complement activation in acquired and hereditary amyloid neuropathy

Journal of the Peripheral Nervous System 5:131–139 (2000)

© 2000 Peripheral Nerve Society, Inc.

131

Blackwell Science Publishers

Complement activation in acquired and hereditaryamyloid neuropathy

Charlene E. Hafer-Macko,

1

Peter J. Dyck,

2

and Carol Lee Koski

1

Department of Neurology, University of Maryland School of Medicine, Baltimore, Maryland,

1

and Department of Neurology, Mayo Clinic and Foundation, Rochester, Minnesota

2

Abstract

The pathogenesis of the axonal degeneration in acquired or hereditary amy-loidosis is unknown. In this immunohistochemistry study, we examined 20 sural nerve bi-opsies from individuals with amyloid neuropathy (14 acquired and 6 hereditary) for evi-dence of complement activation. Complement activation products were detected on andaround amyloid deposits within peripheral nerves. We found no difference in the extent,location or pattern of complement activation products between the 2 forms of amyloido-sis. The presence of early classical pathway activation markers in the absence of antibodyin hereditary cases suggests an antibody-independent activation of the classical pathwaythrough binding of C1q. The lack of Factor Bb–suggested alternative pathway activationwas not significant in these cases. The detection of C5b-9 neoantigen on amyloid depos-its demonstrated that the full complement cascade was activated. Complement activa-tion on amyloid deposits and the generation of C5b-9

in vivo

may contribute to bystanderinjury of axons in the vicinity of amyloid deposits.

Key words:

complement, peripheral neuropathy, hereditary amyloidosis, acquired amyloid-

osis, immunohistochemistry

Introduction

Amyloidosis affects multiple organ systems as a re-sult of the deposition of amyloidogenic proteins as aninsoluble fibrillar material in a beta-pleated sheet config-uration. The deposition and accumulation of these ag-gregated molecules are enhanced by their resistance todegradation and by the abnormal processing of amyloidfibrils by macrophage lysosomes

(Durie et al., 1982;Sommer and Schroder, 1989; Toyooka et al., 1995)

.Within peripheral nerves, amyloid deposition in andaround unmyelinated, small myelinated and large my-elinated fibers is associated with dysfunction of sen-sory, autonomic, and motor nerves. Peripheral neuropa-thy occurs in both the acquired and hereditary forms ofamyloidosis that are caused by the overexpression of

lambda (

λ

) and kappa (

κ

) light chains or point mutationsof the transthyretin (TTR) gene, respectively. Despite

differences in the constituent fibril precursor proteins,these two forms of neuropathy have similar pathologicfeatures in terms of location and degree of amyloiddeposition and severity of axonal degeneration and loss

(Li et al., 1992)

.The precise mechanisms underlying the pathogen-

esis of fiber damage in amyloid peripheral neuropathyremain unknown, but have been postulated to reflectischemia

(Kernohan and Woltman, 1942)

, compression

(Dyck and Lambert, 1969)

and inflammatory

(Chamberset al., 1955)

or toxic effects of amyloidogenic proteins

(Sullivan et al., 1955)

. In Alzheimer’s disease, neuriticplaques contain a different form of amyloid,

b

-amyloid.Aggregated

b

-amyloid activates the classical comple-ment cascade by binding to the collagen tail of C1qcomponent of complement in the absence of specificantibody

(Jiang et al., 1994)

and leads to the formationof the terminal complement complex, C5b-9

(Rogers etal., 1992)

. It is postulated that C5b-9 generated by the

b

-amyloid plaques may contribute to bystander injury ofneural and glial tissue in the vicinity

(Rogers et al.,

Address correspondence to:

Charlene Hafer-Macko, M.D., Depart-ment of Neurology, University of Maryland School of Medicine, 22South Greene Street, Baltimore, MD 21201-1595, USA. Tel: 410-328-6484; Fax: 410-328-5899; E-mail: [email protected]

Hafer-Macko et al. Journal of the Peripheral Nervous System 5:131–139 (2000)

132

1992)

. In the peripheral nerve literature, there is a con-flict regarding a possible role for complement in thepathogenesis of the acquired amyloidosis

(Sommer andSchroder, 1989; Zanusso et al., 1992)

. To clarify this,we examined whether complement activation productswere associated with amyloid deposits in the peripheralnerve and whether complement activation occurred viathe classical pathway in both acquired and hereditaryamyloid neuropathy.

