Transforming growth factor-β is involved in the pathogenesis of dialysis-related amyloidosis

Kidney International, Vol. 57 (2000), pp. 697–708

Transforming growth factor-b is involved in the pathogenesisof dialysis-related amyloidosis

KENZO MATSUO, T. ALP IKIZLER, RICHARD L. HOOVER, MASAHIKO NAKAMOTO,CHIKAO YASUNAGA, LARA B. PUPIM, and RAYMOND M. HAKIM

Department of Medicine, Division of Nephrology, Department of Pathology, Vanderbilt University Medical Center, Nashville,Tennessee, USA, and Kidney Center, Saiseikai Yahata Hospital, Kitakyushu, Japan

In turn, TGF-b suppresses the proinflammatory activation ofTransforming growth factor-b is involved in the pathogenesismacrophages, suggesting a dual role for TGF-b in the inflam-of dialysis-related amyloidosis.matory process of DRA. These observations may provide aBackground. Advanced glycation end product-modifiedpathophysiologic link between TGF-b and DRA.b2-microglobulin (AGE-b2m) is an important component of

dialysis-related amyloidosis (DRA). Its presence induces mono-cyte chemotaxis and the release of the proinflammatory cyto-kines through macrophage activation. Transforming growth Dialysis-related amyloidosis (DRA) is a disablingfactor-b (TGF-b) is a multifunctional cytokine that also has

complication in long-term chronic hemodialysis (CHD)chemotactic activity for monocytes at very low (0.1 to 10 pg/mL)patients [1]. The osteoarticular manifestations of DRAconcentrations and inhibits proinflammatory cytokine produc-

tion of macrophages. In this study, we investigated the role of include carpal tunnel syndrome (CTS), bone cysts, andTGF-b in the pathogenesis of DRA. destructive spondylarthlopathy [2]. The major constit-

Methods. We performed an immunohistochemical study of uent of amyloid fibril has been identified to be b2-micro-DRA tissues (8 cases) to confirm the existence of TGF-bs andglobulin (b2m) [3]. Recent studies have demonstratedtheir receptors; we also performed a chemotaxis assay of humanthat advanced glycation end products (AGE) formedmonocytes as well as enzyme-linked immunosorbent assay

(ELISA) of TGF-b1, tumor necrosis factor-a (TNF-a), inter- by nonenzymatic Maillard reaction between protein andleukin-1b (IL-1b), and interleukin-1 receptor antagonist sugar are present in b2m-amyloid fibrils of DRA [4, 5].(IL-1Ra) in the supernatant of human monocyte-derived mac- AGE-modified b2m (AGE-b2m) has also been detectedrophage cell culture under varying conditions of incubation

in the urine and sera of these patients with DRA, butwith TGF-b1, AGE-b2m, and TGF-b1 antibody additions.not in that of healthy individuals [4]. The pathogenesisResults. There was positive staining for TGF-bs (types 1, 2,

and 3) and their receptors (types I, II, and III) in infiltrated of DRA is initiated by the presence of AGE-inducedmacrophages (CD681), synovial lining cell, as well as vascular monocyte chemotaxis and the release of bone-resorbingwalls around amyloid deposition. AGE-b2m also induced TGF-b1 cytokines from macrophages, such as tumor necrosisproduction by macrophages in a dose-dependent manner (410 6

factor-a (TNF-a) and interleukin-b (IL-b), which stimu-80 pg/mL at 12.5 mg/mL, 621 6 62 pg/mL at 25 mg/mL, andlate collagenase synthesis in cultured synovial cells [6].776 6 62 pg/mL at 50 mg/mL of AGE-b2m). AGE-b2m inducedThe release of these cytokines may subsequently lead tosignificant TNF-a and IL-1Ra production by macrophage. The

addition of exogenous TGF-b1 (0.1 to 10 ng/mL) decreased collagen and connective tissue breakdown.AGE-b2m–induced TNF-a production and increased IL-1Ra Transforming growth factor-b (TGF-b) is a well-production in a dose-dependent fashion. IL-1b production was known multifunctional cytokine produced by variety ofnot effected by any experimental conditions. In chemotaxis

cell types: platelets, monocytes, macrophages, lympho-assay, anti–TGF-b1 antibody (0.1 to 10 mg/mL) attenuatedcytes, fibroblasts, osteoblasts, osteoclasts, endothelialAGE-b2m–induced monocyte chemotaxis.

Conclusions. These results provide the first evidence to our cells, and neoplastic cells [7]. TGF-b is involved in theknowledge for the presence of TGF-b in DRA tissue, as well regulation of cell proliferation, differentiation, angiogen-as the stimulatory action of AGE-b2m on tissue macrophages. esis, as well as extracellular matrix protein production

[8, 9]. Among its multifunctional effects, TGF-b is alsostrongly implicated in the inflammatory process and tis-Key words: amyloidosis, chemotaxis, cytokines, hemodialysis, TGF-b,

AGE-b2m. sue repair from its onset to its resolution [10]. Releasedat the site of inflammation, TGF-b promotes leukocyteReceived for publication December 15, 1998recruitment, including monocytes, neutrophils, and mastand in revised form July 12, 1999

Accepted for publication September 28, 1999 cells, as well as the activation and cytokine synthesis bythese cells [11–16]. For example, TGF-b rapidly induces 2000 by the International Society of Nephrology

697

Matsuo et al: TGF-b in dialysis-related amyloidosis698

Table 1. Profile of patients in this study plex (ABC) method using Vectastain ABC kit (VectorCo., Burlingame, CA, USA). Rabbit antihuman TGF-bsDuration of Cause of renal

Case Age years Sex HD years failure (isoform 1, cat# sc-146; 2, cat# sc-90; 3, cat# sc-82; SantaCruz Biotechnology, Santa Cruz, CA, USA), TGF-bR-I1 60.5 F 15.3 CGN

