Kidney International, Vol. 55 (1999), pp. 899–906

Impaired lysosomal processing of b2-microglobulin byinfiltrating macrophages in dialysis amyloidosis

MAR GARCIA-GARCIA, ANGEL ARGILES, ANNIE GOUIN-CHARNET, MERCE DURFORT,JOSE GARCIA-VALERO, and GEORGES MOURAD

Institut de Genetique Humaine, Montpellier, France; Unit of Cell Biology, Department of Biochemistry and Physiology,University of Barcelona, Barcelona, Spain; and Department of Nephrology, University Hospital Lapeyronie, Montpellier, France

Impaired lysosomal processing of b2-microglobulin by infil- are positive for Congo red staining, showing green bire-trating macrophages in dialysis amyloidosis. fringence under polarized light and fibrillar appearance

Background. Macrophages may participate in amyloid fibril by electron microscope analysis [1]. Although amyloidformation by processing the protein precursor. Although thisdeposition has been classically considered extracellulartheory seems to apply for amyloidosis, in which proteolytic

cleavage is a prerequisite for amyloid fibril formation, it has [1], a few studies showing intracellular amyloid fibrilsnot been demonstrated for b2-microglobulin (b2m) amyloidosis. have been reported [2–6]. In some of these amyloidoses,We aimed to establish the role played by macrophages in b2m it has been suggested that the cells participate in amyloidamyloidosis.

fibril formation by processing the precursor protein [1].Methods. We used a double immunogold electron micros-Recently, the unprocessed protein precursor, serumcopy technique, including mouse antihuman CD68, rabbit anti-

human b2m, amyloid P component, and lysosome-associated amyloid A (AA), and the proteolyzed protein, AA, havemembrane protein (LAMP-1) antibodies. Differential density been found within the lysosomes in AA amyloidosis [7].labeling studies of b2m and amyloid P component were per-

Because AA fibrils contain proteolyzed amyloid protein,formed extra- and intracellularly to assess protein processingit has been suggested that intralysosomal proteolysisby macrophages.

Results. The cells surrounding amyloid fibrils were found to would be a prerequisite for amyloid fibril formation [7].be mostly CD68 positive, suggesting that they were of mono- However, the possibility exists that intralysosomal amy-cyte–macrophage lineage. Intracellular accumulation of amy-

loid fibrils were the product of phagocytosis. Althoughloid fibrils was also observed; these fibrils were constantly sur-most of the reports dealing with the monocytic involve-rounded by LAMP-1–linked gold particles, demonstrating that

intracellular b2m was almost exclusively lysosomal. The rough- ment in the pathogenesis of amyloidosis propose an ac-surface endoplasmic reticulum was not labeled by b2m anti- tive participation in the formation and subsequent shed-body, suggesting that there was no active synthesis of b2m by

ding of amyloid fibrils by these cells, there are nothe cells. As a marker of endocytosis, protruded cytoplasmicconclusive studies allowing the rejection of the hypothe-processes in close relation with the intracellular accumulations

of b2m amyloid fibrils were observed. No difference in density sis of a simple phagocytosis of already formed amyloidlabeling (extracellular vs. intracellular) was observed for b2m, fibrils. If the latter were true, the macrophage participa-whereas intracellular P component labeling was significantly

tion would be part of the reactive phenomena commonlydecreased.seen in other inflammatory processes rather than theConclusions. All of these data are strongly suggestive of

phagocytosis and not synthesis of amyloid fibrils by macro- cause of amyloidosis.phages. Further, they demonstrate an impaired lysosomal pro- Dialysis-related amyloidosis (DRA), described in thecessing specific for b2m, as other compounds of the amyloid

1980s, is a type of amyloidosis with a clear predilectionfibrils (P component) are significantly cleared.for osteoarticular structures, which is mainly observedin long-term dialysis patients [8, 9]. Carpal tunnel syn-drome, scapulohumeral periarthritis, and pain are theAmyloidosis is a disease characterized by the tissuemain clinical features [9, 10]. Biochemically, the maindeposition of insoluble protein fibrils. Amyloid depositsprotein component of DRA fibrils has been identifiedas b2-microglobulin (b2m) [11]. The mechanisms byKey words: amyloid fibrils, phagocytosis, dialysis-related amyloidosis,

toxicity. which b2m precipitates into amyloid fibrils remain un-clear.Received for publication May 5, 1998

