Kidney International, Vol. 47 (1995), pp. 935—944

Cellular events in the evolution of experimental diabetic

nephropathyBEs5IE A. YOUNG, RICHARD J. JOHNSON, CHARLES E. ALPERS, EUDORA ENG, KATHERINE GORDON,

JURGEN FLOEGE, and WILLIAM G. COUSER, with statistical assistance from KRIsT'Y SEIDEL

Division of Nephrology, Department of Medicine and Department of Pathology, University of Washington and the Fred Hutchinson Cancer ResearchCenter, Seattle, Washington

Cellular events in the evolution of experimental diabetic nephropathy.In several models of progressive glomerular disease, mesangial cellproliferation, phenotypic change and increased growth factor expressionprecede up-regulation of genes for extracellular matrix components(ECM) and mesangial expansion. To examine these events in diabeticnephropathy (DN) we conducted sequential studies of glomeruli in ratswith streptozotocin induced DN. We found prominent mesangial cellproliferation at three days (4.34 2.24PCNA + cells/glom vs. 1.6 0.74in controls, P < 0.001) associated with increased n-actin expression.PDGF B-chain mRNA was slightly increased at day one, and PDGFB-chain immunostaining was slightly increased at days one and six.Staining for bFGF was significantly increased at three days (2.2 0.6 vs.1.2 0.1 in controls, P < 0.01). There was also an early increase inplatelets in glomeruli of diabetic animals, and platelet depletion signifi-cantly inhibited the early phase of proliferation. In addition to mesangialcell proliferation, a prominent glomerular macrophage infiltration beganat day three and peaked at day 30 (3.94 1.47 vs. 2.08 1.13 in controls,P < 0.01). TGF-J3 mRNA increased at days 14 and 30. Insulin treatmentprevented mesangial cell proliferation, actin expression, and macrophageinfiltration, and normalized TGF-13 expression at 14 and 30 days. Thesemultiple cellular events preceded any detectable increases in glomerulargene expression or deposition of collagen I, IV or laminin.

The most common cause of end-stage renal disease in theUnited States is diabetic nephropathy (DN), accounting foralmost a third of all new patients requiring dialysis [1]. Thepathogenesis of DN is incompletely understood but is generallybelieved to involve a disorder of glomerular cell metabolismleading to accumulation of both basement membrane and non-basement membrane extracellular matrix (ECM) components,resulting in expansion of the mesangial matrix and thickening ofglomerular basement membrane (GBM) [2]. Although hypergly-cemia is the underlying metabolic abnormality, multiple othermechanisms have also been proposed including abnormal glomer-ular hemodynamic factors [3, 4], aberrant metabolic pathways [5,6], glycosylation of glomerular structures [7], and abnormalgrowth factor expression leading to ECM accumulation [8].

Recent studies from our laboratory have suggested that an earlyphase of mesangial cell proliferation and alteration in phenotype

Received for publication May 31, 1994and in revised form October 26, 1994Accepted for publication October 31, 1994

© 1995 by the International Society of Nephrology

may also be central to the development of later mesangial matrixexpansion and glomerular sclerosis in models of both immune andnon-immune glomerular injury [reviewed in 9, 10]. Thus, in theThy 1.1 model of immune-mediated mesangial proliferative gb-merulonephritis, mesangial matrix expansion is preceded by aphase of active mesangial cell proliferation, which is mediated bycomplement and platelets, and involves increased expression orrelease of growth factors including basic fibroblast growth factor(bFGF), platelet-derived growth factor (PDGF) and its receptor,and transforming growth factor beta (TGF-13) [11]. Other char-acteristics of this activated mesangial cell phenotype include denovo expression of a-smooth muscle actin and increased extracel-lular matrix production accompanied by over production of matrixmodifying enzymes [10]. Mesangial cell proliferation also pre-cedes matrix expansion in non-immune models of progressiveglomerular disease such as the remnant kidney model [12], andgbomeruloscierosis in mice transgenic for growth hormone (GH)[13, 14] and SV4O [15, 16]. We therefore hypothesized thatglomerular cell proliferation might precede, and perhaps predict,the development of glomeruloscierosis in DN as well.

In this study we report that mesangial cell proliferation occursearly in experimentally induced DN, is associated with a glomer-ular platelet infiltrate and increased expression of growth factors(bFGF, PDGF), and precedes detectable changes in mesangialECM deposition. A later phase of glomerular macrophage infil-tration was associated with increased expression of TGF-13 thatmay also be related to the development of chronic glomensloscle-rosis in diabetic nephropathy. These newly recognized cellularevents in early diabetic nephropathy may provide importantinsights into the pathogenesis of this condition.

Methods

Experimental design

Experimental diabetes was induced in 96 male, 120 to 140 g,Sprague-Dawley rats (Simonsen Laboratories, Gilroy, CA, Bantinand Kingman, Edmonds, WA, USA), with streptozotocin (Sigma,St. Louis, MO, USA), 65 mg/kg, dissolved in sodium citrate buffer(pH 4.5), given intravenously. Forty-two rats were treated withdaily injections of insulin (protamine zinc and ultralente, Eli Lillyand Company, Indianapolis, IN, USA) in doses of 5 to 12 unitssubcutaneously every 24 hours to maintain blood glucose less than150 mg/dl. Fifty-four rats served as controls and were injected with

935

936 Younget al: Cellular events in experimental diabetic nephropathy

equivalent doses of sodium citrate buffer. Animals were housed incages with standard rat chow and water ad libitum. Serum glucosewas measured at multiple time points by glucometer reading, anddaily in rats receiving insulin by serum chemstrip bG(R) (Boehr-inger Mannheim Corp., Indianapolis, IN, USA) and/or with urineglucose measurements using Tes-Tape (R) glucose enzymaticstrips (Eli Lilly). Serum glucose was measured in all animals atsacrifice by a Paramax ZY 720 Analyzer (Baxter, McGaw Park,IL, USA). Urine protein was measured at two and four weeks.Renal tissue was obtained in separate groups of animals at one,three, six, 14, and 30 days. Biopsies and glomerular mRNA wereobtained at the above selected time points to stain for cellproliferation, growth factor expression and extracellular matrixcomponents, and to assess gene expression of growth factors andmatrix components.

