: (a) the presence of focal i m m u n o r e a c t i v i t y for various c o m p l e m e n t c o m p o n e n t s ( i n c l u d i n g C l q , C4d, C3b, C3c, C3d a n d C5b-9) o n d e g e n e r a t i n g e l e m e n t s ( i n c l u d i n g senile plaques, tangles a n d dystrophic n e u r i t e s ) in b r a i n s of A D p a t i e n t s b u t n o t in those of a g e - m a t c h e d n o r m a l subjects [6,21,22,32,33]; (b) a higher level of expression of C l q in b r a i n regions displaying m o r e A D pathology, e.g. s u p e r i o r frontal cortex versus c e r e b e l l u m [2,19]; (c) an i n c r e a s e d expression of m R N A for C l q B , C3 a n d C4 in b r a i n s of A D p a t i e n t s t h a n in a g e - m a t c h e d controls [16,34]; (d) A f t e n h a n c e s p r o d u c t i o n of comp l e m e n t C3 by microglial cells in c u l t u r e (9); (e) A/3 b i n d s to C l q a n d activates the classical c o m p l e m e n t cascade in a n a n t i b o d y i n d e p e n d e n t m a n n e r [32]. Thus, the d e p o s i t i o n of A f t in amyloid p l a q u e s may trigger the activation of the classical c o m p l e m e n t cascade a n d c o n t r i b u t e to the p a t h o g e n e s i s in A D . However, in spite of the p r e s e n c e of b o t h early a n d late c o m p l e m e n t c o m p o n e n t s in A D brains, i n d i c a t i n g the activation of the full classical c o m p l e m e n t cascade, t h e r e is n o direct evidence that c o m p l e m e n t activation could kill c e n t r a l n e u r o n s . I n d e e d , most n u c l e a t e d cells protect themselves against a u t o l o g o u s c o m p l e m e n t attack by expressing m e m b r a n e - a s s o c i a t e d c o m p l e m e n t regulatory p r o t e i n s [18] a n d by ' s h e d d i n g ' the M A C from their cell surfaces [4]. I m p o r t a n t l y , in parallel with a n i n c r e a s e d expression of c o m p l e m e n t c o m p o n e n t s , b r a i n s of A D p a t i e n t s exhibit a n i n c r e a s e d expression of c o m p l e m e n t r e g u l a t o r y p r o t e i n s such as clusterin a n d CD59 [5,20,24,25]. U p - r e g u l a t i o n of c o m p l e m e n t regulatory p r o t e i n s has b e e n hypothesized to limit further c o m p l e m e n t attack o n cells in the vicinity of c o m p l e m e n t activation. W h e t h e r n e u r o n s express adeq u a t e a m o u n t s of these r e g u l a t o r y p r o t e i n s to protect themselves from c o m p l e m e n t attack r e m a i n s unexplored. I n the p r e s e n t study, we address the q u e s t i o n w h e t h e r c o m p l e m e n t activation is n e u r o t o x i c a n d w h e t h e r this process is r e g u l a t e d by c o m p l e m e n t regulatory p r o t e i n s in a species d e p e n d e n t m a n n e r (hom o l o g o u s restriction).

283

(DMEM), containing 7% heat-inactivated horse serum (GIBCO, Grand Island, NY, USA), 7% fetal calf serum, 10 nM /3-NGF (Becton Dickinson Labware, Bedford, MA, USA) and 50 /~g/ml penicillin-streptomycin (GIBCO, Grand Island, NY, USA). Cells were seeded at 5000 cells/well in matrigel-treated 24 well plates (Costar, Cambridge, MA, USA). The medium was replaced every 3 days. Human neuroblastoma SH-SY5Y cells were cultured in 50% MEM and 50% F12 media, 15% fetal calf serum and 10 ~M retinoic acid. The 24-well plates were coated with matrigel (Collaborative Biomedical Products, Two Oak Bedford, MA, USA) at 1:20 dilution for 3 h at room temperature and then cells were seeded at 25,000 cells/well. The medium was replaced every 3 days. Both cells were differentiated with either NGF or RA for 6 days. On day 7 after seeding, cells became neurotypic with multiple neurites. Various treatments were given to the cells in serum-free media with 100 x N2 supplement (GIBCO, Grand Island, NY, USA). 2.2. Pl-specific phospholipase C (PI-PLC) treatments