Although the protein constituents of DRA have beenand in revised form September 10, 1998Accepted for publication September 25, 1998 well assessed, the cellular participation on the genesis

of the amyloid fibrils has been less investigated. Only 1999 by the International Society of Nephrology

899

Garcıa-Garcıa et al: Lysosomal processing and b2m900

a few reports have characterized the cells surrounding ded in Lowicryl K4M (Chemische Werke Lowi, Wald-Kraiburg, Germany) as previously described [14]. Ul-amyloid deposits in DRA [3, 5, 12], and it is not known

if the presence of these cells is at the origin of the amyloid trathin sections (60 to 90 nm) were obtained using anUltracut E system (Reichert-Jung, Wien, Austria) andfibrils or if they accumulate in amyloid deposits because

of the existence of amyloid fibrils. Our aim in this study were mounted on formvar-coated and etched gold grids.Before labeling, sections were rinsed twice with 0.1 mwas to clarify the cellular participation on the pathogene-

sis of b2m amyloidosis. We performed electron micros- glycine-PBS for five minutes and were incubated with2% ovalbumin in PBS for 30 minutes to block unspecificcopy studies with simple and/or double immunogold

labeling, as well as with differential density labeling quan- antibody–antigen complexes. The grids were then incu-bated with polyclonal anti-b2m (dilution 1:500; Nordic,titation. These techniques allowed us to assess the struc-

tural relations between amyloid deposits and the sur- Tilburg, The Netherlands) or anti-amyloid P component(Dako, Glostrup, Denmark) antibodies in ovalbumin-rounding cells. Furthermore, by identifying the cytosolic

endomembranes and locating the intracellular b2m, we PBS for two hours. After three 15-minute rinses in glycine-PBS, bound antibodies were visualized with 10 nm pro-were able to hypothesize on the participation of the

cellular organelles (those related with synthesis as well tein A gold (provided by Dr. J.W. Slot, University ofUtrecht, Utrecht, The Netherlands). The sections wereas those related with the degradative pathways) on the

pathogenesis of amyloidosis. Finally, the intracellular finally rinsed with PBS and were double distilled in waterprior to counterstaining with aqueous uranyl acetate andand extracellular behavior of b2m was compared with

those of other constituents of amyloid deposits, such as lead citrate.Double immunolabeling. For double immunolabe-the amyloid P component, to know whether these pro-

teins are handled differently by the incriminated cells. ling, the samples were fixed as for single labeling, cryo-protected by infusing 2.1 m sucrose in PBS, snap freezing,and storing in liquid nitrogen until analysis. Ultrathin

METHODScryosections were obtained using an FC-40 cryosystem

Patients and samples (Reichert-Jung) and transferred to formvar-coated andetched gold grids in a drop of 2.3 m sucrose in PBS.Amyloid deposits were obtained from carpal tunnel

of three men and one woman, who were 65 6 4 years Before labeling, sections were rinsed twice with 0.1 mglycine-PBS for 10 minutes and were incubated with 2%old, and had dialysis-related amyloidosis. Their renal

diseases were chronic glomerulonephritis (three pa- ovalbumin in PBS for 20 minutes. For immunolabeling,sections were first incubated with anti-b2-microglobulintients) and autosomal-dominant polycystic renal disease

(the remaining one). None of the patients had systemic antibodies for 30 minutes and were labeled with 15 nmprotein A gold after three glycine-PBS rinses. Then thediseases known to be associated with amyloidosis.sections were stabilized for five minutes with 0.5% glu-