2 48.3 F 14.8 CGN (cat# sc-398), TGF-bR-II (cat# sc-220) polyclonal anti-3 50.7 M 17.1 CGN bodies, goat antihuman TGF-bR-III (cat# sc-6199) poly-4 56.2 M 20.2 CGN

clonal antibody (Santa Cruz Biotechnology), rabbit anti-5 56.3 M 16.8 CGN6 65.2 F 16.5 CGN human b2m polyclonal antibody, mouse antihuman7 49.0 M 16.2 CGN macrophage (CD68) monoclonal antibody (Dako, Car-8 50.1 M 15.4 CGN

pinteria, CA, USA), and mouse antihuman AGEs mono-Abbreviations are: M, male; F, female; CGN, chronic glomerulonephritis.

clonal antibody (Wako Chemical, Richmond, VA, USA)were used as the primary antibodies. After inhibition ofthe endogenous peroxidase activity, the sections wereincubated with the primary antibodies in a humidifiedincreased levels of steady-state mRNA for TNF-a andchamber at 48C overnight. The secondary antibody reac-IL-1b, and secretion of TNF-a and IL-1b protein intion was accomplished at room temperature by incubat-monocytes [11, 17, 18]. On the other hand, TGF-b alsoing with the biotinylated antibodies for 60 minutes; there-antagonizes the biological activity of IL-1b by the down-after, the sections were incubated with ABC at roomregulation of IL-1 receptors [19] and synthesis of inter-temperature for 40 minutes. The immunocomplexes inleukin 1 receptor antagonists (IL-1Ra) [20, 21]. As mono-the sections were visualized with 0.001% (wt/vol) H2O2cytes differentiate to macrophages, TGF-b has beenand 0.005% diaminobenzidine in Tris-buffered salineshown to suppress macrophage release of reactive oxy-(pH 7.6), and the sections were then counterstained withgen intermediates [22], reactive nitrogen intermediateshematoxylin. Negative controls were treated with nonim-[23, 24], and the production of TNF-a and IL-1 [25, 26].munized IgG as the primary antibody.Thus, TGF-b inhibits the proinflammatory role of differ-

entiated macrophages, thereby promoting termination In vitro preparation of advanced glycation endof the inflammatory response. product proteins

These well-known properties of TGF-b, along withAdvanced glycation end product proteins were pre-the inflammatory process of DRA, suggest a potential

pared in vitro as previously described [27]. Briefly, 1link between TGF-b and DRA. In an effort to identifymg/mL of purified normal human b2m (Cortex Biochem,

the role of TGF-b in the pathogenesis of DRA, we stud- San Leando, CA, USA) or bovine serum albumin (BSA;ied the presence of TGF-bs, as well as their receptors Sigma, St. Louis, MO, USA) were incubated at 378Cin the synovial tissues of DRA. In addition, we investi- for 60 days with 100 mmol/L d-glucose in 100 mmol/Lgated the complex interplay between TGF-b, its chemo- phosphate buffer (pH 7.4) containing 200 U/mL of peni-tactic activity, and activity to modify cytokine release. cillin, 80 mg/mL of gentamicin, and 1.5 mmol/L of phenyl-Our results indicate the presence of TGF-bs and their methylsulfonyl fluoride under sterile conditions. b2m, in-receptors in the synovial tissues of DRA; in addition, cubated in an identical manner in the absence of glucose,we demonstrate the ability of AGE-b2m to induce secre- was used as control. Unincubated b2m was also used astion of TGF-b from monocyte-derived macrophages, an additional control. After incubation, all samples werewhich, in turn, down-regulate proinflammatory cytokine dialyzed against phosphate buffer to remove free glu-production by these same macrophages. cose. AGE proteins were brownish and were character-

ized by a fluorescence spectrophotometer at a proteinconcentration of 0.2 mg/mL in 10 mmol/L phosphateMETHODSbuffer (pH 7.2). The maximum emission fluorescencePathological investigationsintensities at 430 nm were determined upon excitation

Resected specimens of the carpal tunnel synovium at 360 nm [4]. AGE proteins also reacted with anti-AGEwere collected from eight CHD patients. The profile of antibody that recognizes AGE-BSA, AGE-HAS, andpatients is summarized in Table 1 (mean age, 54.5 6 6.1 AGE-hemoglobin, but not nonglycated protein or theyears; range, 48.3 to 65.2; mean duration of hemodialysis, early products of the Maillard reaction. Endotoxin levels16.5 6 1.7 years; range, 14.8 to 20.2). These were fixed in all samples were measured by LimLus amoebocytein formalin, embedded in paraffin, and cut at 4 mm serial lysate assay (E-toxate; Sigma) and were found to besections. The existence of the amyloid deposition was below the detection limit (,0.2 ng/mL).confirmed by Congo red staining in all eight samples.

Chemotaxis assaysAn immunohistochemical study for TGF-bs (isoforms 1,2, and 3), TGF-b receptors (types I, II, and III), b2m, Human peripheral blood mononuclear cells (PBMCs)

from the blood of four normal healthy volunteers wereAGE, and CD68 was performed by the avidin-biotin com-

Matsuo et al: TGF-b in dialysis-related amyloidosis 699

Table 2. TGF-b1 inhibited AGE-b2m-induced TNF-a Statistical analysissecretion by macrophages

Data are expressed as mean 6 SD. Differences wereTGF-b1 ng/mL

AGE-b2m Ab examined by the unpaired t-test and Mann–Whitneylg/ml 0 0.1 1 10 10 lg/mL U-test. A P value , 0.05 was considered statistically0 1366 1366 1064 963 1064 significant.12.5 372641 306626 196616 148629 42865425 474660 431647 270634 175620 54965250 770673 675679 452644 356646 823630 RESULTS

Levels of TNF-a in supernatants from in vitro-derived macrophages culturedImmunohistochemical detection of TGF-bs andunder various conditions of AGE-b2m (12.5, 25, and 50 mg/mL), TGF-b1 (0.1,