Renal morphology

Renal biopsies were fixed in methyl Carnoy's solution andembedded in paraffin. Four micrometer sections were stained withperiodic acid and Schiff's reagent (PAS), and counter stained withhematoxylin. Biopsies were assessed for basement membranechanges and sclerosis. Total glomerular cellularity was deter-mined by counting the number of nuclei per glomerular cross-section and subtracting the number of macrophage/monocytesstained by ED-i antibody (see below). Extracellular matrix expan-sion was assessed on 4 m sections of tissue using a modifiedGomori's silver methenamine stain, and specific extracellularmatrix antibodies.

Electron microscopy

Tissues were prepared and processed as previously described[17]. Biopsies were studied at days one and three for ultrastruc-tural changes and platelet infiltration.

Immunohistochemical stainingTissue was sectioned (4 tm) and stained using an indirect immu-

noperoxidase or immuno-alkaline phosphatase method (PDGF Bchain) as previously described [12, 18, 19]. Antigens evaluated andantibodies used included the following: Proliferating cell nuclearantigen (PCNA) using 19A2 (American Biotech Inc., Plantation,FL, USA) [20]; a-smooth muscle actin using a-SM-i (gift of G.Gabbiani) [21]; macrophages, monocytes and dendritic cells usingthe murine monoclonal antibody ED-i (Bioproducts for Science,Indianapolis, IN, USA); platelets using the murine monoclonalantibody PL-1 (gift of W.W. Baker, Groning, Netherlands) [22];mesangial cells using the anti-thy 1.1 monoclonal antibody OX-7(Serotec Ltd., Blackhorn Bicester, UK) [23]; common leukocyteantigen expressed by B lymphocytes and some T lymphocytesusing OX-22 (Accurate Chemical Corporation, Westbury, NY,USA); rat neutrophils using the RP-3 antibody (gift of F. Sendo,Yamagata, Japan) [24]; bFGF using the monoclonal antibodyDE6, (gift of T. Reilly, DuPont-Merck, Wilmington, DE, USA)[25]; PDGF B-chain using the monoclonal antibody PDGF-007(Mochida Pharmaceutical, Tokyo, Japan) [26]; and PDGF 13-re-ceptor using a rabbit polyclonal antibody to the 13-subunit of thePDGF-receptor (gift of Ron Seifert, Seattle, WA, USA) [19].Polyclonal antibodies to laminin (Chemicon, Temecula, CA,USA); heparan sulfate proteoglycan (HSPG; gift of J.R. Couch-man, Birmingham, AL, USA) [27]; collagen type I (gift of L.Iruela-Arispe, Seattle, WA, USA) [28]; and collagen type IV

(Southern Biotech, Birmingham, AL, USA) were also utilized.Controls included known positive and negative tissue as well asstaining with an irrelevant antibody of the same isotype. Allbiopsies within a group were stained at the same time.

Each biopsy was reviewed and scored by an observer blinded towhich group the animal belonged. At least 20 consecutive cross-sections of cortical glomeruli containing more than 20 discretecapillary segments were examined using each of the above anti-bodies. Mean scores of positive cells/glomerulus were calculatedfor proliferation (PCNA + cells), and for the various leukocytepopulations. Semiquantitative scores of 0 to 4+ were used toquantitate staining for the ECM components, a-actin, bFGF,PDGF B-chain and PDGF 13-receptor and platelet infiltrates aspreviously described [12, 29].

Double labeling

To determine the identity of proliferating (PCNA positive)cells, a double immunogold staining method [12, 19] was utilizedin which tissue was stained first for PCNA and then stained witheither the OX-7 antibody to mesangial cells or the ED-I antibodyto macrophages and monocytes. PCNA positive cells were thenclassified as either mesangial, macrophage, or non-classifiable aspreviously described [12, 19].

Preparation of glomendar mRNA/Northem analysisGlomeruli were isolated by differential sieving as described

previously [18]. All preparations studied had less than 10%tubular contamination. Messenger RNA was isolated according tothe method of Chomczynski and Sacchi [30], modified usingRNAsoI B® (Tel-Test, Inc. Laboratories, Friendswood, TX, USA)following the manufacturer's directions, with an additional over-night precipitation in 3 M lithium chloride for further purification[311. RNA was analyzed for content using standard 260/280 ODspectrophotometric readings and corroborated on a 1% agaroseminigel. Northern analysis was performed with purified RNAusing 15 WIane, electrophoresed through a 1% agarose/1.1%formaldehyde gel with ix MOPs buffer, and transferred to a nylonfilter (Hybond N, Amersham, Arlington Heights, IL, USA) asdescribed previously [321. The following cDNA probes were usedfor analysis:

PDGF B-chain. The 3.4 kb rat PDGF B-chain transcript wasdetected using a 3 kb EcoR I fragment of rat PDGF B-chaincDNA clone p3-4a (gift of C.M. Giachelli, Seattle, WA, USA)[321.

TGF-/31. A 985 bp Hind III/Xba I fragment of rat TGF-131cDNA plasmid was used to detect the 2.4 kb rat TGF-f31transcript (gift of Dr. Anita Roberts, NIH, Bethesda, MD, USA)[33].

bFGF. A 539-bp Xho I/Nco I fragment of rat ovarian bFGFcDNA from plasmid RObFGF-477 was used to detect the 6.8-kbrat bFGF transcript [34].

Type IVcollagen, The 6.2 and 6.8 kb type IV collagen transcriptswere detected using a 1.8 kb EcoR I/Hind III cDNA fragment ofmouse al (IV) collagen (gift of M. Kurkinen) [35].

Laminin B2 chain. The 8 kb transcript was detected using a 1.7kb EcoR I/Xba I fragment of mouse laminin B2 cDNA (gift of Y.Yamada) [36].

28S. A bovine 280 bp cDNA probe was used to detect 28Sribosomal RNA (gift of Dr. Luisa Iruela-Arispe and Dr. HeleneSage).

Young et al: Cellular events in experimental diabetic nephropathy 937

Variables measured

Time in weeks

2 4

Serum glucose mg/dlDM 430 56 443 174DM-INS 137 108 145 167Control 146 12 147 22

Weight gDM 237.0 9.6 279.1 24.8DM-IN 246.7 14.3 300.4 29.2Control 289.6 24.6a 370.9 34.la

Serum BUN mgldlDM 30.2 5.0 27.8 4.96DM-IN 25.0 5.7 30.0 5.4Control 22.8 4.5 23.5 3.3

Serum creatinine mgldlDM 0,32 0.04 0.30 0.13DM-IN 0.37 0.05 0.35 0.19Control 0.33 0.10 0.35 0.14

Urine protein mg/24 hrDM 4.6 1.3 12.1 2.2Control 2.9 •53 10.6 2.1

Urine volume mlDM 23.1 6.5 23.7 2.3Control 11.7 2.6a 12.9 2.8a

Results

All probes were labeled with [a-32P]-deoxycytidine 5-triphos-phate (3000 Cilmmol, New England Nuclear, Boston, MA, USA)by random primer extension. Filters were prehybridized andhybridized as described previously [31]. Autoradiograms wereread by laser scanning using the Biosoft scan analysis program(Burcham, Biosoft, Ferguson, MO, USA). All Northern analyseswere repeated with RNA preparations obtained from differentsets of animals. Some filters were rehybridized with additionalprobes up to a maximal of four.