Sample preparation and immunohistochemistry taraldehyde in PBS, blocked with 150 mm NH4Cl in PBS,and incubated with a rabbit anti-human lysosome-associ-Congo red staining. Unfixed amyloid deposits surgi-

cally obtained were immersed in O.C.T. compound ated membrane protein (LAMP-1) antibody (dilution1:120; provided by Dr. S.R. Carlsson, University of Umea,(Miles Inc., Diagnostics Division, Elkhart, IN, USA),

frozen, and stored at 2708C. Cryostat sections of 10 mm Umea, Sweden) or monoclonal anti-CD68 antibody (di-lution 1:100; Dako) followed by unlabeled rabbit anti-were transferred onto gelatin-coated (0.3%) glass slides

and were stained with Congo red (Searle Scientific Ser- mouse Ig (dilution 1:50; Dako). The ultrathin sectionswere rinsed with PBS and double-distilled water, werevices, High Wycombe Bucks, UK) following a slight

modification of Puchtler’s method [13]. Briefly, sections then counterstained with uranyl acetate, and were em-bedded in methyl cellulose [15].were dehydrated in 70% ethanol for five minutes and

were then treated with an alkaline-saturated salt solu- Controls were performed with the omission of the pri-mary and/or the secondary antibodies, and/or the omissiontion. The stained sections were mounted in a medium

compatible with organic solvents and were observed un- of the rabbit antimouse Ig. Examination was performedin a Hitachi H-600 AB and a Philips EM-300 transmis-der polarized light in a Zeiss microscope (Oberkochen,

Germany). sion electron microscopes.Immunogold methods. Blocks of approximately 1 mm3

Morphology studiesof amyloid deposits were fixed in 4% (vol/vol) paraform-aldehyde and 0.1% (vol/vol) glutaraldehyde in 0.1 m Semithin sections of 1 mm were obtained from small

blocks of amyloid deposits embedded in Lowicryl K4Mphosphate-buffered saline (PBS), pH 7.4, for two hoursat 48C. resin. Sections were transferred onto glass slides and were

stained with 0.5% methylene blue 0.5% Borax for 30Single immunolabeling. Samples were progressivelydehydrated in graded ethanols and were finally embed- seconds at 908C. The stained sections were then mounted

Garcıa-Garcıa et al: Lysosomal processing and b2m 901

in DPX medium and were observed in a Polyvar 2 optical pathways of cells (endoplasmic reticulum or the Golgicompartment; Fig. 3b).microscope (Reichert-Jung) under immersion oil.

For ultrastructural studies, small pieces of amyloid Two series of studies of double immunolabeling wereperformed using the LAMP-1 or KP1 in addition to b2mdeposits were fixed as for Lowicryl embedding and were

postfixed in 2.5% glutaraldehyde (vol/vol) in PBS. Then and amyloid P component antibodies, to characterize thecytoplasmic membranes containing amyloid fibrils. Whenthe samples were progressively treated in graded ace-

tone-resin solutions, postfixed with 0.1% osmium tetrox- anti-b2m antibody was incubated along with a polyclonalantihuman LAMP-1 antibody (lysosome marker), bothide, and embedded in Spurr (Agar Scientific LTD,

Stansted, UK). Ultrathin sections were obtained as for antibodies colocalized on the same vesicular structures(Fig. 3c). Some vesicles labeled with anti-b2m antibodyLowicryl embedded samples using an Ultracut E system

and were then counterstained with uranyl acetate and were only weakly labeled with LAMP-1 antibody. Accord-ing to their location (marginal, close to the plasma mem-lead citrate before examination in the transmission elec-

tron microscope. brane) and to their immunolabeling pattern (near con-stant LAMP-1 label), these vesicles had an endocytic

Quantitative evaluation of labeling density character (that is, early endosomes or endocytic vesicles;Fig. 3c).Label density was estimated as the number of gold

particles per area of amyloid deposit by stereological When anti-b2m antibody was incubated along withKP1 antibody, their labeling again colocalized withinmethods [16]. Differences were checked using the Stu-

dent’s t-test. A P of less than 0.05 was considered signifi- the same vesicles. Because KP1 antibody is a lysosomalmarker specific for monocytes/macrophages that is di-cant.rected against the CD68 antigen, these experiments con-firmed both the lysosomal character of the vesicles con-