1, and 10 ng/mL) anti–TGF-b1 antibody (10 mg/mL) for 24 hours. TNF-a levels TGF-b receptors in the amyloid tissues(pg/mL) were determined by ELISA. Data are shown as means 6 sd from fourseparate representative experiments using monocytes from different donors (N 5 Immunostaining with anti-AGE antibody demon-4). Ab is anti–TGF-b1 antibody.

strated that most of AGE was detected in the b2m-posi-tive amyloid deposited area (Fig. 1 A, B). CD68-positivemacrophages were the major infiltrating cells around theamyloid deposits and in the synovial lining layer (Fig.isolated by centrifugation on Ficoll-Paque (Pharmacia1C) in agreement with the previous report [30]. All eightLKB, Uppsala, Sweden) and suspended in Hank’s bal-samples of DRA tissues demonstrated the presence ofanced salt solution at 4 3 105 cells/mL. Chemotaxis assayall three human TGF-bs and their receptors. TGF-b1,of human monocyte was performed in 24 wells of doubleTGF-b2, and TGF-b3 were mainly detected in the infil-chambers separated by polycarbonate membranes (5 mm;trated CD68-positive macrophages and synovial liningCorning Coster Corp., Cambridge, MA, USA) [28]. Acells (Fig. 1 D–F). They were also localized in the vesselPBMC suspension (0.1 mL) was added to the upperwalls and faintly detected in the synovial intimal cells.compartment of the chamber, and the test substances inTheir respective receptors (types I, II, and III) wereHank’s balance salt solution were added to the upperalso localized in the same area as TGF-bs (Fig. 1 G–I).compartment or the lower compartment (0.6 mL), asNonimmunized IgG antibody was used as a negativeindicated in Table 2 and Figure 6 (only lower compart-control (Fig. 1J).ment used). The chambers were incubated at 378C for

four hour. After removing nonmigrating cells, mem-Cytokine production of humanbranes were fixed in methanol and stained with Giemsa.monocyte-derived macrophagesEach experiment was performed twice.

Effect of AGE-b2m on TGF-b1 secretion by macro-Cell culture and measurement of cytokines phages. AGE-b2m induced TGF-b1 production from hu-

man monocyte-derived macrophages in a dose-dependentHuman monocytes were isolated from PBMCs by themanner (410 6 80 pg/mL at 12.5 mg/mL, 621 6 108adherence method [29]. For this experiment using fullypg/mL at 25 mg/mL, and 776 6 107 pg/mL at 50 mg/mLdifferentiated human monocyte-derived macrophages,of AGE-b2m, P , 0.01 vs. normal b2m, as well as com-monocytes were placed in the 24-well culture plates and

allowed to differentiate in RPMI 1640 containing 10% pared with those at concentrations of 12.5 and 50 mg/mLof AGE-b2m; Fig. 2). Similarly, AGE-BSA (50 mg/mL)fetal bovine serum (Sigma Chemical Co.) for seven days.

All wells were thereafter washed with phosphate-buf- also induced TGF-b1 secretion (832 6 97 pg/mL), sug-gesting that the AGE moiety is the primary effectorfered saline and incubated overnight in RPMI 1640 with-

out serum. Finally, cells (4 3 105 cells/well) were incu- agent. However, under similar conditions, normal b2minduced very low levels of TGF-b1 production comparedbated for 24 hours at 378C in various conditions using

several concentrations of b2m, AGE-b2m, recombinant with AGE-b2m (20 6 2 pg/mL at 12.5 mg/mL, 21 6 12pg/mL at 25 mg/mL, and 28 6 14 pg/mL at 50 mg/mLhuman TGF-b1 (R&D Systems, Minneapolis, MN, USA),

anti–TGF-b1 antibody (cat# AB-101-NA; R&D Sys- of b2m). Anti–TGF-b1 antibody (10 mg/mL) completelysuppressed AGE-b2m–induced TGF-b1 production. LPStems), lipopolysaccharide (LPS; 0128:B12; Sigma) as de-

scribed in the figure legends; supernatants were collected used a positive control (0.1 mg/mL) induced 192 6 24pg/mL of TGF-b1. Neither anti–TGF-b1 antibody norand assayed for TGF-b1, TNF-a, IL-1b, IL-1Ra, using

enzyme-linked immunosorbent assay (ELISA) kits (R&D medium alone induced a detectable level of TGF-b1production in the supernatant.Systems, Minneapolis, MN, USA). According to TGF-b1

measurement, the acidification procedure was performed Effect of TGF-b1 on TNF-a secretion by macro-phages. Table 2 and Figure 3 depict the effect of AGE-following the manufacturer’s instructions. The minimum

detectable level was 7.0 pg/mL for TGF-b1, 4.4 pg/mL b2m on TNF-a production. As can be seen, AGE-b2minduced significant amounts of TNF-a production in afor TNF-a, 1.0 pg/mL for IL-1b, and 14.0 pg/mL for

IL-1Ra. Each experiment was repeated four times. dose-dependent manner compared with normal b2m-



Fig. 1. Immunohistochemical study for transforming growth factor be-tas (TGF-bs) and their receptors in the synovial tissues of dialysisrelated amyloidosis (DRA). (A) Anti-AGE immunostaining. (B) Anti-b2-microglobulin (anti-b2m) immunostaining. (C) Anti-CD68 immuno-staining. (D) Anti–TGF-b1 immunostaining. (E) Anti–TGF-b2 immuno-staining. (F) Anti–TGF-b3 immunostaining. (G) Anti–TGF-b receptortype I immunostaining. (H) Anti–TGF-b receptor type II immunostain-ing. (I) Anti–TGF-b receptor type III immunostaining. (J) Antinon-immunized IgG immunostaining. Most of AGE was detected in the amy-loid deposition. CD68-positive macrophages mainly infiltrated aroundthe amyloid deposition and in the synovial lining layer. TGF-b1, -2,and -3 were mainly detected in the infiltrated CD68-positive macro-phages and synovial lining cells. They were also localized in the vesselwalls, although faintly detected in the synovial intimal cells. Their recep-tors of types I, II, and III were also almost localized in the same areaof TGF-bs. Original magnifications of (A–J) were 3100.