PI-PLC stock (20 units/ml, Funakoshi Co. Ltd., Tokyo) was diluted to 5 units/ml with serum free media. Two different concentrations of PI-PLC (0.25 unit/well and 0.04 unit/well) were used. SH-SY5Y cells were preincubated with PI-PLC at 37°C for 60 min prior to treatment with human serum or reconstituted C5b-9. 2.3. Complement preparations Human serum preparation. Human blood specimens were collected from 8 healthy donors in separate containers. The samples were allow to clot at room temperature for 30 min and then incubated at 4°C for 45-90 min. They were centrifuged at 800 X g for 15 min at 4°C. Each serum was transferred into a separate 15 ml polypropylene tube and spun down at 1000× g for 15 min at 4°C. At this point, all sera were pooled into a 50 ml polypropylene tube, aliquoted (250/xl) into microfuge tubes and then stored at -80°C. Heat-inactivated human serum: to inactivate complement, human serum was incubated at 56°C for 45 min. The heated serum, unlike normal serum, was unable to lyse red blood cells in the CH50 assay (Sigma, St. Louis, MO, USA) indicating inactivation of complement. Normal serum or heat-inactivated serum was diluted into various concentrations (3.3, 5, 10 and 20%) with serum-free media and incubated with the differentiated neuronal cells at 37°C for 1, 2, or 3 days. Reconstituted C5b-9 preparation. C5b-9 was reconstituted as previously described [11]. Briefly, C5b-6 was formed from C5, C6, factor B and cobra venom factor and recombinant factor D and isolated by HPLC on a diethyl aminoethyl column. Then the C5b-6 mixture was added into buffer A (60 mM NaCI, 10 mM sodium phosphate, pH 7.6) and eluted with a linear 60 min gradient of of buffer A and buffer B (500 mM NaCI, 10 mM sodium phosphate, pH 7.6). C5b-6 was eluted as a distinct peak with a characterizatic absorption spectra at 220, 250 and 280 nm (Diode Array Spectrophotomer 1040A; Hewlett-Packard Co., Avondale, PA) and then was titrated to 4 hemolytic units with serum free media. To reconstitute C5b-9, C5b-6 was incubated with C7 (30 ~g/ml) for 5 min at 37"C. C8 and C9 (30 mg/ml, Quidel, San Diego, CA, USA) were then added and incubated with target ceils at 37°C for 1, 2 or 3 days. 2.4. Cytotoxicity assay

2. Materials and methods 2.1. Cell culture

Rat pheochromocytoma PC12 cells were grown in the presence of 95% O 2 and 5% CO 2 in Dulbecco's modified Eagle's medium

The cytotoxic effect of complement activation was examined morphologically and biochemically. Morphologically, a degenerating neuron was characterized by a swollen, chromatic or centric cell body with broken or swollen neurites. Degenerating neurons have a tendency to round-up and detach from the substratum they grew on. Biochemically, the release of lactate dehydrogenase (LDH) from

Fig. 1. Neurotoxic effect of human serum on NGF-differentiated PC12 cells. Phase-contrast micrographs of neuronal cells. A: control, no treatment. B,C and D: cells were treated with 5%, 10% and 20% of human serum for 3 days. Swollen cell bodies, disintegrating neurites are apparent in human serum treated cells.

b,a

b,a

~a

2

4~

t,a

Y. Shen et al. / Brain Research 671 (1995) 282-292 degenerating neurons was measured using CytoTox 96 non-radioactive cytotoxicity assay kit (Promega, Madison, WI, USA) as described by the supplier. LDH is a stable cytosolic enzyme which is released upon cell lysis. Fifty microliters of cell supernatant or cell lysate were added into a substrate mix which contains a tetrazolium salt (INT). LDH converts INT into a red formazan product. The intensity of color formed (recorded as absorbance at 492 nm) is proportional to the number of lysed cells. Percentage of LDH release was calculated as the portion of LDH in supernatant over total LDH from both supernatant and cell lysate.

285

100.

Day I 80

•

Day 2

~

Day 3

_J

2.5. iC3b Irnmunoassay Complement activation leads to the activation of C3 convertase which cleaves C3 into two fragments, C3a and C3b. The nascent C3b reacts covalently through its thioester bond onto activating surfaces and becomes the focus of further proteolytic processing giving rise to iC3b. In our experiments, iC3b was measured after cells were incubated with human serum for 24 h at 37°C with the Quidel iC3b EIA kit as described by the supplier (Quidel, San Diego, CA). Control samples consist of human serum incubated under the same condition in the absence of cells. The IgG we used for positive control was heat aggregated.