RESULTStaining amyloid fibrils and the monocyte–macrophage

Amyloid diagnosis lineage of the cells associated with the amyloid deposits.The surgically obtained deposits were of amyloid ori-

Density labeling of b2m and amyloid P componentgin, as they were positively stained with Congo red anddisplayed apple-green birefringence under polarized The presence of numerous lysosomes filled with b2mlight. An example of Congo red staining without and with fibrils suggests a failure in degrading the engulfed mate-polarized light is given in Figure 1 A and B, respectively. rial by the macrophages. To verify this hypothesis, differ-

ential density labeling studies between intracellular andDeposits and related cells extracellular material were performed for b2m. Further-

Amyloid deposits were observed as nodules, consisting more, the same studies were performed assessing theof intercellular amorphous material and long-shaped extracellular and intracellular density labeling of amy-cells as well as macrophages. The long-shaped cells con- loid P component. Amyloid P component is a commontained in the nodules presented an enlarged perinuclear constituent of most types of amyloidoses, irrespective ofregion with long and narrow cytoplasmic processes ex- the characteristics of their major protein component.tending to the amyloid material (Figure 2). Electron Extracellular density labeling for b2m was not signifi-microscopy showed that the extracellular material had cantly different from the intracellular one (Fig. 4 a, b).a heterogeneous composition consisting of amyloid On the contrary, extracellular density labeling for amy-clumps and some scattered collagen fibers. The macro- loid P component was significantly stronger than thephage-like cells had an irregular surface with numerous intracellular one (Fig. 4 c, d), suggesting that the engulfedinvaginations that were usually occupied by the amyloid amyloid P component is processed by the cells, whereasclumps (Fig. 2 a, b). The cytoplasm had a granular ap- there is no cellular processing of b2m. The analysis ofpearance and showed a well-developed Golgi compart- the density labeling is summarized in Table 1.ment, endoplasmic reticulum, and vesicle compartments.The vesicles seemed generated by endocytosis of the

DISCUSSIONfibrillar clumps (Fig. 2c). The extracellular amyloid clumpsOur study shows, by immunogold labeling, that mostwere heavily labeled with anti-b2m antibody (Fig. 2b).

of the cells around and between b2m amyloid depositsAmyloid fibrils immunoreactive for anti-b2m antibodywere monocyte–macrophages, presenting cytoplasmicwere also observed within the cytoplasm of macrophage-extension processes surrounding extracellular b2m fibrils.like cells. These intracellular fibrils were usually con-Some cells, also positive for macrophage markers, con-tained in vesicles, and cytoplasmic amyloid clumps weretained intracellular b2m amyloid fibrils. Double immuno-seldom devoid of membrane (Figs. 2c and 3). Intracellu-gold labeling allowed us for the first time to localizelar labeling of b2m was clearly not associated with the

compartments related to the synthesis and secretory precisely intracellular b2m fibrils within the lysosomes:

Garcıa-Garcıa et al: Lysosomal processing and b2m902

Fig. 1. Amyloid criteria: Congo red stainingof a b2-microglobulin amyloid deposit ob-served under light microscopy (A) and underpolarized light (B). The publication of thisfigure in color was made possible by a grantfrom Gambro Renal Care R&D, Lund,Sweden.

LAMP-1 antibody was distributed on and underneath surrounding the amyloid deposits [12]. However, therewere no previous immunohistochemical studies at thethe membranes of the vesicles containing b2m amyloid

fibrils. These results were fully confirmed by KP1 anti- ultrastructural level confirming the identity of the cells.Because in most of the different amyloidoses the amy-body immunostaining. Finally, and more importantly,

the analysis of differential density labeling of b2m and loid proteins result from partial proteolytic digestion oftheir precursors, proteolysis has been considered for aamyloid P component showed that monocyte–macro-

phage cells are unable to degrade b2m once it is in the long time as a prerequisite for amyloid fibril formation[1]. Because monocytes and macrophages are enrichedlysosomes and that this impaired lysosomal function is

specific for b2m. in lysosomal structures that may proteolyze the differentamyloid precursors, these cells have been thought to beThe studies aiming to characterize the cells participat-

ing in amyloid formation and/or persistence resulted in responsible for the amyloid fibril formation [1, 2, 4, 19–21]. However, the opposite may also hold true, and mac-the identification of a variety of cells with a common

feature: their monocyte–macrophage lineage [2, 5, 12, 17, rophages could participate in the reaction processes ofamyloidosis by trying to remove the amyloid fibrils as18]. We have previously established, by specific antigen

typing, the frequency of the different leukocyte cell types they clear other exogenous or endogenous material [22,