Platelet depletion studies

Platelet depletion was performed in seven rats with polyclonalgoat anti-platelet IgG as described elsewhere [18]. Control rats(N = 6) received normal goat IgG. Streptozotocin was adminis-tered eight hours after the anti-platelet antibody or normal goatIgG serum injections. Platelet counts in platelet depleted animalswere less than 10,000 per mm3 at the time of streptozotocininjection, and were maintained at this level until sacrifice at threedays.

Miscellaneous measurements

Urine protein was measured by the sulfosalicylic acid method[37], using a whole serum standard (Lab Trol, Dade Diagnostics,Aquado, Puerto Rico).

Statistical analysis

Data on macrophages, PCNA, PDGF, and platelets were testedfor statistically significant group differences using the nonpara-metric Wilcoxon rank-sum test [38]. Data on other variablesstudied are expressed as means and standard deviations andanalyzed for group differences using the two-sample Student'sI-test.

Metabolic characteristics of streptozotocin diabetes

Hyperglycemia exceeding 300 mg/dl was induced in the STZtreated rats within 24 hours of STZ administration and remainedelevated throughout the 30 day study. The serum glucose indiabetic animals was significantly different from insulin treatedanimals at all time points from day three, except day six whereinsulin treated animals had elevated serum glucose (243 128

mgldl compared to 352 140 in diabetes). Diabetic rats gainedless weight than controls (Table 1). Insulin treated diabeticanimals also gained less weight than controls (Table 1) but slightlymore than diabetic animals.

Diabetic animals did not have more proteinuria than controlsover the time course of the study except at two weeks, but weresignificantly more polyuric than controls at two and four weeks(Table 1).

Glomerular cellularity

Total glomerular cellularity was assessed by counting the totalnumber of nuclei per glomerulus in PAS stained sections andsubtracting the number of infiltrating cells (assessed by ED-istaining for macrophages which were the primaly infiltratingleukocyte population). There was no significant increase in totalnuclei in glomeruli from diabetic rats compared to control ani-mals. However, cellularity was significantly higher in diabetic ratscompared to insulin treated diabetic animals at day six and 14(data not shown).

PCNA staining was increased in diabetic rats at day three andwas normal in control and insulin-treated diabetic animals (Fig.1). Increased proliferation continued in diabetic rats through day30 and was significantly different from control animals andinsulin-treated diabetic animals except at day 30. Double immu-nostaining (Table 2) confirmed that the majority (that is, 61 to

Table 1. Variables measured during the first thirty days of STZ-diabetes

(a,E5 .2

0—0+0z0a-

5

4

3

2

1

0

Abbreviations are: DM, diabetic animals; DM-IN, insulin-treated dia-betic animals; BUN, blood urea nitrogen. Values are mean so.

ap < 0.05 versus diabetic animals.

0 5 10 15 20 25 30

Day

Fig. 1. Graph of proliferating cells determined by PCNA stain in diabeticanimals (— —), diabetic animals treated with insulin (— — —), and normalcontrol (—) animals during the first 30 days after STZ induction ofdiabetes. Values are mean savi. * P < 0.05 vs. control; P < 0.05 vs.insulin.

938 Younget al: Cellular events in experimental diabetic nephropathy

Day PCNAJTHY-1 PCNAJED-1 Unclassified

3 69.16 1.266 61.08 2.41

14 61.95 4.79

8.92 3.69 21.7 9.918.96 2.84 30.4 7.315.05 2.06 33.0 10.55

To establish the nature of proliferating (PCNA-positive) cells in dia-betic animals, double labeling was done with ED-i (monocyte/macro-phages) and Thy 1 (mesangial cells) antibodies. Most PCNA-positive cells(60 to 70%) were mesangial. PCNA-positive cells that were negative forED1 and Thy 1 were designated as unclassified.

a p < 0.05 diabetic animals vs. controls'P < 0.05 diabetic animals vs. insulin treated diabetic animals

70%) of PCNA positive cells were mesangial cells at days three,six, and 14, with 7 to 13% of the proliferating cells representingmacrophages and the rest unclassifiable.

Increased numbers of platelets were found to be present inglomeruli of diabetic rats by day three compared to controlanimals and insulin treated animals (Table 3). Platelets were alsoidentified by EM in glomerular capillaries of diabetic rats at daythree along with mononuclear leukocytes (Fig. 7). Platelet deple-tion significantly decreased mesangial cell proliferation at daythree in diabetic rats compared to normal serum IgG treatedcontrols (1. 42 0.28 vs. 0.27 0.11 PCNA + cells/glom, respec-tively, P < 0.003).

To determine if a change in mesangial cell phenotype occurs inexperimental diabetes, tissue was stained for cs-actin. De novomesangial expression of a-actin was documented in diabeticanimals at days one and three, and was significantly increasedcompared to control and insulin treated animals (Fig. 2). Nodifference in cw-actin expression was noted at later time points,with all groups showing minimal to no a-actin expression.

A glomerular infiltrate of monocytes/macrophages was appar-ent as early as day three, and was maximal at day 30 (Fig. 3). Thismonocyte/macrophage infiltrate was not associated with neutro-phils, or with B cells and T cells that express the commonleukocyte antigen. Insulin therapy reduced the number of infil-trating cells to control levels except at one time point (day six)when transient hyperglycemia occurred (Fig. 3).

Growth factor expression

A significant increase in bFGF immunostaining was noted inthe mesangium of diabetic rats at day three, and decreased tobaseline levels at days 14 and 30, when compared to controls andinsulin treated diabetic animals (Fig. 4). However, no increase inbFGF mRNA could be documented in glomeruli at any timepoint.