0

c,;,,,,

; Huemn

" serum

2'0

,', ¢on©entrstlen

(%)

Fig. 2. Effects of human serum on LDH release in NGF-differentiated PC12 cells. Cells were exposed to various concentrations of human serum (3.3%, 5%, 10% and 20%). LDH release was measured at the indicated time intervals. Values represent the mean + S.D. of 6 separate determinations ( * P < 0.05, * * P < 0.001 when compared to untreated control cells by Student's paired t-test).

2.6. PCR blot experiments First strand eDNA reverse-transcribed from total RNA isolated from differentiated SH-SY5Y, human liver, PC12 cells or rat liver was amplified by PCR to detect the expression of CD59. Briefly, equal amounts of (10/~g) total RNA were converted to first-strand cDNA by RNaseH-reverse transcriptase with oligo(dT) primer (Promega, cDNA Kit). Each template eDNA sample was amplified using the same master PCR reaction mixture which contains 10 mM Tris-HCl (pH 8.3), 3 mM MgCI 2, 0.1% glycogen, 5 mM of each dNTP and 2.5 units of Taq DNA polymerase (Perkin-Elmer/Cetus) and 1 /.LM each of the following primers: 5'-GTTGACTTAGGGATGAAGGCTCCAG-3' and 5'-CAATGGGAATCCAAGGAGGGTCTG-3'. After a first denaturing step at 94°C for 5 min, the cycle was 68°C annealing for 2 min, extension at 72°C for 2 min, denaturing step at 94°C for 1 min. The total cycles were 35 with a final extension step of 15 min. A portion (25%) of the reaction mixture was run in 1% agarose gel and then vacuum transfered to GeneScreen Plus membrane as recommended by the manufacturer (Pharmacia, Piscataway, N J). The blot was prehybridized at 42°C for 60 min in 0.5 M phosphate buffer (pH 7.8), 7% SDS, 1 mM EDTA, 5 × denhardt's reagent, 50% formamide, 6× SSC, 10/zg/ml Salmon sperm DNA and then hybridized in the above solution at 42°C for 20 h with a human CD59 specific oligonucleotide probe: 5'-GTTAGGACAGTI'GTAGCACTGCAGG CTATGACC TGAATGG-3'. The oligonucleotide was labelled with 32p-dATP and tailed by TdT and the labelled probe was cleaned by Nuc Trap Column (Stratagene). The blot was washed twice in 2 × SSC at 65°C and exposed to XAR-5 autoradiographic film (Kodak) at - 80°C with an intensifying screen.

ondary antibody at 1:200 dilution (Fluorescein Isothiocyanate, FITC, Sigma, St. Louis, MO, USA) at 37°C for 30 min. The cells were then rinsed 3 times in PBS and distilled water, air-dried and mounted in Pro-tex mounting medium (Lerner Laboratories, Pittsburg, PA, USA). The slides were examined under a fluorescent microscope.

3. Results To determine whether complement activation could cause neuronal cell death, initial experiments were conducted on NGF-differentiated rat pheochromocytoma PC12 cells using normal human serum as a source

100

[-I [ ] Human ~

8o

•

(s'/.)

MAC

60 e o -e

4O

q t

2. 7. CD59 Immunocytochemistry SH-SY5Y cells were differentiated with retinoic acid as described above in chamber slides (NUNC, Naperville, IL, USA). The cells were washed in phosphate-buffered saline (PBS) and then the cells on slides were fixed with 1% formaldehyde at room temperature for 10 min. After rinsing in PBS, cells were incubated in 100/~i of 4% bovine serum albumin (BSA) at 37°C for 10 min to block non-specific binding sites. The cells were incubated with a rat monoclonal antibody against human CD59 at 1:400 dilution (Bioproducts for Science, Indianapolis, IN, USA) at 37°C for 30 rain and then washed 3 times in PBS. This was followed by incubation with a goat anti-rat sec-

t

2O

0

Day I

Day 2

Tlme

Day 3

course

Fig. 3. Comparison of LDH release induced by reconstituted human C5b9 (MAC) and human serum in PC12 cells. Cells were exposed to either reconstituted C5b9 (equivalent to 4 hemolytic units) or 5% human serum. LDH release was measured at the indicated time intervals. Values represent the mean + S.D. of 6 separate determinations. * P < 0.05, * * P < 0.001 when compared to untreated control cells by Student's paired t-test.

toxicity is apparent in human serum treated cells.