Fig. 2. Effect of advanced glycation end-prod-uct modified b2m (AGE-b2m) on TGF-b1 se-cretion by macrophages. Levels of TGF-b1 insupernatants from in vitro-derived macro-phages cultured under various conditions ofnormal b2m (12.5, 25, 50 mg/mL), AGE-b2m(12.5, 25, and 50 mg/mL), anti–TGF-b1 anti-body (10 mg/mL), AGE-BSA (50 mg/mL), andLPS (0.1 mg/mL) for 24 hours. TGF-b1 levels(pg/mL) were determined by ELISA. Data areshown as means 6 SD from four separate rep-resentative experiments using monocytes fromdifferent donors (N 5 4). Abbreviation Ab isanti–TGF-b1 antibody. *P , 0.01 vs. normalb2m. **P , 0.01 between 12.5 and 50 mg/mLof AGE-b2m.

Fig. 3. Effect of TGF-b1 on TNF-a secretion by macrophages. Levels of TNF-a in supernatants from in vitro-derived macrophages culturedunder various conditions of normal b2m (12.5, 25, and 50 mg/mL), AGE-b2m (12.5, 25, and 50 mg/mL), AGE-b2m and TGF-b1 (0.1, 1, and 10ng/mL), AGE-b2m and anti–TGF-b1 antibody (10 mg/mL), AGE-BSA (50 mg/mL), and LPS (0.1 mg/mL) for 24 hours. TNF-a levels (pg/mL)were determined by ELISA. Data are shown as means 6 sd from four separate representative experiments using monocytes from different donors(N 5 4). Ab is anti–TGF-b1 antibody. *P , 0.001 vs. normal b2m. **P , 0.005 between those of 12.5 and 50 mg/mL of AGE-b2m.


Table 3. TGF-b1 enhanced AGE-b2m-induced IL-1Ra secretion by macrophages

TGF-b1 ng/mLAb

AGE-b2m lg/mL 0 0.1 1 10 10 lg/mL

0 69628 431639 788677 1,2406229 9061212.5 304625 461653 1,1046193 1,6236135 8562125 4276194 745652 1,5096370 1,8866428 236610450 1,9096118 2,2246335 3,1506313 3,3886411 5476199

Levels of IL-1Ra in supernatants from in vitro-derived macrophages cultured under various conditions of AGE-b2m (12.5, 25, and 50 mg/mL), TGF-b1 (0.1, 1,and 10 ng/mL), and anti–TGF-b1 antibody (10 mg/mL) for 24 hours. IL-1Ra levels (pg/mL) were determined by ELISA. Data are shown as means 6 SD from fourseparate representative experiments using monocytes from different donors (N 5 4). Ab is anti–TGF-b1 antibody.

induced TNF-a production (372 6 41 pg/mL in response report [32]. Neither anti–TGF-b1 antibody nor mediumalone stimulated IL-1Ra secretion. TGF-b1 alone stimu-to 12.5 mg/mL of AGE-b2m vs. 41 6 9 pg/mL in response

to 12.5 mg/mL of b2m; 474 6 60 pg/mL in response to 25 lated IL-1Ra secretion as previously reported (Table 3)[20, 21].mg/mL of AGE-b2m vs. 45 6 17 pg/mL in response to 25

mg/mL of b2m; 770 6 73 pg/mL in response to 50 mg/mL Effect of TGF-b1 on IL-1b secretion by macrophages.A low level (less than 4 pg/mL) of IL-1b was secretedof AGE-b2m vs. 91 6 37 pg/mL in response to 50 mg/mL

of b2m, P , 0.001 all three different concentrations) [6]. into the culture media by culture-derived macrophagesunder all conditions, including LPS (Fig. 5). There wasAGE-BSA (50 mg/mL) also induced TNF-a production

(690 6 110 pg/mL) as previously reported [31]. no effect of TGF-b1 on IL-1b secretion.The addition of exogenous TGF-b1 (0.1 to 10 ng/mL)

Involvement of TGF-b in AGE-b2m–induceddecreased AGE-b2m–induced TNF-a production in amonocyte chemotaxisdose-dependent manner (Table 2 and Fig. 3). On the other

hand, the combination of AGE-b2m and anti–TGF-b1 To determine the chemotactic activity of TGF-b1 formonocytes at different concentrations [11], recombinantantibody (10 mg/mL) slightly but not significantly in-

creased TNF-a production compared with AGE-b2m human TGF-b1 was tested at concentrations rangingfrom 0.1 to 10 ng/mL (Fig. 6A). Compared with N-alone. Similar results were observed with AGE-BSA–

induced TNF-a production (data not shown). LPS used Formyl-Met-Leu-Phe(FMLP) (1026 m) as a positive con-trol, TGF-b1 had a lower chemotactic activity that wasas a positive control (0.1 mg/mL) induced 1798 6 324

pg/mL of TNF-a production. Anti–TGF-b1 antibody and present even at a concentration as high as 10 ng/mL, andwas not concentration dependent. Checkerboard analy-medium alone induced a very low level of TNF-a produc-

tion in the supernatant (10 6 4 and 13 6 6 pg/mL, sis indicated that cell migration was not random (chemo-kinetic) but directional (chemotactic; Table 4). The pres-respectively). TGF-b1 alone did not stimulate TNF-a

production (Table 2). ence of b2m did not appear to affect the chemotacticactivity of TGF-b1 at either of the three different concen-Effect of TGF-b1 on IL-1Ra secretion by macrophages.