Immunostaining for PDGF B-chain showed a slight increase indiabetic animals on day one and was persistently elevated com-pared to controls and insulin treated diabetic animals over the 30day time course (data not shown). PDGF B-chain mRNA wasminimally (10%) increased compared to control and insulintreated diabetic animals at days one, six, 14, and 30 (Table 4).PDGF a-subunit receptor staining was also slightly increased inglomeruli of diabetic animals at day three compared to controlsbut not thereafter (data not shown).

TGF-131 mRNA was also increased in glomeruli of diabeticanimals compared to controls or insulin treated diabetic animalsat day 14 and 30 (Fig. 5).

Extracellular matrix

Biopsies from diabetic animals were stained with PAS and silvermethenamine to assess changes in morphology, ECM expansion,and progression to sclerosis. Staining with PAS showed mildmesangiolysis as early as day three only in diabetic animals.Mesangiolysis was not seen in controls or in insulin treatedanimals. Silver and PAS staining confirmed some mesangialmatrix expansion by day 30 without sclerosis (data not shown). Nomorphologic changes were seen in control or insulin treateddiabetic animals.

To determine if the cellular events described above precededchanges in individual extracellular matrix components immuno-histochemical staining for type I collagen, type IV collagen,HSPG, and laminin was quantitated in mesangial matrix. Theabundance of glomerular mRNA for type IV collagen and lamininwas also assessed in diabetic, control, and insulin treated animalsat days one, three, six, 14 and 30. Immunohistochemical stainingfor type I collagen was found not to be elevated over the first 30days of streptozotocin induced diabetes. In contrast, type IVcollagen staining was found to be significantly increased in amesangial distribution in glomeruli by day 30 but not before (Fig.6). However, this increase in immunohistochemical staining wasnot associated with a significant increase in mRNA expression fortype IV (al) collagen (Table 4).

Laminin staining was also increased in diabetic animals whencompared to insulin-treated diabetic animals, but not controlanimals at day 30 (data not shown). Messenger RNA for lamininB2 was not significantly elevated in STZ diabetic rats at day 30(Table 4).

A significant reduction in glomerular staining for HSPG wasfound at days 14 and 30 in diabetic animals when compared tocontrol and insulin treated diabetic animals. The decreasedstaining of HSPG was in both a mesangial and capillary looppattern (data not shown).

Discussion

Diabetic nephropathy (DN) exceeds hypertension as a cause ofend-stage renal disease in the United States [1]. However, thecellular mechanisms of injury remain poorly understood. Multiplemechanisms have been proposed for DN including abnormalitiesin glomerular hemodynamics [3, 4, 39, 40], mesangial matrixexpansion [41] with glomerular hypertrophy [42, 43], aberration inglomerular polyol/sorbitol pathways [5, 44], glycosylation of gb-merular structures [7], increased sodium-lithium countertransport[6, 45], abnormal growth factor expression leading to ECMaccumulation [8], and directly toxic effects of exogenous insulin

Table 2. Double immunostaining results Table 3. Immunostaining for platelets (0 to 4+) in glomeruli

Day Diabetes Control Insulin

1 0.79 0.15 0.65 0.26 —3 1.0 0.05 0.68 0.16 0.66 025b6 1.0 0.30 0.90 0.08 1.03 0.21

14 0.76 0.18 0.53 0.31 0.75 0.0730 0.98 0.12 0.94 0.33 0.76 0.13

S

.5. p

Fig. 2. Alpha actin expression at day three in diabetes (A), insulin-treated diabetic animals (B), and normal control animals (C). There is de novo expressionof alpha actin in a mesangial pattern in the diabetic animal (A), which is abolished by insulin therapy and not seen in controls. Alpha actin staining incontrol and insulin-treated animals is in smooth muscle cells of the afferent arteriole (X400).

previously been recognized in vivo. Our study documents animmediate change in mesangial cell phenotype as early as day oneafter STZ administration manifested as de novo expression ofa-smooth muscle actin. This phenomenon has previously beenshown to be a feature of mesangial cell activation in severalexperimental models of glomerular disease including the ATSmodel of mesangial proliferative nephritis [31], Habu snakevenom glomerulonephritis (GN) [31], the remnant kidney modelof progressive glomerulosclerosis [12], and angiotensin TI-medi-ated hypertension [47]. Moreover, we have recently recognizedmesangial cell a-actin expression as a feature of diabetes andother glomerular diseases in humans [48]. The significance of thisphenotypic change is uncertain. In other models, it is accompa-nied by de novo expression of type I collagen indicating thatmesangial cells acquire the characteristics of "myofibroblasts" inresponse to several different forms of glomerular injury [10, 12].Alpha-actin may also augment the mesangial contractile response

0 5 10 15 20 25 30 to the increased glomerular pressures and flows which occur earlyin diabetes [49], thereby modulating local hemodynamics. The

Day prevention of a-actin expression by insulin administration sug-gests that it occurs in response to the diabetic milieu rather thanas a consequence of STZ administration, a conclusion whichapplies to the early hemodynamic changes in STZ diabetes as well[50]. There was also a mild mesangiolysis seen in diabetic animalsthat was not found in control or insulin treated animals and isanother recognized feature of diabetic nephropathy in humans[51].

An additional finding in our study was a prominent glomerularcell proliferation which was evident by day one, and peaked at daythree but persisted throughout the remaining 30 days of the study.This confirms and extends the recent observation by Nakamura etal of increased abundance of mRNA for PCNA in early strepto-zotocin nephropathy in rats [52]. Double labeling studies demon-strated that a majority of the proliferating cells were mesangial,with a minor contribution from infiltrating macrophages. Theearly proliferative response was totally abolished by insulin ther-apy making any role for STZ in initiating mesangial cell prolifer-ation less likely. Several recent studies from our laboratory havedescribed a similar wave of mesangial cell proliferation that

administration [46]. Our study describes two new mechanisms:phenotypic activation and proliferation of the glomerular mesang-ial cell, and a prominent glomerular monocyte/macrophage infil-trate. Both of these events preceded the development of ECMexpansion and sclerosis.