Fig. 4. Effect of human serum on RA-differentiated SH-SY5Y cells. A: control, no treatment; B,C and D: cells were treated with 5%, 10% and 20% of human serum for 3 days. Little or no

t,o

t,o

I

~e t,~

e~

e~

Y. Shen et al. / Brain Research 671 (1995) 282-292

of human complement. As illustrated in Fig. 1, treatment of PC12 cells with various concentrations of human serum at 5, 10 and 20% for 3 days resulted in cell injury characterized by swollen cell body, degenerating neurites and tendency to round-up and detach from the culture substratum. In addition to these morphological changes, cell injury by human serum was further substantiated by L D H measurements. As depicted in Fig. 2, normal complement sufficient human serum caused a time (1-3 days) and dose (3.3 to 20%)-dependent increased in L D H release from rat PC12 cells. Significant increase in L D H release was observed after 1 day exposure to 5, 10 and 20% human serum. To further explore the neurotoxic effect of complement, we studied the effects of reconstituted human membrane attack complex (MAC). A time-dependent increase in L D H release from PC12 cells was observed upon treatment with MAC. The effect of MAC (equivalent of 4 hemolytic units) was comparable to that induced by 5% human serum and significant increase in L D H release was observed after 2 and 3 days of treatment (Fig. 3). Unlike the rat PC12 cells, complement sufficient human serum showed no significant toxic effect on RA-differentiated human SH-SY5Y ceils on the basis of morphology (Fig. 4) and L D H measurements (Fig. 5). This differential susceptibility to human serum toxicity could be due to the presence of species specific complement regulatory proteins on the cell surface that protect human SH-SY5Y cells against human complement attack (a phenomenon commonly referred to as homologous restriction). Since homologous restriction complement inhibitors identified to date are GPI-anchored membrane proteins, we tested the sus100. Oily 1

80.

A •" - - G - - •

o~y2 01y8 Drt 1-NO PL¢

* *

60.

| 40"

0

Conlrol

313 Human

; serum

concentration

1;

2;

(%)

Fig. 5. Effect of human serum on LDH release in RA-differentiated SH-SY5Y cells with or without PI-PLC (0.25 U/ml) pretreatment. Cells were exposed to various concentrations of human serum (3.3, 5, 10 and 20%). LDH release was measured at the indicated time intervals. Values represent the mean + S.D. of 6 separate determinations ( * P < 0.05 and * * P < 0.001 when compared with untreated controls by Student's paired t-test).

287

ceptibility of SH-SY5Y cells to human complement attack after treatment with PI-PLC. If SH-SY5Y cells are protected by homologous restriction, it is expected that PI-PLC pretreatment will render them complement sensitive by removing the GPI-anchored complement regulatory proteins. Indeed, complement sufficient human serum induced a significant increase in L D H release from SH-SY5Y cells after PI-PLC treatment (0.25 U / m l , 60 min, Fig. 5). This effect was apparent with as little as 3.3% serum after 1 day. In agreement with the L D H measurements, the neurotoxic effect of human serum on PI-PLC pretreated SH-SY5Y cells was confirmed by morphological changes typical of injured neurons (Fig. 6). As expected, PI-PLC pretreated SH-SY5Y cells were also susceptible to human reconstituted MAC-induced lysis and a significant increase in L D H release was observed after a 3 days incubation with the equivalent of 4 hemolytic units of MAC (Fig. 7). The neurotoxic effect of human serum on PC12 and PI-PLC pretreated SH-SY5Y cells as measured by L D H release can be inhibited by heat-inactivation of serum complement (Table 1) and this was confirmed by morphological changes (data not show). These results further support the notion that serum-induced neurotoxicity is complement-mediated. To monitor complement activation, we measured the level of iC3b, which is an activation product of C3, after incubations with human serum. As a positive control, incubation of 5% human serum with 1 t z g / m l of heat aggregated human IgG at 37°C for 40 min induced the formation of 10 t z g / m l of iC3b (Fig. 8). Incubation of 5% human serum in culture dishes alone or in the presence of matrigel (substratum for cells) for 24 h at 37°C induced little or no iC3b formation ( < 2 /zg/ml). However, a large increase (40 /zg/ml) of iC3b level was detected after incubating PC12 and SH-SY5Y cells with 5% human serum for 24 h (Fig. 8). These results confirmed that exposure of neuronal cells to human serum activates the complement cascade. To examine whether the homologous restriction complement membrane inhibitor CD59 is expressed in the neuronal cells used in our experiment, we isolated total R N A from these cells. Then the R N A was reverse transcribed to cDNA with oligo primer and finally the c D N A mixture was amplified by polymerase chain reaction using oligo primers specific for human CD59. With this procedure we detected the expected 396 base pairs (bp) fragment in the SH-SY5Y but not PC12 cells (data not show). The segment Size is identical to the fragment obtained from human liver cDNA which contains abundant CD59 and also to that previously described for CD59 in brains of AD patients (McGeer et al., 1991) [24]. The hybridization of PCR blot with a specific oligonucleotides probe for CD59 further confirmed this result (Fig. 9). The expression of CD59 at