As shown in Table 3 and Figure 4, AGE-b2m also in- trations. b2m alone did not enhance the chemotactic ac-tivity and was no different than medium alone (Fig. 6A).duced significant amounts of IL-1Ra from human mono-

cyte-derived macrophages in a dose-dependent manner Next, the interaction between TGF-b and AGE-b2min determining monocyte chemotaxis was tested. Asin the culture media compared with that induced by

normal b2m (304 6 25 pg/mL in response to 12.5 mg/mL shown in Figure 6B, AGE-b2m induced monocyte che-motaxis in a dose-dependent manner. Twenty-five mg/mLof AGE-b2m vs. 69 6 22 pg/mL in response to 12.5

mg/mL of b2m, P , 0.05) Similar differences were ob- of AGE-b2m induced comparable chemotactic activityto TGF-b1. The addition of TGF-b1 (0.1 to 10 ng/mL)served in response to 50 mg/mL of AGE-b2m and b2m

(P , 0.01). Furthermore, like AGE-b2m, AGE-BSA (50 did not significantly change the chemotactic activity ofAGE-b2m alone. AGE-BSA (50 mg/mL) also showedmg/mL) also induced IL-1Ra production (2160 6 254

pg/mL). The addition of exogenous TGF-b1 (0.1 to equivalent chemotactic activity to that of AGE-b2m.On the other hand, the addition of anti–TGF-b1 anti-10 ng/mL) enhanced AGE-b2m–induced IL-1Ra produc-

tion in a dose-dependent manner. In contrast, anti– body at concentrations of greater than 0.1 mg/mL slightlyinhibited AGE-b2m–induced monocytes chemotaxis. ForTGF-b1 antibody (10 mg/mL) inhibited but did not com-

pletely suppress AGE-b2m–induced IL-1Ra production. example, at a concentration of 10 mg/mL of anti–TGF-b1antibody, AGE-b2m–induced chemotaxis at a concentra-Similar results were observed with AGE-BSA–induced

IL-1Ra production (data not shown). LPS (0.1 mg/mL) tion of 12.5 mg/mL was reduced by 55%, whereas at aconcentration of 50 mg/mL, AGE-b2m chemotaxis wasdid not stimulate IL-1Ra secretion from culture-derived

macrophage, and this result is consistent with a previous reduced by 50%. There was no difference between the


Fig. 4. Effect of TGF-b1 on IL-1Ra secretion by macrophages. Levels of IL-1Ra in supernatants from in vitro-derived macrophages culturedunder various conditions of normal b2m (12.5, 25, and 50 mg/mL), AGE-b2m (12.5, 25, and 50 mg/mL), AGE-b2m and TGF-b1 (0.1, 1, and 10ng/mL), AGE-b2m and anti–TGF-b1 antibody (10 mg/mL), AGE-BSA (50 mg/mL), and LPS (0.1 mg/mL) for 24 hours. IL-1Ra levels (pg/mL)were determined by ELISA. Data are shown as means 6 SD from four separate representative experiments using monocytes from different donors(N 5 4). Ab is anti–TGF-b1 antibody. *P , 0.05 vs. 12.5 mg/mL normal b2m. **P , 0.01 vs. 50 mg/mL normal b2m. ***P , 0.005 between thoseof 12.5 and 50 mg/mL of AGE-b2m.

extent of inhibition by anti–TGF-b1 antibody at concen- matory process [10]. By using human culture-derivedtrations of 0.5 versus 10 mg/mL. IgG did not have an affect macrophages, our study demonstrated the possible in-on AGE-b2m–induced monocyte chemotaxis (Fig. 6C). volvement of TGF-b in cytokine traffic related to the

formation of DRA. Thus, although AGE-b2m is a well-known proinflammatory mediator demonstrated by itsDISCUSSIONability to induce large concentrations of TNF-a, it also

The results of this study have several important impli- induces large amounts of TGF-b1 production from hu-cations on the pathogenesis of DRA. Specifically, to our man monocyte-derived macrophages in a dose-depen-knowledge, this is the first study to demonstrate that all dent fashion (Fig. 2). Our study also demonstrated thatthree human isoforms of TGF-bs and their receptors

TGF-b1 decreases AGE-b2m–induced TNF-a produc-are localized in the infiltrated macrophages and synovialtion, as shown in Figure 3. This is consistent with the well-lining cells around AGE-modified amyloid deposits inknown bifunctional actions of TGF-b, which displayssynovial tissues of DRA. The pathobiology of DRA isinhibitory as well as stimulatory effects of cytokine re-initiated by AGE-b2m formation on AGE-modified col-lease in diverse cells. Specifically, TGF-b has been re-lagen [33]. This study immunohistochemically indicatedported to stimulate human mononuclear leukocytes tothat AGE formation is implicated in the production ofaccumulate mRNA for TNF and IL-1 [11, 12, 17, 18], asTGF-bs and their receptors in DRA tissues. Once se-well as to secrete TNF and IL-1 protein [18]. On the othercreted into the extracellular environment, TGF-b inter-hand, in inflammatory peritoneal murine macrophages,acts with specific membrane receptors in which receptorTGF-b neither elicited TNF nor IL-1 protein productiontype I and type II forming a heteromeric complex arenor induced their respective cytokine mRNA [26]. Simi-required for signal transduction [34]. Thus, we speculatelarly, when human PBMCs or murine macrophages werethat this interaction mechanism may occur in DRA tissues.stimulated with LPS, the release of TNF and IL-1 wasTransforming growth factor-b is a well-known multi-

functional cytokine that is often implicated in the inflam- inhibited by TGF-b [17, 25]. In this study, TGF-b elicited


Fig. 5. Effect of TGF-b1 on IL-1b secretion by macrophages. Levels of IL-1b in supernatants from in vitro-derived macrophages cultured undervarious conditions of normal b2m (12.5, 25, and 50 mg/mL), AGE-b2m (12.5, 25, and 50 mg/mL), AGE-b2m and TGF-b1 (0.1, 1, and 10ng/mL),AGE-b2m and anti–TGF-b1 antibody (10 mg/mL), AGE-BSA (50 mg/mL), and LPS (0.1 mg/mL) for 24 hours. IL-1b levels (pg/mL) were determinedby ELISA. Data are shown as means 6 sd from four separate representative experiments using monocytes from different donors (N 5 4). Ab isanti–TGF-b1 antibody.

an inhibitory effect on AGE-induced TNF-a production TGF-b1 may be indirectly involved in IL-1b activity. Asignificant finding in our study is that AGE proteinsby monocyte-derived macrophages.