It is generally accepted that the principal glomerular lesion indiabetic nephropathy is mesangial matrix expansion which re-duces the area for filtration and leads eventually to sclerosis andrenal failure [2]. While it seems probable that this process reflectsan abnormality in mesangial cell function leading to over produc-tion or impaired degradation of new or novel matrix components,no direct evidence for mesangial cell dysfunction in diabetes has

Young et al: Cellular events in experimental diabetic nephropathy 939

a)E0c,)

i5 .Q

U)G)

0U)0000.

aw

5

4

3

2

0

Fig. 3. Time course of the glomerular monocytelmacrophage infiltrate indiabetic animals (— —), insulin-treated diabetic animals (— — —) and controlanimals (—). There was a persistent macrophage infiltrate in diabeticanimals that increased over 30 days and was decreased to control levels byinsulin therapy (except for one episode of transient hyperglycemia at daysix). Values are mean SCM. * P < 0.05 vs. control; t P < 0.05 vs. insulin.

940 Young et al: Cellular events in experimental diabetic nephropathy

Fig. 4. Immunohistochemical staining for bFGF at day three in diabetic animals (A), insulin-treated diabetic animal (B), and normal controls (C). Thereis an apparent increase in staining in the diabetic animal in a mesangial distribution which is abolished by insulin therapy (X400).

occurs in both immunologic [53] and non-immunologic [12] formsof glomerular injury, and precedes increased extracellular matrixdeposition and later development of sclerosis. Mesangial cellproliferation appears to be platelet-mediated, and may be initi-ated by bFGF [17, 54], and maintained by PDGF [55] through anautocrine mechanism involving increased mesangial cell produc-tion of PDGF B-chain and PDGF 13-receptor [12, 32]. In diabeticnephropathy we were able to demonstrate that glomerular cellproliferation was reduced by platelet depletion and was accom-panied by some increased staining for bFGF, as well as a modestincrease in glomerular PDGF B-chain staining and mRNA ex-pression. However, the magnitude of these changes in bFGF andPDGF was small, and further studies would be required to assesstheir functional significance. Other stimuli for mesangial cellproliferation in diabetes should be considered.

Of note was the finding that glomerular cell proliferation wasreduced in insulin treated diabetic animals compared to normalcontrols. Insulin treated controls also gained significantly lessweight than normal animals suggesting that nutritional factorsmay have affected cell proliferation. Protein restriction is knownto be reno-protective in diabetic nephropathy [56, 57] and toinhibit expression of cytokines such as TGF-f3 [58]. However,diabetic animals gained even less weight than insulin treatedcontrols which should have diminished rather than increasedcellular events. Thus the cellular events observed in diabeticanimals may have been even more prominent if not modified bynutritional factors.

The finding of platelets early in STZ DN is significant sinceplatelets have been implicated in glomerular injury in humandiseases including DN [59]. Furthermore, use of platelet inhibi-tors in DN decreases urinary protein excretion and thromboxaneB2 levels in humans [60, 61]. Thrombocytopenia induced withanti-platelet serum significantly inhibited proliferation in thismodel. Thus, platelets may be important in mediating the prolif-erative response of mesangial cells in vivo, perhaps throughrelease of growth factors such as PDGF or bFGF [62].

In addition to the mesangial cell changes, our study alsoestablished a significant increase in glomerular macrophageswhich occurred as early as day three and increased progressivelythrough day 30. Like mesangial cell proliferation, the macrophage

Table 4. PDGF B chain, collagen IV and laminin B2 messenger RNAexpression in glomeruli from streptozotocin diabetic ratsa

Day Diabetes Control Insulin

PDGF BChaint 136

1430

1.130.611.111.151.10

1.01.01.01.01.0

—0.721.40C0.920.83

Collagen Type IV" 1

36

1430

1.070.681.081.040.81

1.01.01.01.01.0

—0.761.31C0.980.69

Laminin" 1

36

1430

—0.581.060.840.77

—1.01.01.01.0

—1.020.950.950.58

Expressed as relative densitometry units compared to control animalsassigned a value of 1.0

bAll values are normalized to a 28S ribosomal RNA probe (see text)The increase in PDGF-B chain and collagen IV mRNA noted on day

six in insulin-treated animals may reflect a one time only increase in bloodsugar in this group on day six (243 128 mgldl vs. 150 22 in controls and352 139 in diabetic animals)

infiltrate was also decreased by insulin therapy. Again, this findingis similar to what we have observed in other models of progressivesclerosis involving the mesangium, including the anti-thymocyteserum and remnant kidney models [18, 12]. Glomerular macro-phages have also been prominently identified in a variety of otherprogressive glomes-ular diseases in both animal models [63, 64],and in humans [65], including diabetic nephropathy [66]. Theabsence of a significant lymphocytic or neutrophilic infiltratesuggests that the stimulus for macrophage accumulation mayderive from resident glomerular cells. Possible mesangial cellderived chemotactic factors would include PDGF [9], TGF-/3 [67],advanced glycosylation end-products (AGES) [68], carbohydrates[69], monocyte chemotactic peptide (MCP-1/JE-protein) [70], ormacrophage inhibitory factor [71].

The consequences of this macrophage infiltrate in diabetic

Young et a!: Cellular events in experimental diabetic nephropathy 941

Dayl4 Day3O

TGF-131

28 S

DM CON INS-OM DM CON INS-DM

Fig. 5. Gene expression for TGF-f31 at days 14and 30 in glomeruli from diabetic animals (DM),normal controls (con) and insulin-treated diabetes(INS-DM). Quantitation by densitometsy inrelative units compared to controls that areassigned a value of 100 is shown below. TGF-pmRNA is increased about 20% in diabeticanimals at both time points.

Fig. 6. Immunostaining for type iT/collagen at 30 days in a diabetic animal (A), insulin-treated diabetic animal (B), and normal control (C). There is a slightincrease in mesangial matrix staining for type IV collagen that preceded any evidence of matrix expansion or sclerosis (X400).

nephropathy were not defined. Macrophages may release a variety have been documented in other models of progressive sclerosisof pro-inflammatory substances such as oxidants, proteases, tissue [73]. TGF-f3 is a potent stimulus for the production of extracel-factor and various cytokines and growth factors [reviewed in 72]. lular matrix by glomerular and extra glomerular cells [74—76], andMacrophages are also major effector cells in the mediation of collagen synthesis by glomeruli in STZ diabetes has been sug-some types of acute immune glomerular injury [64]. However, a gested to be TGF-/3 dependent [76]. Moreover, neutralization ofmore likely role for macrophages in the non-inflammatory early TGF-/3 in vivo using specific antibody or neutralizing protogly-phase of diabetic nephropathy is as a source for TGF-f3. TGF-p cans has been reported to diminish glomerular matrix accumula-mRNA was increased in glomeruli by day 14 and persisted for 30 tion and sclerosis in the ATS model of glomerulonephritis [77].days, a finding similar to that recently reported by Border and The source of the increase in glomerular TGF-/3 mRNA has notcolleagues [8]. Similar increases in TGF-/3 mRNA and protein been established in glomerulonephritis or diabetic nephropathy,