Garcıa-Garcıa et al: Lysosomal processing and b2m 903

Fig. 2. Amyloid deposit location and b2-microglobulin distribution. (Inset) Methyleneblue staining showing the cell distributionaround the amyloid fibrils. (a) Electron mi-croscopy showing a cell engulfing an amyloiddeposit (short thick arrows). Cellular pro-cesses are indicated by thin arrows (318,000).Immunogold labeling with anti-b2-microglob-ulin antibody showing an extracellular (b) andintracellular (c) location of an amyloid deposit(327,000).

23]. If this were the case, extracellular amyloid deposi- [28]. Finally, atrial natriuretic factor is frequently foundin intact form in amyloid fibrils of the senile type oftion would take place when the rate of amyloid produc-

tion overcomes the clearing capacity of macrophages. amyloidosis [29].Our findings are very much supportive of a phagocyto-Interestingly, there are a few exceptions to the proteol-

ysis rule in amyloidosis. The first known exception has sis role for monocyte macrophages and render less likelythe participation of these cells in the synthesis of amyloidbeen b2m amyloidosis. Although a few reports suggest

that the amyloidogenic b2m may be proteolyzed in its fibrils. (a) We found the extracellular amyloid fibrilsfrequently surrounded by cellular extensions at differentN-terminal side [24], several groups have shown that the

amyloid deposits in b2m amyloidosis are mainly consti- stages of endocytosis. (b) We observed intracellular b2munevenly distributed and located the amyloid fibrilstuted of intact b2m [25, 26]. Other examples of unproteo-

lyzed precursor proteins have been reported subse- within the lysosomes, and (c) we observed no label forb2m in the synthesis pathways of the cells.quently by Buxbaun for the AL type of amyloidosis [27]

and by Ericsson et al for the AA type of amyloidosis Because the amyloidogenic b2m is intact in its N-termi-

Garcıa-Garcıa et al: Lysosomal processing and b2m904

Fig. 3. Intracellular immunolocalization ofb2-microglobulin. No labeling with anti-b2-microglobulin antibody was observed in en-doplasmic reticulum (er) (a, 333,000) nor inthe Golgi compartment (g) (b, 333,000). (c)Double immunolabeling for 32-microglobu-lin (15 nm) and lysosome-associated protein(LAMP-1; 10 nm) showing that amyloid fibrilsare within the lysosomes (340,000). (d) Dou-ble immunolabeling for b2-microglobulin (15nm) and CD68 (10 nm) confirming that amy-loid fibrils are within the lysosomes and thatthe cells involved are of macrophage-mono-cytic lineage (340,000).

nal side, it seems unlikely that macrophages with lyso- ger RNA of amyloid protein precursor (APP) has beenfound by in situ hybridization in plaque-associated micro-somes filled with b2m participate in synthesis of amyloid

fibrils. glia, suggesting no synthesis of the amyloid precursor [34].Once the ultrastructural studies enabled us to con-A second line of evidence supporting the phagocytic

role for the cells surrounding amyloid fibrils comes from clude that amyloid fibrils are contained in lysosomes, weaimed to step further and assessed the degradation ofAlzheimer’s disease. Cataldo et al reported the up-regu-

lation of cathepsin D synthesis and the accumulation of the different proteins of these amyloid fibrils. We com-pared the density labeling of the extracellular amyloidhydrolase-laden lysosomes in pyramidal neurons, indi-

cating that the endosomal-lysosomal system is abnor- fibrils in regard to the intracellular ones. It came outclearly from our studies that macrophages were able tomally activated in Alzheimer’s disease [30]. On the other