Unlike previous reports of IL-1b secretion by AGE- induced IL-1Ra production from monocyte-derivedmacrophages (Fig. 4). The robust inhibition of IL-1Rab2m [6, 28], our study indicates that AGE-proteins do

not stimulate IL-1b secretion from human monocyte- production by the addition of anti–TGF-b1 antibodystrongly suggests the involvement of TGF-b in this pro-derived macrophages. Furthermore, TGF-b had no stim-

ulatory effect on IL-1b secretion by these macrophages cess, most likely secondary to its anti-inflammatory ef-fect. Our in vitro model therefore suggests that, although(Fig. 5). A possible explanation for this result is that

human monocytes cultured over six days exhibited low AGE proteins promote a proinflammatory action via itspromotion of TNF-a release, they also have an anti-levels of both IL-1b mRNA and protein production in

comparison with freshly isolated monocytes under all inflammatory action via its promotion of TGF-b release.Nevertheless, it is likely that the net effect is a proin-experimented conditions, including LPS stimulation [35].

Our results are thus consistent with recent studies that flammatory action, resulting in the presence of chronicinflammation in DRA.have suggested that various populations of human tissue

macrophages, including alveolar and peritoneal cells, are In this study, we have also demonstrated that AGEproteins may induce the accumulation of monocytes viapoor agonists for IL-1b production in response to LPS

stimulation [35–37]. In addition, the previous reports its production of TGF-b. TGF-b is a well-known potentchemotactic factor for monocytes at concentrations rang-have shown that TGF-b antagonizes the biologic activity

of IL-1b in two ways: down-regulation of IL-1 receptors ing from 0.1 to 10 pg/mL [11]. In our study, even at a highconcentration of TGF-b1 (0.1 to 10 ng/mL), TGF-b1 stillbecause of a direct effect of TGF-b [19] and induction

of IL-1Ra, which competitively binds to the IL-1 receptor had chemotactic activity (Fig. 6A), although it was weakercompared with the previous report [11]. AGE-b2m alsowithout triggering an activation signal [20, 21].

Although we were not able to show any direct effect had a direct chemotactic activity at a low concentration(12.5 mg/mL), which increased further at high concen-of TGF-b1 on IL-1b secretion, our results suggest that


Fig. 6. Involvement of TGF-b1 in AGE-b2minduced monocyte chemotaxis. Monocyte mi-gration was tested in response to increasedconcentrations of (A) TGF-b1 (0.1, 1, and 10ng/mL), normal b2m (12.5, 25, and 50 mg/mL)alone and those various combinations (B)AGE-b2m (12.5, 25, and 50 mg/mL), TGF-b1(0.1, 1, and 10 ng/mL) alone and those variouscombinations, AGE-BSA (50 mg/mL; (C) vari-ous combinations of AGE-b2m (12.5, 25, and50 mg/mL) and anti–TGF-b1 antibody (Ab;0.1, 0.5, 1, and 10 mg/mL), and nonimmuneIgG. FMLP (1026 mol/L) and medium itselfwere used as a positive, negative control, re-spectively. Chemotactic activity is defined asthe mean number of monocytes that had mi-grated to the lower surface of the filters. Thenumber of cells per eight high-power fields(hpf; 3400) was counted through a micro-scope. Cell migration assay was duplicated us-ing healthy human PBMCs. TGF-b1 had thesame level of chemotactic activity during thisrange. Data are shown as means 6 SD fromfour representative experiments (N 5 4).

trations (25, 50 mg/mL). Although the combination of that TGF-b1 is involved in AGE-b2m–induced monocytechemotaxis to certain extent, but there are other factors,TGF-b and AGE-b2m did not increase chemotactic activ-

ity, the addition of anti-TGF-b1 antibody suppressed it yet to be identified, that play a role in this process.A recent report suggested that the interaction of AGEby approximately 50% (Fig. 6C). This result suggests


Table 4. Checkerboard analysis of human monocytes migration REFERENCESresponse to TGF-b1

1. Drueke TB: Beta-2-microglobulin amyloidosis and renal bone dis-TGF-b1 ng/ml in upper compartment ease. Miner Electrolyte Metab 17:261–272, 1991TGF-b1 ng/ml in

2. Koch KM: Dialysis-related amyloidosis [clinical conference]. Kid-lower compartment 0 0.1 1 10ney Int 41:1416–1429, 1992

0 6.263.4 8.463.7 8.663.2 6.962.8 3. Gejyo F, Yamada T, Odani S, Nakagawa Y, Arakawa M, Kuni-0.1 21.363.4 9.364.1 8.363.4 7.163.2 tomo T, Kataoka H, Suzuki M, Hirasawa Y, Shirahama T, Cohen1 20.762.1 19.964.2 7.563.9 6.864.0 AS, Schmid K: A new form of amyloid protein associated with10 20.962.9 21.163.8 19.763.6 6.764.3 chronic hemodialysis was identified as beta 2-microglobulin. Bio-

chem Biophys Res Commun 129:701–706, 1985Data are shown as means 6 sd from four representative experiments and4. Miyata T, Oda O, Inagi R, Iida Y, Araki N, Yamada N, Horiuchichemotactic activity is defined as mean number of monocytes that had migrated

to lower surface of the filters in response to the TGF-b1 (0.1, 1, and 10 ng/mL). S, Taniguchi N, Maeda K, Kinoshita T: Beta 2-microglobulinThe number of cells per eight high-power fields (3400) was counted through a modified with advanced glycation end products is a major compo-microscope. Cell migration assay was duplicated using healthy human PBMC. nent of hemodialysis-associated amyloidosis. J Clin Invest 92:1243–