Methods and Materials

Nerve biopsy studies

A series of 20 sural nerve biopsies from patientswith amyloid neuropathy were collected at the MayoClinic in Rochester, Minnesota, from 1976 to 1992. Thesural nerve pathology and percent of axonal degenera-tion in these biopsies were previously characterizedbased on examination of paraffin and semi-thin epoxysections, and teased fiber nerve preparations (Table 1)

(Li et al., 1992)

. Within this patient population, 14 caseswere classified as acquired based on the presence oflight chains in the serum or urine, and 6 hereditarybased on a positive family history or the presence ofpoint mutations in the TTR gene. Classification was re-confirmed in the current study by immunoreactivity ofthe amyloid deposits for

λ

or

κ

light chains (Novocas-tro), and/or TTR (Dako) on the 6-micron paraformalde-hyde-fixed paraffin-embedded sural nerve specimens

(Liet al., 1992)

. Mononuclear cell infiltrates were identifiedby hematoxylin and eosin staining and specific immu-nohistochemical markers for lymphocytes (anti-LCA,Vector) and macrophages (anti-HAM56, Enzo Diagnos-

tics). Congo red and methyl violet staining and immuno-histochemistry for serum amyloid P (Calbiochem) wereused to determine the number and location of amyloiddeposits. A series of 7 control nerves consisted of 6cases of axonal degeneration and 1 case of diabeticneuropathy.

Immunohistochemistry

Complement activation fragments and componentsof the classical and alternative complement pathwayswere detected by immunohistochemistry using specificmonoclonal and polyclonal antibodies and a modified bi-otin-avidin peroxidase method

(Hsu et al., 1981)

. Paraffin-embedded sections were deparaffinized, rehydrated, andmicrowave-treated in 10 mM sodium citrate buffer (pH6.0). Serial incubation with 3% H

2

O

2

: methanol and 10%serum/0.5% nonfat milk/phosphate buffered saline (PBS)blocked endogenous peroxide and nonspecific proteinbinding, respectively. Between each incubation step,slides were rinsed 3 times in PBS. Sections were incu-bated with appropriate dilutions of primary antibody(anti-C1q, C4d, Factor Bb, C5b-9 from Quidel, anti-C3dfrom Dako, and anti-C4 from Sigma) at 4

o

C for 16hours. Slides were sequentially incubated with specificbiotinylated secondary antibodies, avidin biotin complex(ABC Vectastain, Vector), and 0.05% 2,3 diaminobenzi-dine with 0.02% H

2

O

2

for a brown colorimetric reaction.The complement activation product, C3d, was also de-tected by immunohistochemistry on semi-thin epoxysections. Epoxy peripheral nerve sections were etchedwith saturated sodium methoxide in propylene oxideand osmium was removed with hydrogen peroxide.

Table 1.

Patient and sural nerve biopsy description

CaseAcquiredfamilial Age Sex

Lightchain Immunostain Fiber loss

Axonaldegeneration Cells

155-77 acquired 56 M none

l

severe (s

.

1) none253-82 acquired 56 M none

l

severe (s

.

1) severe (50%) none372-90 acquired 77 M IgG

l l

mod/severe active none290-89 acquired 65 F IgG

l l

weak TTR mod/severe (s

.