140

120

100

D 80Ct 60

40

20

0

Day 30

Diabetic Control INS-DM

Day 14

on ,C °•60

Diabetic Control INS-DM

I I

I''

'N B1

4

raa !t%' __942 Younget a!: Cellular events in experimental diabetic nephropathy

Fig. 7. Electron micrograph demonstrating circulating platelets (arrow) and mononuclear leukocytes (M) within glomerular capillaries in a diabetic animalat day three (X4000).

and both mesangial cells and macrophages may be involved.TGF-13 has been shown to be increased in vascular smooth musclecells in response to glucose [78], as well as in mesangial cellswhere it is also a potent inhibitor of proliferation induced by highglucose [79]. Thus, up-regulation of glomerular TGF-p precedesdeposition of ECM components such as collagen IV, and coin-cides with continued glomerular macrophage influx and decreasedglomerular cell proliferation.

The cellular events described above preceded any detectableincrease in either staining or gene expression for ECM compo-nents. However, some increased mesangial staining for type IVcollagen and laminin was detectable by 30 days, although noincrease in mRNA for these proteins was observed. Othersstudying changes in ECM early in STZ diabetes have reportedconflicting results including a decrease in type IV (151) collagenmRNA with normal laminin Bi mRNA levels [80] and a gener-alized increase in mRNA for collagen I, IV and laminin Bi and B2[81]. These differences probably relate to variations in experimen-tal design and will require additional studies to reconcile.

To summarize and sequence our findings, we have documenteda very early (days 1 to 3) phase of mesangial cell proliferationassociated with de novo expression of cs-actin in STZ-induceddiabetic nephropathy. Mesangial cell proliferation is associatedwith glomerular platelet localization and is blocked by plateletdepletion. These mesangial cell changes are associated with someincrease in glomerular mesangial content of bFGF and PDGF butare probably mediated primarily by other metabolic and/or hor-monal/hemodynamic factors. After mesangial cell proliferation isin progress, there develops a prominent and persistent glomerularmacrophage infiltrate associated with an increase in glomerularmRNA for TGF-f3. All of these cellular events precede any

detectable changes in glomerular gene expression or mesangialaccumulation of normal extracellular matrix components. It there-fore seems likely that these cellular events are important in thelater development of diabetic glomeruloscierosis and may evenserve as early prognostic markers in this disease.

Acknowledgments

Support for this study was provided by research grants from the USPublic Health Service (DK34198, DK07487, DK43422, DK02142 andDK47659) and from the Northwest Kidney Foundation.

Reprint requests to William G. Couser, M.D., Division of Nephrology,RM-11, University of Washington, Seattle, Washington 98195, USA.

References

1. U.S. RENAL DATA SYSTEM: USRDS 1992 Annual Data Report. TheNational Institutes of Health, National Institute of Diabetes andDigestive Kidney Diseases, Bethesda, August 1992

2. MAUER SM: Structural-functional correlations of diabetic nephropa-thy. Kidney mt 45:612—622, 1994

3. HOSTETTER TH, TROY JL, BRENNER BM: Glomerular hemodynamicsin experimental diabetes. Kidney mt 19:410—415, 1981

4. BRENNER BM: Hemodynamically mediated glomerular injury and theprogressive nature of kidney disease. Kidney mt 23:647—655, 1983

5. KIRKAWA R, UMEMURA K, HANEDA M, ARIMURA T, EBATA K,SHIGETA Y: Evidence for the existence of polyol pathway in culturedrat mesangial cells. Diabetes 36:240—243, 1987

6. MANGILI R, BENDING JJ, Scorr 0, LIK L, GUPTA A, VIBERTI G:Increased sodium-lithium countertransport activity in red cells ofpatients with insulin dependent diabetes and nephropathy. N EngI JMed 318:146—150, 1988

7. MAKITA Z, RADOFF S, RAYFIELD RI, YANG Z, SKOLNIK E, DELANEYV, FRIEDMAN EA, CERAMJ A, VLASSARA H: Advanced glycosylation

Young et al. Cellular events in experimental diabetic nephropathy 943

end products in patients with diabetic nephropathy. N Engl J Med325:836—842, 1991

8. YAMAMOTO T, NAKAMURA T, NoBLE NA, RUOSLkHTI E, BORDER W:Expression of transforming growth factor /3 is elevated in human andexperimental diabetic nephropathy. Proc NatlAcad Sci USA 90:1814—1818, 1993

9. ENG E, FLOEGE J, YOUNG BA, COUSER WG, JOHNSON RJ: Doesextracellular matrix expansion in glomerular disease require mesang-ial cell proliferation? Kidney mt 45:S45—S47, 1994

10. JOHNSON RJ, FLOEGE J, YOSHIMURA A, IIDA H, COUSER WG, ALPERSCE: The activated mesangial cell: A glomerular "myofibroblast"? JAm Soc Nephrol 2:S190—S197, 1992

11. FLOEGE J, ENG E, YOUNG BA, JOHNSON RJ: Factors involved in theregulation of mesangial cell proliferation in vitro and in vivo. Kidneymt 43:547—554, 1993

12. FLOEGE J, BURNS MW, ALPERS CE, YOSHIMURA A, PRITZL P.GORDON K, SEIFERT RA, BOWEN-POPE DF, COUSER WG, JOHNSONRJ: Glomerular cell proliferation and PDGF expression precedeglomerulosclerosis in the remnant kidney model. Kidney mt 41:279'-309, 1992

13. Doi T, STRIKER LI, GIBSON CC, AGODOA LY, BRINSTER RL, STRIKERGE: Glomerular lesions in mice transgenic for growth hormone andinsulinlike growth factor-I. I. Relationship between increased glomer-ular size and mesangial sclerosis. Am J Pathol 137:541—552, 1990

14. Doi T, STRIKER U, KIMATA K, PETEN EP, YAMADA Y, STRIKER GE:Glomeruloscierosis in mice transgenic for growth hormone. Increasedmesangial extracellular matrix is correlated with kidney mRNA levels.JExp Med 173:1287—1290, 1991

15. MAcKAY K, STRIKER LI, PINKERT CA, BRINSTER RL, STRIKER GE:Glomerulosclerosis and renal cysts in mice transgenic for the earlyregion of SV4O. Kidney mt 32:827—837, 1987