1252, 19935. Niwa T, Miyazaki S, Katsuzaki T, Takemichi N, Takei Y, Miya-

zaki T, Morita T, Hirasawa N: Immunohistochemical detectionof advanced glycation end products in dialysis-related amyloidosis.and the receptor for AGE (RAGE) on monocyte con- Kidney Int 48:771–778, 1995

tributed to monocyte chemotaxis and was prevented by 6. Miyata T, Inagi R, Iida Y, Sato M, Yamada N, Oda O, MaedaK, Seo H: Involvement of beta 2-microglobulin modified with ad-soluble RAGE or anti-RAGE antibody [38]. Althoughvanced glycation end products in the pathogenesis of hemodialysis-this study showed that AGE proteins induced monocyte associated amyloidosis: Induction of human monocyte chemotaxis

chemotaxis through TGF-b, it is unclear how TGF-b and and macrophage secretion of tumor necrosis factor-alpha andinterleukin-1. J Clin Invest 93:521–528, 1994RAGE interact with each other. Further investigations

7. Roberts AB, Thompson NL, Heine U, Flanders C, Sporn MB:are necessary to clarify this complex interaction. Transforming growth factor-beta: Possible roles in carcinogenesis.The exact mechanism by which TGF-b may be in- Br J Cancer 57:594–600, 1988

8. Massague J: The transforming growth factor-beta family. Annuvolved in DRA of dialysis patients is not clear at thisRev Cell Biol 6:597–641, 1990time. Several studies have shown that the circulating

9. Roberts AB, Sporn MB: Physiological actions and clinical applica-TGF-b level is significantly higher in hemodialysis pa- tions of transforming growth factor-beta (TGF-beta). Growth Fac-

tors 8:1–9, 1993tients compared with normal subjects [39–41]. In the10. Wahl SM, McCartney-Francis N, Mergenhagen SE: Inflamma-circulation, TGF-b essentially exists in an inactive com-

tory and immunomodulatory roles of TGF-beta. Immunol Todayplex with the protease inhibitor a2-macroglobulin (a2M). 10:258–261, 1989

11. Wahl SM, Hunt DA, Wakefield LM, McCartney-Francis N,Consequently, TGF-b may exist in two distinct, biologi-Wahl LM, Roberts AB, Sporn MB: Transforming growth factorcally inactive forms: the original latent complex with itstype beta induces monocyte chemotaxis and growth factor produc-

precursor polypeptide and a second complex with a2M, tion. Proc Natl Acad Sci USA 84:5788–5792, 1987which can be dissociated by heparin [42, 43]. Techne- 12. Wiseman DM, Polverini PJ, Kamp DW, Leibovich SJ: Trans-

forming growth factor-beta (TGF beta) is chemotactic for humantium-labeled heparin kinetics have revealed that the ac-monocytes and induces their expression of angiogenic activity.cumulation of radioactivity was significantly higher at Biochem Biophys Res Commun 157:793–800, 1988

the knee and on the shoulder in hemodialysis patients 13. Brandes ME, Mai UE, Ohura K, Wahl SM: Type I transforminggrowth factor-beta receptors on neutrophils mediate chemotaxis tocompared with that of normal subjects [44]. These find-transforming growth factor-beta. J Immunol 147:1600–1606, 1991ings indicate that TGF-b may be repeatedly activated 14. Reibman J, Meixler S, Lee TC, Gold LI, Cronstein BN, Haines

by the use of heparin in hemodialysis treatment, particu- KA, Kolasinski SL, Weissmann G: Transforming growth factorbeta 1, a potent chemoattractant for human neutrophils, bypasseslarly in the osteoarticular structures, which is the pre-classic signal-transduction pathways. Proc Natl Acad Sci USAdominant lesion of DRA. Further research is needed to 88:6805–6809, 1991

clarify this process. 15. Gruber BL, Marchese MJ, Kew RR: Transforming growth factor-beta 1 mediates mast cell chemotaxis. J Immunol 152:5860–5867,In summary, this study showed that TGF-b is involved1994in the pathogenesis of DRA. While enhancing the che-

16. Wahl SM: Transforming growth factor beta (TGF-beta) in in-motaxis of monocyte, TGF-b at the same time inhibited flammation: A cause and a cure. J Clin Immunol 12:61–74, 1992

17. Chantry D, Turner M, Abney E, Feldmann M: Modulation ofthe proinflammatory role of in vitro differentiated mac-cytokine production by transforming growth factor-beta. J Immu-rophages, suggesting that it may play a dual role in thenol 142:4295–4300, 1989

pathogenesis of DRA in CHD patients. 18. McCartney-Francis N, Mizel D, Wong H, Wahl L, Wahl S:TGF-beta regulates production of growth factors and TGF-betaby human peripheral blood monocytes. Growth Factors 4:27–35,ACKNOWLEDGMENTS1990

19. Dubois CM, Ruscetti FW, Palaszynski EW, Falk LA, Oppen-This study was supported in part by National Institutes of Healthgrant DK-45604-7. We gratefully thank Ms. Rhoda Jones and Ms. heim JJ, Keller JR: Transforming growth factor beta is a potentJanice Harvell for their assistance. inhibitor of interleukin 1 (IL-1) receptor expression: Proposed

mechanism of inhibition of IL-1 action. J Exp Med 172:737–744,1990Reprint requests to Raymond M. Hakim, M.D., Ph.D., Division of

Nephrology, Vanderbilt University Medical Center, S-3223, Medical 20. Turner M, Chantry D, Katsikis P, Berger A, Brennan FM,Feldmann M: Induction of the interleukin 1 receptor antagonistCenter North, Nashville, Tennessee 37232-2372, USA.