Fig. 6. Effect of human serum on RA-differentiated SH-SY5Y cells with PI-PLC (0.25 U / m l ) pretreatment. A: control, no treatment. B,C and D: cells were treated with 5%, 10% and 20% of human serum for 3 days. Swollen cell bodies, disintegrating neurites are apparent in human serum treated cells.

t

t~a

I

t,,a

x._

~z

e~

}i.. Shen et al. / Brain Research 671 (1995) 282-292

289

100.

[]

Comrd

+ 80.

Human liver

| 40.

O

*

*

S

S H -S Y S Y Rat

20"

liver

PC1 2 0"

Day 1

Day 2 Tlmo

Day 3

course

Fig. 7. Effects of reconstituted M A C (C5b-9) on L D H release in RA-differentiated SH-SY5Y cells. Cells were exposed to 4 hemolytic units of C5b-9 for 1 - 3 days. L D H release was m e a s u r e d at the indicated time intervals. Values represent the m e a n + S . D , of 6 separate determinations. * P < 0.001 when compared to untreated control cells by Student's paired t-test. Table 1 Comparison of cytotoxic effects of normal and heat-inactivated hum a n s e r u m on PC12 cells and PI-PLC pretreated SY5Y cells

Fig. 9. P C R blot analysis of cell expression of CD59 complement inhibitor. First-strand c D N A was transcribed from total R N A by using oligo (dT) as a primer and finally the c D N A mixture was amplified by polymerase chain reaction using oligo primers specific for h u m a n CD59. The hybridization of P C R blot with a specific oligonucleotides for CD59 further confirmed P C R result. With this procedure we detected the expected 396 base pairs (bp) fragment in the SH-SY5Y but not PC12 cells. Lanes (from left to right): h u m a n liver, SH-SY5Y, and rat liver and PC12.

Cell types and treatments

L D H release (%) M e a n + SD (n = 6)

NGF-differentiated PC12 cells

Normal serum

Heat-inactivated serum

the protein level in SH-SY5Y cells was further confirmed by immunofluorescent staining of CD59. Control SH-SY5Y cells exhibited a clear CD59 immunofluorescent signal (Fig. 10A) which was greatly reduced upon PI-PLC treatment (0.25 U / m l ) at 37°C for 60 min (Fig. 10B).

No treatment control 3.3% 5% 10%

6.8 + 1.6 12.8+2.7 23.6+ 1.3 * 32.1+2.6 * *

6.3 + 0.5 6.0+ 1.5 8.0-1-2.1 13.8+3.2 *

4. Discussion

7.8 + 1.2 25.2+5.7 * 24.2 + 7.1 * 38.5+9.2 * *

7.8 + 1.2 7.1+2.0 9.9 + 0.9 18.5+4.8 *

RA-differentiated S Y S Y cells (with PI-PLC pretreatment) No treatment control 3.3% 5% 10% • P < 0.05; * * P < 0.001

Note: comparison was m a d e between no treatment control and various treatment groups.

50

_~E

• •

40

• [] []

_ ~o

5% Serum: SH-SY5Y 5% Serum:PC12 s% .Serum: ~gG 5% Serum Alone 5% Serum: Matrigel

'3

_g 20-

g, 11]

L

Serum

.........

Condalons

Fig. 8. Increase in iC3b immunoreactivity in h u m a n s e r u m induced by SH-SY5Y cells, PC12 cells and heat aggregated h u m a n IgG. Cells were exposed to 5% h u m a n serum for 24 h. T h e cell supernatant in different groups were collected and iC3b formation was measured using a commercial iC3b E L I S A kit as described by the supplier.