1) moderate (19%) none45-92 acquired 63 F none

l

weak TTR few fibers remain moderate (22%) few503-92 acquired 73 F none

l.k

mod/severe moderate none534-92 acquired 69 F none

l.k

mod/severe severe (42%) none200-75 acquired 53 M IgD

l l/k

few fibers remain active none249-79 acquired 64 F IgG

l l/k

mod/severe severe none342-86 acquired 63 F IgG

l l/k

severe few216-78 acquired 73 M

k k

mod/severe active none68-86 acquired 56 M

k

BJP

k

mod/severe moderate (26%) none33-88 acquired 68 M IgG

k k

few fibers remain late few115-90 acquired 51 M IgG

k k

few fibers remain active few60-76 familial 57 M none TTR mod/severe active few142-80 familial 67 F none TTR few fibers remain few perineurium428-80 familial 68 M none TTR severe few epineurium296-84 familial 56 M none TTR mod/severe moderate (18%) few122-87 familial 55 F none TTR mod/severe (s

.

1) severe (62%) none127-90 familial 64 M none TTR mod/severe active none

s

.

1, fiber loss is greater in the unmyelinated and smaller myelinated fibers than large myelinated fibers; BJP, Bence Jones proteins.Source:

Li et al., 1992.


133

The sections were treated with 1% sodium borohy-dride, then immunostained as described above for theparaffin-embedded sections.

Results

The 20 amyloid neuropathy cases were classifiedas acquired based on the presence of light chains or he-reditary based on point mutations in the TTR gene orpositive family history (Table 1). Light chain immunore-activity was detected in 14 cases of acquired amyloidneuropathy (

λ

[7],

κ

[4], and both

λ

and

κ

[3]) and tran-sthyretin (TTR) immunoreactivity was detected in 6cases of hereditary amyloid neuropathy

(Li et al., 1992)

.Two cases with

λ

light chains immunoreactivity alsohad weak immunoreactivity with the anti-TTR antibody.In these 2 cases, the weak TTR immunoreactivity wasmost likely nonspecific staining.

The acquired and hereditary neuropathies were notdistinguishable on the basis of pathologic abnormalitiesfrom examination of paraffin and semi-thin epoxy sec-tions, or from teased fiber nerve preparations (Table 1)

(Li et al., 1992)

. In both forms of peripheral neuropathy,axonal degeneration was the predominant pathologicprocess. Most cases had moderate-to-severe loss ofunmyelinated and myelinated fibers. Active ongoing ax-onal degeneration was present to a variable degree.Only a limited number of demyelinated fibers werepresent in a few cases. Mononuclear cells were absentor infrequently observed around epineurial vessels orwithin the endoneurium. In these paraffin sections,macrophages were not associated with amyloid depos-its. In both acquired and hereditary neuropathy, amyloiddeposits were detected in the endoneurium betweennerve fibers by methyl violet metachromasia (Fig. 1)and by amyloid P immunoreactivity (Figs. 2-A and 3-A).Amyloid was also deposited within vessel walls and ina perivascular location around epineurial and endoneur-ial vessels. There were no distinctions in terms of num-ber or location of deposits between the 2 forms of amy-loid neuropathy. There was no correlation with thenumber of amyloid deposits and severity of the patho-logic changes.

Products of complement cascade activation weredetected on or around the amyloid deposits within theendoneurium between nerve fibers, within vessel wallsand in perivascular locations as summarized in Table 2.Complement immunoreactivity was detected on amy-loid deposits and in the endoneurial tissues immedi-ately surrounding the deposits. C3d, a stable degrada-tion fragment of C3b, which covalently binds to targetmembranes, was detected on the surface of amyloiddeposits (Fig. 1). C3d was detected in 16 acquired andhereditary cases. Amyloid deposits immunostained dif-fusely with antibodies directed against C1q (Figs. 2-B

and 3-B) and C4 (Fig. 2-C), markers of early classicalpathway activation, in 11 cases of acquired and heredi-tary amyloid neuropathy. The immunoreactivity for C3dwas more distinct and discrete, relative to the immu-noreactivity for C1q and C4. No Factor Bb immunostain-ing was detected in either acquired or hereditary cases,suggesting that alternative pathway of complementcascade activation was not significant in these cases.The terminal complement complex (C5b-9) was de-tected on a proportion of the amyloid deposits. C5b-9staining was less intense compared with the stableC3d (Fig. 3-C). Although complement activation prod-ucts were clearly associated with endoneurial vesselsand were present on the surface of amyloid deposits,staining appeared to extend beyond the deposit withinthe endoneurium. On these paraffin sections the se-vere axonal loss made it difficult to clearly discernwhether complement activation products were also de-posited on Schwann cells or axons.