hand, the addition of inhibitors of lysosomal proteinases degrade the amyloid P component through phagocytosisand lysosomal processing. On the contrary, b2m, afterdoes not block the formation of amyloid fibrils in the in

vitro setting [31–33]. In other words, the endosomal- being internalized by the same cells, was not processedfurther. Therefore, the lysosomes of macrophages werelysosomal system is activated, but its blockade does not

result in preventing the formation of amyloid fibrils. unable to pursue the processing of b2m. Whether thelack of processing of b2m is due to the conformationalThus, such activation would be a consequence rather

than the cause of amyloid fibril formation. In keeping situation of the molecule within amyloid fibrils or to anintrinsic characteristic of b2m is not known. Very re-with the phagocytic hypothesis, no detection of messen-

Garcıa-Garcıa et al: Lysosomal processing and b2m 905

Fig. 4. Density labeling studies of b2-micro-globulin and amyloid P component (APC):Comparison of intracellular and extracellularprotein density. It can be observed that extra-cellular labeling of b2-microglobulin (a,327,000) was similar to that observed intracel-lularly (b, 327,000), whereas a clear decreasein intracellular APC labeling (d, 333,000) wasobserved when comparing with extracellularAPC labeling (c 337,800).

Table 1. Intracellular and extracellular density labeling with anti- degradation. Whether the accumulated b2m behaves likeb2-microglobulin and anti-amyloid P substance antibodies

a cell toxin remains to be demonstrated.Extracellular Intracellular In summary, although the biochemical analysis of the

different types of amyloidosis is quite well established,b2-microglobulin 299 6 53 266 6 53 (NS)Amyloid P substance 12 6 1 5 6 1a the cellular participation has been less assessed. There

Values represent the mean 6 standard error fo the mean of the patients. The is some controversy as to whether the cells surroundingvalue for each patient was obtained by averaging 5 random readings of 3 different the amyloid deposits synthesize the amyloid fibrils ortissue preparations.

P , 0.001 (Student’s T test intracellular versus extracellular) are there to degrade them. Previous studies on AL andAA amyloidoses suggested a synthesis role. Our analysisof b2m amyloid deposits showed that the phagocytic cellsfrequently observed in amyloid deposits have cytoplas-

cently, Inoue et al have nicely shown the arrangement mic extension processes in regards to the b2m amyloidof b2m, amyloid P component, and chondroitin sulfate fibril-enriched zones. Further, it showed that the intracel-proteoglycan within amyloid fibrils [35]. According to lular b2m is preferentially and almost exclusively intraly-these elegant ultrastructural studies, b2m is associated sosomal, and there were no signs of synthesis of b2m orwith the surface of the amyloid fibrils [35] and would be amyloid fibrils. We have previously reported the identi-accessible to lysosomal proteases. Therefore, the struc- fication of the major serum antiprotease a2-macroglobu-ture of b2m, comprising of two b-pleated sheets of three lin in b2m amyloid deposits [36] and proposed a role for

antiproteases in the pathogenesis of amyloidosis [37].and four strands, respectively, may render it resistant to

Garcıa-Garcıa et al: Lysosomal processing and b2m906

J: Cells surrounding haemodialysis-associated amyloid deposits areThese data represent an extension from the protein tomainly macrophages. Nephrol Dial Transplant 9:662–667, 1994

the cellular level on the assessment of the catabolism of 13. Putchler H, Sweat F, Levine M: Binding of Congo Red by amy-loid. J Histochem Cytochem 10:355–364, 1962amyloid fibrils. We hypothesize that failure of lysosomes

14. Carlemaln E, Garavito RM, Villiger W: Resin developmentto degrade intact b2m may be one of the key pointsfor electron microscopy and an analysis of embedding at low tem-

in the pathogenesis of b2m amyloidosis. Our data are perature. J Microsc 126:123–143, 198215. Tokuyasu KT: Immunocytochemistry on ultrathin sections. Histo-supportive of a phagocytic role for macrophages and

chem J 12:381–403, 1980provide some evidence that lysosomal function is not16. Weibel ER: Stereological methods, in Practical Methods for Bio-

effective in processing and clearing b2m, suggesting a logical Morphometry, New York, Academic Press, 197917. Kisilevsky R, Lyon AW, Young I: A critical analysis of postulatedputative b2m toxicity on these cells.