The present results show a time- (1-3 days) and dose-dependent (3.3 to 20%) neurotoxic effect of complement sufficient human serum on NGF-differentiated PC12 rat pheochromocytoma cells. Cells plated at a higher cell density (35,000cells/well) were more resistant to human serum although higher serum concentrations (10-20%) were toxic even in the high cell density culture (data not show). Three different lines of evidence which suggest that the toxic effect of human serum on neuronal cells is complement-mediated. First, this serum toxicity can be inhibited by heat-inactivation of serum complement. Secondly, the level of iC3b, an activation product of complement cascade, was increased after incubating the neuronal cells with human serum containing media. Thirdly, reconstituted human C5b-9 also exerts a time-dependent cytotoxic effect on PC12 cells. On the other hand, complement sufficient human serum exerted no significant toxic effect on RA-differentiated SH-SY5Y human neuroblastoma cells. We have provided two different lines of evidence to suggest that these human neuronal cells are protected from autologous complement attack by homologous restriction. First, we have demonstrated by PCR and immunocytochemistry that the SH-SY5Y cells do express CD59 at both m R N A and protein levels. CD59 is a 18-20-kd glycoprotein discovered first in erythro-

290

Y Shen et aL /Brain Research 671 (1995) 282-292

Fig. 10. lmmunocytochemistry of CD59 in differentiated SH-SY5Y cells. A: expression of CD59 protein with no treatment. B: lower expression of CD59 with PI-PLC treatment (0.25 U / m l , 37°C, 60 min).

Y. Shen et al. / Brain Research 671 (1995) 282-292

cyte cell membrane [14]. It protects red blood cells from autologous complement attack by inhibiting the cytolytic activity of complement and hence is referred to as a homologous restriction factor, membrane inhibitor of reactive lysis or human protectin [27]. Upregulation of CD59 has been demonstrated in AD brains [24], however, its cellular origin (neurons, glial cells or endothelial cells) has not yet been defined. Astrocytes but not oligodendrocytes have been demonstrated to express CD59 in culture [37,8]. The present study suggests that neuronal cells are able to express CD59 in culture and they are resistant to autologous complement attack. Secondly, treating SH-SY5Y cells with PI-PLC to remove CD59 made them sensitive to human serum and reconstituted C5b-9 toxicity. Complement-mediated lysis of rat sympathetic nerves has been reported after treatment with dopamine-/3-hydroxylase antibodies in vitro [7] and acetylcholinesterase antibodies in vivo [3]. The ability to activate complement has been reported for oligodendrocytes in the CNS [35,39]. However, unlike SH-SY5Y ceils, oligodendrocytes are sensitive to autologous complement attack presumably due to lack of homologous restriction factors. In contrast to the complement-mediated lysis of erythrocytes where the event is 'colloidosmotic' and fast in onset, the killing of neuronal cells in our experiments appears to take place over a longer time course (1-3 days). The mechanism underlying this 'delay' killing remains to be studied. The ability of sublytic concentrations of MAC to induce changes in membrane permeability to Na + , K ÷ and Ca 2+, deplete ATP, release cytokines, eicosanoids and reactive oxygen radicals [28,10,31,13,38] may contribute to a delayed nucleated cell death observed in the present study. In conclusion, we have provided evidence that complement activation can kill neuronal cells and that homologous restriction plays an important role in regulating this process. If these in vitro observations are representative of the complement-neuron interactions in vivo, they would have important therapeutic implications for AD. In addition to develop direct complement inhibitors to reduce the /3-amyloid mediatedcomplement activation, up-regulating neuronal homologous restriction factor would probably represent an alternative therapeutic approach for this common neurodegenerative disorder.

Acknowledgments We would like to thank Laura Benzaquen, Marc Brande, Diane Casuto, Terry Pederson, Mike Russell and Traci Sullivan for providing cells and technical assistance. We also thank Dr. June Biedler of Memo-

291

rial Sloan-Kettering Cancer Center, New York for providing an initial stock of SH-SY5Y cells.