Immunohistochemistry on etched-plastic sectionsprovided more precise localization of the complementactivation products in relation to the amyloid depositsaround vessels (Fig. 4). The sural nerve biopsy totalamyloid load did not correlate with the severity of theneuropathologic changes. The severity of the fiber lossand axonal degeneration in these cases probably reflectthe accumulative pathologic changes in the dorsal rootganglia and the proximal portion of these nerves. Therewere no differences in extent of amyloid deposition orthe detection of complement activation products be-tween hereditary and acquired amyloid neuropathycases. Complement immunoreactivity was not detectedin any of the control sural nerve biopsies.

In 16 primary and hereditary amyloid neuropathyspecimens, C5b-9 immunoreactivity was present withinvessel walls, which were thickened by amyloid (Fig. 5-A)and on the endothelial surface with extension into thevessel wall (Fig. 5-B). Complement activation productswere deposited on amyloid deposits within the perivas-cular space surrounding many vessels within the epi-neurium (Fig. 5-C) and endoneurium (Fig. 5-D). Similarstaining was not present in any of the control nerve bi-opsies. The perineurium of 11 sural nerve biopsies hadC5b-9 immunoreactivity (Fig. 5-B). C5b-9 immunoreac-tivity was present in the perineurium of 2 of the controlsural nerve biopsies including 1 individual with diabetes.

Discussion

The pathogenesis of the progressive sensory, mo-tor and autonomic dysfunction associated with amyloidneuropathy is not conventionally considered inflamma-tory in nature. In the current study, we demonstratedcomplement activation products including the terminalactivation complex, C5b-9, on and around amyloid de-


134

Figure 4. Immunohistochemistry on etched plastic sections provides a useful tool to localize complement activation products in rela-tion to amyloid deposits. C3d immunoreactivity (arrow; B, mag 7003 – insert from A) is present in a perivascular location around an en-doneurial vessel (v). A serial plastic section stained with toluidine blue (A) demonstrates an amorphous material surrounding the en-doneurial vessel. Only a few myelinated fibers (arrowheads) remain in this case with severe fiber loss (Case 372-90 λ, mag 2003).

Figure 1. This paraffin-embedded sural nerve from an individual with acquired amyloidosis (λ) demonstrates C3d immunore-activity (brown reaction product – arrowhead) around the periphery of amyloid deposits stained with methyl violet. C3d immu-noreactivity is present around an endoneurial vessel (v). C3d is also observed within the endoneurium, however the precise lo-calization could not be discerned because of the severe axonal loss (Case 249-79, mag 7003) (P – perineurium, En – endoneurium).


135

posits in both acquired and hereditary forms of amy-loidosis. The formation of C5b-9 and its insertion intomembranes in the vicinity of amyloid deposits may me-diate bystander injury of axons, Schwann cells and neu-rons, similar to the mechanisms proposed for the dele-terious effects of complement activation by

b

-amyloidin Alzheimer’s disease

(McGeer and McGeer, 1992)

.One of our principal findings was the substantial de-

tection of C1, C4 and C3d in association with amyloid de-posits in the peripheral nerves of patients with both ac-quired and hereditary amyloid neuropathy. Classicalpathway activation normally occurs through interaction ofimmune complexes to C1q. The binding of the CH3 do-main of an IgM or multiple CH2 domains of IgG