16. MAcKAY K, STRIKER U, STAUFFER JW, AGODOA LY, STRIKER GE:Relationship of glomerular hypertrophy and sclerosis: Studies in SV4Otransgenic mice. Kidney list 37:741—748, 1990

17. JOHNSON RJ, ALPERS CE, PRITZL P, SCHULZE M, BAKER P, PRUCHNOC, COUSER WG: Platelets mediated neutrophil-dependent immunecomplex nephritis in the rat. J Clin Invest 82:1225—1235, 1988

18. JOHNSON RJ, GARCIA RL, PRITZL P, ALPERS CE: Platelets mediateglomerular cell proliferation in immune complex nephritis in the ratinduced by anti-mesangial cell antibodies. Am J Pathol 136:369—374,1990

19. IIoA H, SEIFERT R, ALPERS CE, GRONWALD RGK, PHILLIPS PE,PRITZL P, GORDON K, GowI'l AM, Ross R, BOWEN-POPE DF,JOHNSON RJ: Platelet-derived growth factor (PDGF) and PDGFreceptor are induced in mesangial proliferative nephritis in the rat.Proc NatlAcad Sci USA 88:6560—6564, 1991

20. KURKI P, VANDERLAAN M, DOLBEARE F, GRAY J, Tr' EM: Expres-sion of proliferating cell nuclear antigen (PCNA)/cyclin during the cellcycle. Exp Cell Res 166:209—219, 1986

21. SKALLI 0, RoPRAz P, TRZECIAK A, BENZONANA G, GILLESSEN D,GABIANI G: A monoclonal antibody against is-smooth muscle actin: Anew probe for smooth muscle differentiation. J Cell Biol 103:2787—2796, 1986

22. BAGCHUS WM, JEUNINK MF, ROZING J, ELEMA JD: A monoclonalantibody against rat platelets. I. Tissue distribution in vitro and in vivo.Clin Exp Immunol 75:317—323, 1989

23. MASON DW, WILLIAMS AF: The kinetics of antibody binding tomembrane antigens in solution and at the cell surface. Biochem J187:1—20, 1980

24. SEKIYA 5, GoToil 5, YAMASHITA T, WATANABE T, SAITOH 5, SENDO F:Selective depletion of rat neutrophils by in vivo administration of amonoclonal antibody. J Leuk Biol 46:96—102, 1989

25. REILLY TM, TAYLOR DS, HERVLIN WF, THOOLON MJ, CFIIU AT,WATSON DW, TIMMERMANS PB: Monoclonal antibodies directedagainst basic fibroblast growth factor which inhibit its biologic activityin vitro and in vivo. Biochem Biophys Res Commun 164:736—743, 1989

26. SHIRAISHI T, M0RIM0TO S, SATO H, SUGIHARA K, ONISHI T, OGIHARAT: Radioimmunoassay of human platelet-derived growth factor usingmonoclonal antibodies toward a synthetic 73-97 fragment of itsB-chain. Clin Chim Acta 184:65—74, 1989

27. COUCFIMAN JR: Heterogeneous distribution of a basement membraneheparan sulfate proteoglycan in rat tissues. J Cell Biol 105:1901—1916,1987

28. FOUSER L, IRUELE-ARISPE L, BORNSTEIN P. SAGE EH: Transcriptionalactivity of the al(I)-collagen promoter is correlated with the forma-tion of capillary-like structures by endothelial cells in vitro. J BiolChem 266:18345—18351, 1991

29. FLOEGE J, ENG E, YOUNG BA, COUSER WU, JOHNSON RJ: Heparinsuppresses mesangial cell proliferation and matrix expansion in ex-perimental mesangioproliferative glomerulonephritis. Kidney mt 43:369—380, 1993

30. CI-IOMCZYNSKI P, SACCHI N: Single-step method of RNA isolation byacid guanidinium thiocyanate-phenol-chloroform extraction. AnalBiochem 162:156—159, 1987

31. JOHNSON RJ, IIDA H, ALPERS CE, MAJESKY MW, SCHWARTZ SM,PRITZL P, GORDON K, GowN AM: Expression of smooth muscle cellphenotype by rat mesangial cells in immune complex nephntis.Alpha-smooth muscle actin is a marker of mesangial cell proliferation.J Clin Invest 87:847—858, 1991

32. YOSHIMURA A, GORDON K, ALFERS CE, FLOEGE J, PRITZL P, Ross R,COUSER WG, BOWEN-POPE DF, JOHNSON RJ: Demonstration ofPDGF B-chain mRNA in glomeruli in mesangial proliferative nephri-tis by in situ hybridization. Kidney mt 40:470—476, 1991

33. Qui SW, KONDALAM P, ROBERTS A, Spoitr' MB: cDNA cloning byPCR of rat transforming growth factor /3-1 (abstract). NuclAcids Res18:3059, 1990

34. SHIMASAKI S, EMOTO N, KOBA A, MERCADO M, SHIBATA F, CooIcsEYK, BAIRD A, LING N: Complementary DNA cloning and sequencing ofrat ovarian basic fibroblast growth factor and tissue distribution studyof its mRNA. Biochem Biophys Res Commun 157:256—263, 1988

35. KURKINEN M, CONDON M, BLUMBERG B, BARLOW D, QUINONES S,SAUS J, PIHLAJANIEMI T: Extensive homology between the carboxyl-terminal peptides of mouse al (IV) and a2 (IV) collagen. J Biol Chem262:8496—8499, 1987

36. SASAKI M, YAMADA Y: The laminin B2 chain has a multidomainstructure homologous to the Bi chain. J Biol Chem 262:17111—17117,1987

37. BRADLEY GM, BENSON ES: Examination of the urine, in Todd-SanfordClinical Diagnosis by Laboratory Methods, edited by DAVIDSON I,HENERY JB, Philadelphia, WB Saunders, 1974, p 74

38. CONOVER WJ: Practical Nonparametric Statistics (2nd ed). New York,John Wiley and Sons, 1980

39. MORGENSEN CE, Smims MW, DECKERT T, CHRISTIANSEN JS: Func-tional and morphological renal manifestations in diabetes mellitus.Diabetologia 21:89—93, 1981

40. ZATZ R, MEYER TW, RENNKE HG, BRENNER BM: Predominance ofhemodynamic rather than metabolic factors in the pathogenesis ofdiabetic glomerulopathy. Proc NatI Acad Sci USA 82:5963—5967, 1985