References [1] Braak, H., Braak, E. and Bohl J., Staging of Alzheimer-related cortical destruction, Eur. Neurol., 33 (1993) 403-408. [2] Brachova, L., Lue, L.-F., Schultz, J., El Rashidy, T. and Rogers, J., Association cortex, cerebellum, and serum concentrations of Clq and factor B in Alzheimer's disease, Mol. Brain Res. 18 (1993) 329-334. [3] Brimijoin S., Moser, V., Hammond, P., Oka, N. and Lennon, V.A., Death of intermediolateral spinal cord neurons follows selective complement-mediated destruction of peripheral preganglionic sympathetic terminals by acetylcholinesterase antibodies, Neuroscience, 54 (1993) 201-223. [4] Carney, D.F., Hammer, C.H. and Shin, M.L., Elimination of terminal complement complexes in the plasma membrane of nucleated cells: influence of extracellular calcium and association with cellular calcium, J. lmmunol., 137 (1986) 263-273. [5] Duguid, J.R., Bohmont, C.W., Liu, N. and Tourtellotte, W.W., Changes in brain gene expression shared by scrapie and Alzheimer disease, Proc. Natl. Acad. Sci., 86 (1989) 7260-7265. [6] Eikelenboom, P., Rozemuller, J.M., Kraal, G., Stam, F.C., McBride, P.A., Bruce, M.E. and Fraser, H., Cerebral amyloid plaques in AIzheimer's disease but not in Scrapie-affected mice are closely associated with a local inflammatory process, V/rchows Arch. B Cell Pathol., 60 (1989) 329-336. [7] Furness, J.B., Lewis, S.Y., Rush, R., Costa, M. and Geffen, L.B., Involvement of complement in degeneration of sympathetic nerves after administration of antiserum to dopamine fl-hydroxylase, Brain Res., 136 (1977) 67-75. [8] Gordon, D.L., Sadlon, T., Hefford, C. and Adrian, D., Expression of CD59, a regulator of the membrane attack complex of complement, on human astrocytes, Mol. Brain Res., 18 (1993) 335-338. [9] Haga, S., Ikeda K., Sato, M. and Ishii, T., Synthetic Alzheimer amyloid/3/A4 peptides enhance production of complement C3 complement by cultured microglial cells, Brain Res., 601 (1993) 88-94. [10] Halperin, J.A., Nicholson-Weller, A., Brugnara C. and Tosteson, D.C., Ca++-activated K ÷ efflux limits complement-mediated lysis of human erythrocytes, J. Clin. Invest., 82 (1988) 594600. [11] Halperin, J.A., Tatatuska, A., Rynkiewicz, M. and NicholsonWeller, A., Transient changes in erythrocyte membrane permeability are induced by sublytic amounts of the complement membrane complex (C5b-9), Blood, 81 (1993) 200-205. [12] Hamilton, K.K., Zhao, J., Rollins, S., Stewart, B.H. and Sims, P.J., Regulatory control of the terminal complement proteins at the surface of human endothelial cells: neutralization of a C5b-9 inhibitor by antibody to CD59, Blood, 76 (1990) 2572-2577. [13] H~insch, G.M., The complement attack phase: control of lysis and non-lethal effects of C5b.9, lmmunopharmacology, 24 (1992) 107-117. [14] Holguin, M.H., Frederick, L.R., Bernshaw, N.J., Wilcox, L.A. and Parker, C.J., Isolation and characterization of a membrane protein from normal human erythrocytes that inhibits reactive lysis of the erythrocytes of paroxysmal nocturnal hemoglobinuria, J. Clin. Incest., 84 (1989) 7-14. [15] Ishii, T. and Haga, S., Complements, microglial cells and amyloid fibril formation, Res. lmmunol., 143 (1992) 614-616. [16] Johnson, S.A., Lampert-Etchells, M., Pasinetti, G.M., Rozovsky, I. and Finch, C.E., Complement mRNA in the mammalian