1

and IgG

3

antibodies with C1 cause C1q conformational changesand auto-activation of C1s and C1r, which sequentiallycleave C4, C2 and C3. Enzymatic cleavage of C3 to C3bexposes a metastable thioester site in the alpha chain ofC3 that covalently binds to target membranes or parti-cles. C3b participates in C5 convertase formation, to ini-tiate C5 cleavage and subsequent terminal complementcomplex generation. In this study we found no detectable

immunoglobulin heavy chain immunoreactivity; therefore,it is unlikely that specific antibodies mediate classicalcomplement cascade activation or the pathogenesis ofacquired amyloid neuropathy as has been previously pro-posed

(Trotter et al., 1977)

. Classical pathway activationof C1 can also occur independent of antibodies throughthe binding of C reactive protein, central nervous systemmyelin

(Vanguri and Shin, 1988)

, serum amyloid P

(Yinget al., 1993)

, and the fibrillar form of

b

-amyloid

(Jiang etal., 1994)

. Despite previous reports that the alternativepathway can be activated by aggregated

b

-amyloid

(Bradtet al., 1998)

and isolated compact peripheral nerve myelin

(Koski et al., 1985)

, Factor Bb was not detected in our bi-opsy material. Together, our data suggest that an antibody-independent activation of the classical pathway of com-plement occurred in association with amyloid in nerve inthese acquired and hereditary cases.

The biologic significance of these early comple-ment components in the pathogenesis of amyloid neu-ropathy is not known. In Alzheimer’s disease, C1q influ-ences the size and complexity of amyloid fibrils throughits ability to accelerate

b

-amyloid aggregation and beta-

Figure 2. Amyloid P (A) and C3d (D) immunoreactivity are detected on amyloid deposits just beneath the perineurium (p) andextending into the endoneurium of this nerve fasicle. The presence of C1q (B) and C4 (C) provided evidence for early classicalcomplement pathway activation (Case 372-90 λ , mag 5003).


136

pleated sheet formation

(Webster et al., 1994)

. Al-though

b

-amyloid peptide self-aggregate

in vitro

underhigh, non-physiologic concentrations, levels

in vivo

areinsufficient for self-aggregation. The presence of C1qunder physiologic conditions is proposed to initiate andpotentiate the formation of aggregated

b

-amyloid fibrilsin the brain and may similarly potentiate the formationof amyloid deposits in peripheral nerve. The diffuse C1qimmunoreactivity throughout the entire amyloid depositwould tend to support the hypothesis.

The presence of terminal activation complex, C5b-9,in a significant number of our cases confirmed that thefull complement cascade has been activated

in vivo.

C5b-9 neoantigen was detected not only on the edgesof amyloid deposits themselves, but also in areasnearby those deposits. C5b-7 shed in the vicinity ofamyloid deposits and not rendered cytolytically inactivethrough interactions with vitronectin would be ex-pected to enter into the lipid bilayer of cell membranessurrounding the deposit, and serve as a nidus for inter-action with C8 and C9 to yield channel forming C5b-9.The efficiency of complement mediated cytolysis in nu-

cleated cells is dependent on the formation of multipleC5b-9 channels

(Koski et al., 1983)

. Deposition of C5b-9is limited on Schwann cells and other targets by the ex-pression of complement regulatory proteins

(Koski et al.,1996; Sawant-Mane et al., 1996)

and by endocytosis orshedding of C5b-9 from the cell surface

(Carney et al.,1986; Morgan et al., 1987)

. Even in the absence of cytol-ysis, sublytic complement channels are capable of induc-ing diverse endoneurial cellular responses that includeconduction block

(Roberts et al., 1995)

, downregulationof myelin gene expression

(Dashiell and Koski, 1999)

,stimulation of Schwann cell proliferation and rescue fromapoptosis

(Dashiell et al., 2000)

, and activation of neutralproteases and phospholipase capable of cleaving struc-tural myelin proteins and disrupting the lipid bilayer

(Vanguri and Shin, 1988)

. Through these mechanisms,C5b-9 insertion into cells in the vicinity of amyloid de-posits may promote more widespread damage throughcell activation and injury.