pathogenetic mechanisms in amyloidogenesis. Crit Rev Clin LabSci 29:59–82, 1992

ACKNOWLEDGMENTS 18. Ohashi K, Hara M, Kawai R, Ogura Y, Honda K, Nihei H,Mimura N: Cervical discs are most susceptible to b2-microglobulin

We thank Baxter Healthcare Corporation for contributing to the amyloid deposition in the vertebral column. Kidney Int 41:1646–funding of this work through its Extramural Grant Program. The Cen- 1652, 1992tre Hospitalier Universitaire of Montpellier also funded this study. 19. Ramadori G, Gieder H, Sipe J, Shirahama T, Meyer ZumM.G. was a recipient of a grant from the CIRIT—Direccio General Buschenfelde KH: Murine tissue macrophages synthesize andde Recerca, Generalitat de Catalunya. We thank Almudena Garcıa secrete amyloid proteins different to amyloid (AA). Eur J Clinand the staff of Serveis Cientıfico-Tecnics (Universitat de Barcelona) Invest 19:316–322, 1989for technical assistance. The publication of Figure 1 in color was made 20. Shirahama T, Cohen AS: An analysis of the close relationship ofpossible by a grant from Gambro Renal Care R&D, Lund, Sweden. lysosomes to early deposits of amyloid. Am J Pathol 73:97–114, 1973A part of the results included in this study was orally presented at the 21. Shirahama T, Cohen AS: Intralysosomal formation of amyloidXXXIIIrd ERA-EDTA meeting in Amsterdam. fibrils. Am J Pathol 81:101–116, 1975

22. Johnston RB: Monocytes and macrophages. N Engl J MedReprint requests to Dr. Angel Argiles, Institut de Genetique Hu- 318:747–775, 1988

maine—CNRS UPR 1142, 141, rue de la Cardonille, 34090 Montpellier 23. Zucker-Franklin D: Immunophagocytosis of human amyloid fi-Cedex 5, France. brils by leukocytes. J Ultrastruc Res 32:247–257, 1970E-mail: Angel.Argiles@igh.cnrs.fr 24. Linke RP, Hampl H, Lobeck H, Ritz E, Bommer J, Waldherr

R, Eulitz M: Lysine-specific cleavage of B2m in amyloid depositsassociated with hemodialysis. Kidney Int 36:675–681, 1989

REFERENCES 25. Argiles A, Garcia-Garcia M, Derancourt J, Mourad G, De-maille JG: b2-microglobulin isoforms in healthy individuals and1. Glenner GG: Amyloid deposits and amyloidosis. N Engl J Medin amyloid deposits. Kidney Int 48:1397–1405, 1995302:1283–1292, 1333–1343, 1980

26. Garcia-Garcia M, Garcia-Valero J, Mourad G, Argiles A:2. Fuks A, Zucher-Franklin D: Impaired Kupffer cell function pre-Urinary and serum b2-microglobulin in living related kidney do-cedes development of secondary amyloidosis. J Exp Med 161:1013–nors and in renal failure. Contrib Nephrol 112:77–82, 19951028, 1985 27. Buxbaun J: Aberrant immunoglobulin synthesis in light chain amy-3. Depierreux M, Goldman M, Fayt I, Richard C, Quintin J, loidosis. J Clin Invest 78:798–806, 1986

Dhaene M, Van Herweghem JL: Osteoarticular amyloidosis asso- 28. Ericsson LH, Eriksen N, Walsh KA, Benditt EP: Primary struc-ciated with haemodialysis: An immunoultrastructural study. J Clin ture of duck amyloid protein A: The form deposited in tissuesPathol 41:158–162, 1988 may be identical to its serum precursor. FEBS Lett 218:16–21, 1987