protein by transforming growth factor-beta. Eur J Immunol 33. Hou FF, Chertow GM, Kay J, Boyce J, Lazarus JM, BraatzJA, Owen WF Jr: The interaction between b2-microglobulin and21:1635–1639, 1991advanced glycation end products in the development of dialysis21. Wahl SM, Costa GL, Corcoran M, Wahl LM, Berger AE:related-amyloidosis. Kidney Int 51:1514–1519, 1997Transforming growth factor-beta mediates IL-1-dependent induc-

34. Franzen P, ten Dijke P, Ichijo H, Yamashita H, Schulz P,tion of IL-1 receptor antagonist. J Immunol 150:3553–3560, 1993Heldin CH, Miyazono K: Cloning of a TGF beta type I receptor22. Tsunawaki S, Sporn M, Ding A, Nathan C: Deactivation ofthat forms a heteromeric complex with the TGF beta type II recep-macrophages by transforming growth factor-beta. Nature 334:260–tor. Cell 75:681–692, 1993262, 1988

35. Janson RW, Joslin FG, Arend WP: The effects of differentiating23. Ding A, Nathan CF, Graycar J, Derynck R, Stuehr DJ, Srimal S:agents on IL-1 beta production in cultured human monocytes.Macrophage deactivating factor and transforming growth factors-J Immunol 145:2161–2166, 1990beta 1, -beta 2 and -beta 3 inhibit induction of macrophage nitrogen

36. Wewers MD, Rennard SI, Hance AJ, Bitterman PB, Crystaloxide synthesis by IFN-gamma. J Immunol 145:940–944, 1990RG: Normal human alveolar macrophages obtained by bronchoal-24. Nelson BJ, Ralph P, Green SJ, Nacy CA: Differential susceptibil-veolar lavage have a limited capacity to release interleukin-1. J Clinity of activated macrophage cytotoxic effector reactions to theInvest 74:2208–2218, 1984suppressive effects of transforming growth factor-beta 1. J Immu-

37. Becker S, Johnson C, Halme J, Haskill S: Interleukin-1 produc-nol 146:1849–1857, 1991tion and antigen presentation by normal human peritoneal macro-25. Espevik T, Figari IS, Shalaby MR, Lackides GA, Lewis GD,phages. Cell Immunol 98:467–476, 1986Shepard HM, Palladino MA Jr: Inhibition of cytokine production

38. Miyata T, Hori O, Zhang J, Yan SD, Ferran L, Iida Y, Schmidtby cyclosporin A and transforming growth factor beta. J Exp MedAM: The receptor for advanced glycation end products (RAGE)166:571–576, 1987is a central mediator of the interaction of AGE-beta2 microglobu-26. Bogdan C, Paik J, Vodovotz Y, Nathan C: Contrasting mecha-lin with human mononuclear phagocytes via an oxidant-sensitivenisms for suppression of macrophage cytokine release by trans-pathway. Implications for the pathogenesis of dialysis-related amy-forming growth factor-beta and interleukin-10. J Biol Chem loidosis. J Clin Invest 98:1088–1094, 1996267:23301–23308, 1992 39. Anderson J, Briefel G, Jones JM, Ryu JH, McGuire M, Yun27. Vlassara H, Brownlee M, Cerami A: High-affinity-receptor- YP: Effects of acetate dialysate on transforming growth factor beta

mediated uptake and degradation of glucose-modified proteins: A 1, interleukin, and beta 2-microglobulin plasma levels. Kidney Intpotential mechanism for the removal of senescent macromolecules. 40:1110–1117, 1991Proc Natl Acad Sci USA 82:5588–5592, 1985 40. Mege JL, Capo C, Purgus R, Olmer M: Monocyte production of

28. Miyata T, Iida Y, Ueda Y, Shinzato T, Seo H, Monnier VM, transforming growth factor beta in long-term hemodialysis: Modu-Maeda K, Wada Y: Monocyte/macrophage response to beta lation by hemodialysis membranes. Am J Kidney Dis 28:395–399,2-microglobulin modified with advanced glycation end products. 1996Kidney Int 49:538–550, 1996 41. Suthanthiran M, Khanna A, Cukran D, Adhikarla R, Sharma

29. Treves AJ, Yagoda D, Haimovitz A, Ramu N, Rachmilewitz D, VK, Singh T, August P: Transforming growth factor-beta 1 hyper-Fuks Z: The isolation and purification of human peripheral blood expression in African American end-stage renal disease patients.monocytes in cell suspension. J Immunol Methods 39:71–80, 1980 Kidney Int 53:639–644, 1998

30. Ehlerding G, Schaeffer J, Drommer W, Miyata T, Koch KM, 42. McCaffrey TA, Falcone DJ, Brayton CF, Agarwal LA, WeltFloege J: Alterations of synovial tissue and their potential role in FG, Weksler BB: Transforming growth factor-beta activity is po-the deposition of b2-microglobulin-associated amyloid. Nephrol tentiated by heparin via dissociation of the transforming growthDial Transplant 13:1465–1475, 1998 factor-beta/alpha 2- macroglobulin inactive complex. J Cell Biol

31. Vlassara H, Brownlee M, Manogue KR, Dinarello CA, Pasa- 109:441–448, 1989gian A: Cachectin/TNF and IL-1 induced by glucose-modified 43. McCaffrey TA, Falcone DJ, Vicente D, Du B, Consigli S, Borthproteins: Role of normal tissue remodeling. Science 240:1546–1548, W: Protection of transforming growth factor-beta 1 activity by1988 heparin and fucoidan. J Cell Physiol 159:51–59, 1994

32. Janson RW, Hance KR, Arend WP: Production of IL-1 receptor 44. Majdalani G, Chomant J, Kachko A, Yanai M, Man NK: Kineticsantagonist by human in vitro-derived macrophages. J Immunol of technetium-labeled heparin in hemodialyzed patients. Kidney

Int Suppl 41:S131–S134, 1993147:4218–4223, 1991

Documents

Transforming growth factor-β is involved in the pathogenesis of dialysis-related amyloidosis