C5b-9 was also deposited in and around the vesselwalls, as well as on the endothelium of epineurial andendoneurial vessels thickened by amyloid. Insertion of

Figure 3. C1q (B) diffusely stains amyloid deposits in this sural nerve fasicle. In contrast, C3d (C), which covalently binds to targetmembranes, is deposited more discretely in this paraffin section. C5b-9 (D) immunoreactivity is less intense than C3d staining (Case249-79 λ, mag 5003). In figure (A), amyloid deposits are detected in a serial section by amyloid P immunoreactivity (mag 3003).


137

Table 2. Location of amyloid deposits and complement activation products

Case TypeAmyloid C3d C5b-9

Endo- Peri- Epi- Endo- Peri- Epi- Endo- Peri- Epi-

155-77 l D + D, pv D Endoth Endoth 1 pv Endoth253-82 l pv D, pv pv pv372-90 l D D, pv 1 Endoth290-89 l D, pv45-92 l pv D, pv503-92 l.k D, pv D, pv D, pv 1 pv534-92 l.k D, pv D, pv 1 pv200-75 l/k D D D, pv D D 1 pv249-79 l/k D, pv D, pv pv D, pv342-86 l/k D, pv pv D 1216-78 k D D Endoth D68-86 k pv D33-88 k D vessel wall D pv 1 pv Endoth115-90 k vessel wall pv pv vessel wall60-76 TTR small D pv pv 1 Endoth142-80 TTR D, pv 1 Endoth, pv428-80 TTR D vessel wall D vessel wall D, pv 1 pv Endoth296-84 TTR D D 1 Endoth122-87 TTR D, pv D pv127-90 TTR D vessel wall D vessel wall vessel wall

Epi, epineurium; Peri, perineurium; Endo, endoneurium; pv, perivascular; D, deposit; Endoth, endothelium.

Figure 5. Complement activation products are detected on amyloid deposits associated with vessels. C5b-9 is localized within thewall of this epineurial vessel (v) thickened by amyloid (A – Case 115-90κ, Mag 7003). C5b-9 immunoreactivity is present on the en-dothelial surface of epineurial vessels and extends into the vessel wall (B - case 428-80 TTR, Mag 7803). C3d is present on depositssurrounding these epineurial vessels (C - Case 372-90, mag 8003). In this longitudinal section, punctate C3d immunoreactivity ispresent along this endoneurial vessel (v) and on deposits within the endoneurium (arrowhead) (D - Case 534-92 λ/κ mag 1303).


138

terminal complement complexes can alter the cytoskel-etal integrity of endothelial cells, leading to transientformation of interendothelial cell gaps that enhancevascular permeability (Saadi and Platt, 1995). Enhancedpermeability would facilitate extravasation of amy-loidogenic proteins into the endoneurium and hencepromote amyloid formation. Further studies are neces-sary to determine if C5b-9 is fundamental in the earlypathogenesis of amyloid neuropathy.

Increased vascular permeability permits the entry ofserum complement components normally excluded fromthe endoneurium by the blood–nerve barrier. However,this may not be the only source of complement compo-nents. Microglia produce C3 in vitro, a process that is en-hanced by b-amyloid A4 peptides (Haga et al., 1993). Fur-thermore, complement increases b-amyloid neurotoxicityin embryonic rat hippocampi cultures (Schultz et al.,1994). These findings suggest that in the brain, b-amyloidstimulates surrounding microglia and astrocytes to se-crete C3, which is incorporated into amyloid deposits andcontributes to neurotoxicity (Mattson et al., 1992). Simi-larly, Schwann cells in peripheral nerves are capable ofproducing complement components including C3 (Dash-iell et al., 1997) and amyloid light chain fibrils are knownto invade the Schwann cell basal lamina (Sommer andSchroder, 1989). Hence, it is possible that amyloid in pe-ripheral nerves may stimulate the local production ofacute-phase reactants such as C3 and contribute to thepathogenesis of the neuropathy.