4. Takahashi M, Yokota T, Kawano H, Gondo T, Ishihara T, 29. Johansson B, Wernstedt C, Westermark P: Atrial natriureticYchino F: Ultrastructural evidence for intracellular formation of peptide deposited as atrial amyloid fibrils. Biochem Biophys Resamyloid fibrils in macrophages. Virchows Arch 415:411–419, 1989 Commun 148:1087–1092, 1987

5. Nishi S, Ogino S, Marayama Y, Horama N, Gejyo F, Morita 30. Cataldo AM, Barnett JL, Berman SA, Li J, Quarless S, Bursz-T, Arakawa M: Electron-microscopic and immunohistochemical tajn S, Lippa C, Nixon RA: Gene expression and cellular contentstudy of b2m-related amyloidosis. Nephron 56:357–363, 1990 of cathepsin D in Alzheimer’s disease brain evidence for early up-

6. de Koning E, Morris EF, Posthuma G, Hoppener J, Morris J, regulation of the endosomal-lysosomal system. Neuron 14:671–680,Capel P, Clark A, Verbeek J: Intra and extracellular amyloid 1995fibrils are formed in cultured pancreatic islets of transgenic mice 31. Haass C, Hung AY, Schlossmacher MG, Teplow DB, Selkoeexpressing human islet amyloid polypeptide. Proc Natl Acad Sci DJ: b-Amyloid peptide and a 3-kDa fragment are derived by dis-USA 91:8467–8471, 1994 tinct cellular mechanisms. J Biol Chem 268:3021–3024, 1993

7. Chronopoulos S, Laird DW, Ali-Khan Z: Immunolocalization 32. Busciglio J, Gabuzda DH, Yankner BA: Generation of b-amy-of SAA and AA amyloid in lysosomes in murine monocytoid cells: loid in the secretory pathway in neuronal and nonneuronal cells.Confocal and immunogold EM studies. J Pathol 173:361–369, 1994 Proc Natl Acad Sci USA 90:2092–2096, 1993

8. Assenat H, Calemard D, Charra B, Laurent G, Terrat JC: 33. Higaki J, Quon D, Zhong Z, Cordell B: Inhibition of b-amyloidHemodialyse, syndrome du canal carpien et substance amyloide. formation identifies proteolytic precursors and subcellular site of(letter) Nouv Press Med 24:1715, 1980 catabolism. Neuron 14:651–659, 1995

9. Charra B, Calemard F, Uzan M, Terrat JC, Vanel T, Laurent 34. Scott S, Johnson S, Zarow C, Perlmutter L: Inability to detectG: Carpal tunnel syndrome, shoulder pain and amyloid deposits b-amyloid protein precursor mRNA in Alzheimer plaque-associ-in long-term hemodialysis patients. Proc Eur Dial Transplant Assoc ated microglia. Exp Neurol 121:113–118, 199321:291–295, 1984 35. Inoue S, Kuroiwa M, Ohashi K, Hara M, Kisilevsky R: Ultra-

10. Bardin T, Kuntz D: The arthropathy of chronic hemodialysis. structural organisation of hemodialysis-associated b2-microglobu-Clin Exp Rheumatol 5:379–386, 1987 lin amyloid fibrils. Kidney Int 52:1543–1549, 1997

11. Gejyo F, Yamada T, Odani S, Nakagawa Y, Arakawa M, Kuni- 36. Argiles A, Mourad G, Axelrud-Cavadore C, Watrin A, Mionmoto T, Kataoka H, Suzuki M, Hirasawa Y, Shirahama T, Cohen C, Cavadore JC: High molecular weight proteins in haemodialysisAS, Schimid K: A new form of amyloid protein associated with associated amyloidosis. Clin Sci 76:547–552, 1989chronic hemodialysis was identified as b2m. Biochem Biophys Res 37. Argiles A, Mourad G, Atkins RC, Mion C: New insights in theCommun 129:701–706, 1985 pathogenesis of hemodialysis associated amyloidosis. (editorial)

12. Argiles A, Mourad G, Kerr P, Garcia M, Collins B, Demaille Semin Dial 3:149–152, 1990