292

Y. Shen et al. ~Brain Research 671 (1995) 282-292

brain: response to Alzheimer's disease and experimental lesioning, Neurobiol. Aging, 13 (1992) 641-648. [17] Kim, S.H., Darney, D.F., Hammer, C.H. and Shin, M.L., Nucleated cell killing by complement: effects of Csb_9 channel size and extracellular Ca ++ on the lyric process, J. Immunol., 138 (1987) 1530-1536. [18] Lachmann, P.J., The control of homologous lysis, Immunol. Today, 12 (1991) 313-315. [19] Lue, L.-F. and Rogers, J., Full complement activation fails in diffuse plaques of the Alzheimer's disease cerebellum, Dementia, 3 (1992) 308-313. [20] May, P.C., Lampert-Etchells, M., Johnson, S.A., Poirier, J., Masters, J.N. and Finch, C.E., Dynamics of gene expression for a hippocampal glycoprotein elevated in Alzheimer's disease and in response to experimental lesions in rats, Neuron, 5 (1990) 831-839. [21] McGeer, P.L., Akiyama, H., Itagaki, S. and McGeer, E.G., Immune system response in Alzheimer's disease, Can. Z Neurol. Sci., 107 (1989a) 341-346. [22] McGeer, P.L., Akiyama, H., Itagaki, S. and McGeer, E.G.. Activation of the classical complement pathway in brain tissue of Alzheimer patients, Can. J. Neurol. Sci., 16 (1989b) 516-527. [23] McGeer, P.L., McGeer, E.G., Kawamata, T., Yamada and Akiyama, H., Reactions of the immune system in chronic degenerative neurological diseases, Can. Z Neurol. Sci, 18 (1991) 376-379. [24] McGeer, P.L., Walker, D.G., Akiyama, H., Kawamata, A. L, Guan, A. L., Parker, C.J., Okada, N. and McGeer, E.G., Detection of the membrane inhibitor of reactive lysis (CD59) in diseased neurons of Alzheimer brain, Brain Res., 544 (1991) 315-319. [25] McGeer, P.L. and McGeer, E.G., Complement proteins and complement inhibitors in Alzheimer's disease, Res. Immunol, 143 (1992) 621-623. [26] McGeer, P.L., Kawamata, T. and Walker, D.G., Distribution of clusterin in Alzheimer brain tissue, Brain Res., 579 (1992) 337341. [27] Meri, S., Morgan, B.P., Davies, A., Danials, R.H., Olavesen, M.G., Waldmann, H. and Lachmann, P.J., Human protein (CD59), an 18,000-20,000 MW complement lysis restricting factor, inhibits C5b-8 catalysed insertion of C9 into lipid bilayers, Immunology, 72 (1990) 1-4. [28] Morgan, B.P., Luzio, G.P. and Campbell, A.K., Intracellular Ca + + and cell injury: a paradoxical role for Ca + + in complement membrane attack, Cell Calcium, 178 (1986) 115-139.

[29] Muller-Eberhard, H.J., Zalman, L.S., Chiu, F.J., Jung, G. and Martin, D.E., Molecular mechanisms of cytotoxicity: comparison of complement and killer lymphocytes, J. Rheumatology 14 (1987) 28-34. [30] Ojcius, D.M., Jiang, S. and Young, J.D., Restriction factors of homologous complement: a new candidate? Immunol. Today, 11(1990) 47-49. [31] Papadimitriou, J.C., Ramm, L.E., Drachenberg, C.B., Trump, B.F. and Shin, M.L., Quantitative analysis of adenine nucleotides during the prelytic phase of cell death mediated by C5b_9 , J. ImmunoL, 147 (1991) 212-217. [32] Rogers, J., Cooper, N.R., Webster, S., Schultz, J., McGeer, P., Styren, S.D., Civin, W.H., Brachova, L., Bradt, B., Ward, P. and Lieberburg, I., Complement activation by /3-amyloid in Alzheimer disease, Proc. Natl. Acad. Sci. USA, 89 (1992a) 10016-10020. [33] Rogers, J., Sehultz, S., Brachova, L., Lue, L.-F. Webster, S., Bradt, B., Cooper, N.R. and Moss, D.E., Complement activation and /3-amyloid-mediated neurotoxicity in Alzheimer's disease, Res. Immunol, 143 (1992b) 624-630. [34] Walker, D.G. and McGeer, P.L., Complement gene expression in neuroblastoma and astrocyttoma cell lines of human origin, Mol. Brain Res., 14 (1992) 109-116. [35] Waren, D.R. and Noble, M., Oligodendrocytes and oligodendroeyte/type-2 astrocyte progenitor cells of adult rats are specifically susceptible to the lytic effects of complement in absence of antibody, Proc. Natl. Acad. Sci. USA, 86 (1989) 9025-9029. [36] Whitehouse, P.J., Price, D.L., Struble, R.G., Clark, A.W., Coyle, Joseph, T. and DeLong, M.R., Alzheimer's disease and senile dementia: loss of neurons in the basal forebrain, Science, 215 (1982) 1237-1239. [37] Wing, M.G., Zajicek, J., Seilly, D.J., Compston, D.A.S. and Lachmann, P.J., Oligodendrocytes lack glycolipid anchored proteins which protect them against complement lysis. Restoration of resistance to lysis by incorporation of CD59, Immunology, 76 (1992) 140-145. [38] Wood, A., Wing, M.G., Benham, C.D. and Compston, D.A.S., Specific induction of intracellular calcium oscillations by complement membrane attack on oligodendroglia, J. Neurosci., 13 (1993) 3319-3332. [39] Zajicek, J., Wing, M.G., Lachmann, P.J. and Compston, D.A.S., Oligodendrocyte interaction with human serum - defining the role for complement, J. Neurol. Sci., 108 (1992) 65-72.