A better understanding of the factors that regulateprecursor proteins synthesis, fibrillar condensation, re-moval or dissolution of the deposits from the tissues,and the requirement for complement is essential forthe development of therapies that will limit disease pro-gression. In the current study, we demonstrated thatcomplement activation products are associated withamyloid deposits in acquired and hereditary amyloidneuropathy. In both forms of amyloidosis, complementwas activated in an antibody-independent fashion viathe classical pathway and led to C5b-9 insertion onamyloid deposits and the surrounding tissue. Comple-ment activation in the endoneurium may enhance amy-loid deposition and cause bystander injury of axons andSchwann cells in the vicinity of the deposit in a mannersimilar to b-amyloid in the brain. Clinical trials in Alz-heimer’s disease are currently investigating the poten-tial benefit of anti-inflammatory agents. It is possiblethat similar agents may be beneficial as adjuvants totreatment of amyloid peripheral neuropathy.

AcknowledgementsThis study was supported in part by the National In-

stitutes of Health Grants (K08NS01959 and 1R01NS/

AI36811) and by a University of Maryland IntramuralResearch Grant.

ReferencesBradt B, Kolb WD, Cooper NR (1998). Complement-dependent

pro-inflammatory properties of the Alzheimer’s disease beta-peptide. J Exp Med 188:431–438.

Carney DF, Hammer CH, Shin ML (1986). Elimination of termi-nal complement complexes in the plasma membrane of nu-cleated cells: Influence of extracellular Ca2+ and associationwith cellular Ca2+. J Immunol 137:263–270.

Chambers RA, Medd WF, Spencer H (1955). Primary amyloid-osis with a special reference to involvement of nervous sys-tem. Q J Med 27:207–226.

Dashiell SM, Koski CL (1999). Sublytic terminal complementcomplexes decrease protein zero expression in Schwanncells. J Neurochem 73:2321–2330.

Dashiell SM, Rus H, Koski CL (2000). Terminal complementcomplexes concomittantly stimulate proliferation and rescueof Schwann cells from apoptosis. Glia 30:187–198.

Dashiell SM, Vanguri P, Koski CL (1997). Dibutyryl cyclic AMPand inflammatory cytokines mediate C3 expression inSchwann cells. Glia 20:308–321.

Durie BGM, Persky B, Soehnlen BJ (1982). Amyloid productionin human myeloma stem-cell culture, with morphologic evi-dence of amyloid secretion by associated macrophages.New Engl J Med 307:1689–1692.

Dyck PJ, Lambert EH (1969). Dissociated sensation in amyloid-osis: compound action potential, quantitative histologic andteased-fiber, and electron microscopic studies of sural nervebiopsies. Arch Neurol 20:490–507.

Haga S, Ikeda K, Sato M, Ishii T (1993). Synthetic Alzheimeramyloid beta/A4 peptides enhance production of comple-ment C3 component by cultured microglial cells. Brain Res601:88–94.

Hsu SM, Raine L, Fanger H (1981). Use of avidin-biotin-peroxi-dase complex (ABC) in immunoperoxidase techniques: Acomparison between ABC and unlabeled antibody (PAP) pro-cedures. J Histochem Cytochem 29:577–580.

Jiang H, Burdick D, Glabe CG, Cotman CW, Tenner AJ (1994).Beta-amyloid activates complement by binding to a specificregion of collagen-like domain of the C1q A chain. J Immunol152:5050–5059.

Kernohan JW, Woltman HW (1942). Amyloid neuritis. Arch Neu-rol 47:132–140.

Koski CL, Estep AE, Sawant-Mane S, Shin ML, Highbarger L,Hansch GM (1996). Complement regulatory molecules onhuman myelin and glial cells: Differential expression affectsdeposition of activated complement proteins. J Neurochem66:303–312.

Koski CL, Padmavathy V, Shin ML (1985). Activation of the alter-native pathway of complement by human peripheral nervemyelin. J Immunol 134:1810–1814.

Koski CL, Ramm EL, Hammer CH, Mayer MM, Shin ML (1983).Cytolysis of nucleated cells by complement: Cell death displaysmulti-hit characteristics. Proc Natl Acad Sci USA 80:3816–3820.

Li K, Kyle RA, Dyck PJ (1992). Immunohistochemical character-ization of amyloid proteins in sural nerves and clinical associ-ations in amyloid neuropathy. Am J Pathol 141:217–226.

Mattson MP, Cheng B, Davis D, Bryant K, Lieberburg I, RydelRE (1992). Beta-amyloid peptides destabilize calcium ho-


139

meostasis and render human cortical neurons vulnerable toexcitotoxicity. J Neurosci 12:376–388.

McGeer PL, McGeer EG (1992). Complement proteins andcomplement inhibitors in Alzheimer’s disease. Res Immunol143:516–527.

Morgan BP, Dankert JR, Esser AF (1987). Recovery of humanneutrophils from complement attack: Removal of membraneattack complex by endocytosis and exocytosis. J Immunol138:246–253.

Roberts M, Willison HJ, Vincent A, Newsom-Davis J (1995).Multifocal motor neuropathy human sera blocks distal motornerve conduction in mice. Ann Neurol 38:111–118.

Rogers J, Cooper NR, Webster S, Schultz J, McGeer PL, StyrenSD, Civen WH, Barchova L, Bradt B, Ward P, Lieberburg I(1992). Complement activation by beta-amyloid in Alzheimer’sdisease. Proc Natl Acad Sci USA 89:10016–10020.

Saadi S, Platt J (1995). Transient perturbations of endothelial in-tegrity induced by natural antibodies and complement. J ExpMed 181:21–31.

Sawant-Mane S, Piddelsden SJ, Morgan BP, Holers VM, Koski CL(1996). CD59 homologue regulates complement dependentcytolysis of rat Schwann cells. J Neuroimmunol 69:63–71.

Schultz J, Schaller J, McKinley M, Bradt B, Cooper N, May P,Rogers J (1994). Enhanced cytotoxicity of beta-amyloid pep-tide by a complement dependent mechanism. Neurosci Lett175:99–102.

Sommer C, Schroder JM (1989). Amyloid neuropathy: Immuno-cytochemical localization of intra- and extracellular immuno-globulin light chains. Acta Neuropathol 79:190–199.

Sullivan JF, Twitchwell TE, Gherardi GJ (1955). Amyloid poly-neuropathy. Neurology 5:847–854.

Toyooka K, Fujimura H, Ueno S, Yoshikawa H, Kaido M, Nish-imura T, Yorifuji S, Yanagihara T (1995). Familial amyloid poly-neuropathy associated with transthyretin Gly42 mutation: Aquantitative light and electron microscopic study of the pe-ripheral nervous system. Acta Neuropathol 90:516–525.

Trotter JL, Engel WK, Ignacjak TF (1977). Amyloidosis withplasma cell dyscrasia. Arch Neurol 34:209–214.

Vanguri P, Shin ML (1988). Hydrolysis of myelin basic protein inhuman myelin by terminal complement complexes. J BiolChem 263:7228–7234.

Webster S, O’Barr S, Rogers J (1994). Enhanced aggregationand beta structure of amyloid beta peptide after coincubationwith C1q. J Neurosci Res 39:448–456.

Ying SC, Gewurz AT, Jiang H, Gerwurz H (1993). Human serumamyloid P component oligomers bind and activate the classi-cal complement pathway via residues 14-26 and 76-92 of theA chain collagen like region of C1q. J Immunol 150:169–176.

Zanusso GL, Moretto G, Bonetti B, Monaco S, Rizzuto N(1992). Complement neoantigen and vitronectin are compo-nents of plaques in amyloid polyneuropathy. Ital J NeurolSci 13:493–499.

complement activation in acquired and hereditary amyloid neuropathy

Documents