41. STEFFEs MW, OSTERBY R, CHAVERS B, MAUER SM: Mesangialexpansion as a central mechanism for loss of kidney function indiabetic patients. Diabetes 38:1077—1081, 1989

42. SATRR-HM4SEN K: Renal hypertrophy in experimental diabetes mel-litus. Kidney mt 23:643—646, 1983

43. Sci-swsmR J, FINE LG: Renal hypertrophy, growth factors, andnephropathy in diabetes mellitus. Semin Nephml 10:242—253, 1990

44. TILTON RG, BAIR LD, HARLOW JE, SMITH SR, OsTRow E, WILLIAM-SON JR: Diabetes-induced glomerular dysfunction: Links to a morereduced cytosolic ratio of NADH/NAD+. Kidney list 41:778—788,1992

45. LOPES DE FARIA JB, FRIEDMAN R, TARIQ T, VIBERTI CC: Prevalenceof raised sodium-lithium countersport activity in type 1 diabeticpatients. Kidney mt 41:877—882, 1992

46. ABRAS5 CK, PETERSON CV, RAUGI GJ: Phenotypic expression ofcollagen types in mesangial matrix of diabetic and nondiabetic rats.Diabetes 37:1695—1702, 1988

47. JOHNSON RJ, ALPERS CE, YOSHIMURA A, LOMBARDI D, PRITZL P,FLOEGE J, SCHWARTZ S: Renal injury from angiotensin Il-mediatedhypertension. Hypertension 19:464—474, 1992

48. ALPERS CE, HUDKINS KL, GowN AM, Jom'sSoN RJ: Enhancedexpression of "muscle-specific" actin in glomerulonephritis. Kidney Int41:1134—1142, 1992

49. CORTES P, DUMLER F, G0WMAN J, LEVIN NW: Relationship betweenrenal function and metabolic alterations in early streptozotocin-induced diabetes in rats. Diabetes 36:80—87, 1987

944 Younget a!: Cellular events in experimental diabetic nephropathy

50. CHURCHILL P, CHURCHILL M, DIDANI A, DUNBAR J: Streptozotocin-induced renal hemodynamic changes in isogenic Lewis rats: A kidneytransplant study. Am J Physiol 264:F100—F105, 1993

51. MORITA T, CHURG J: Mesangiolysis. Kidney mt 24:1—9, 198352. NARurt T, FUKUI M, EBIHARA I, OSADA S, NAGAOKA I, Totmto

Y, K0IDE H: mRNA expression of growth factors in glomeruli fromdiabetic rats. Diabetes 42:450—456, 1993

53. FLOEGE J, JOHNSON RJ, GORDON K, JIDA H, PRITZL P, YOSHIMURA A,CAMPBELL C, ALPERS CE, COUSER WG: Increased synthesis ofextracellular matrix in mesangial proliferative nephritis. Kidney mt40:477—488, 1991

54. FLOEGE J, ENG E, LINDNER V, ALPERS CE, YOUNG BA, REIDY MA,JOHNSON RJ: Rat glomerular mesangial cells synthesize basic fibro-blast growth factor. J Clin Invest 90:2362—2369, 1992

55. JOHNSON RJ, RAINES E, FLOEGE J, YOSHIMURA A, PRITZL P, ALPERSCE, Ross R: Inhibition of mesangial cell proliferation and matrixexpansion in glomerulonephritis in the rat by antibody to plateletderived growth factor. J Exp Med 175:1413—1416, 1992

56. WISEMAN MJ, BOGNETFI E, DODDS R, KEEN H, VIBERTI GC: Changesin renal function in response to protein restricted diet in type 1(insulin-dependent) diabetic patients. Diabetologia 30:154—159, 1987

57. DIAMOND JR: Effects of dietary intervention on glomerular patho-physiology. Am J Physiol 258:F1—F8, 1990

58. OKUDA S, NAKAMURA T, YAMAMOTO T, RUOSLAHTI E, BORDER WA:

Dietary protein restriction rapidly redices transforming growth factor131 expression in experimental glomerulonephritis. Proc NatlAcad SciUSA 88:9765—9769, 1991

59. TOTH L, SZENASI P, VAR5ANYL M, AZILVASI I, LEHOCZKY E, KAM-MERER L, R0MIC5 L: Elevated levels of plasma and urine beta-thromboglobulin or thromboxane-b2 as markers of real platelethyperactivation in diabetic nephropathy. Hemostasis 22:334—339, 1992

60. DoN.DlO JV, ILSTRUP DM, HOLLEY KE, ROMERO JC: Platelet-inhibitor treatment of diabetic nephropathy: A 10-year prospectivestudy. Mayo Clinic Proc 63:3—15, 1988

61. DAVI G, CATALONO I, AVERMA M, NOTARBARTOLO A, STRANO A,CIABAITONI G, PATRONO C: Thromboxane biosynthesis and plateletfunction in type II diabetes mellitus. N Engl J Med 322:1769—1774,1990

62. JOHNsON RJ: Platelets in inflammatory glomerular injury. SeminNephrol 11:276—284, 1991

63. SCHREINER GF, COTRAN RS, UNANUE ER: Modulation of Ia andleukocyte common antigen expression in rat glomeruli during thecourse of glomerulonephritis and aminonucleoside nephrosis. LabInvest 51:524—533, 1984

64. SCHREINER GF, COTRAN RS, UNANUE ER: Macrophages and cellularimmunity in experimental glomerulonephritis. Springer Semin Immu-nopathol 5:251—267, 1982

65. BOLTON WK, INNEs DJ, STURGILL BC, KAISER DL: T-cells andmacrophages in rapidly progressive glomerulonephritis: clinicopatho-logical correlations. Kidney mt 32:869—876, 1987

66. FURUTA T, SAITO T, OOTAKA T, Sor J, OBARA K, ABE K, YO5HI-NAGA K: The role of macrophages in diabetic glomerulosclerosis. AmJKidDis 21:480—485, 1993

67. WA}IL SM, HUNT DA, WAKEFEILD LM, MCCARTNEY-FRANCIS N,WAHL LM, ROBERTS AB, SPORN MB: Transforming growth factorbeta (TGF/3) induces monocyte chemotaxis and growth factor pro-duction. Proc Nat! Acad Sci USA 84:5788—5792, 1987

68. SKOLNIK EY, YANG Z, MAKITA Z, RADOFF S, KIRSTEIN M, VIssARAH: Human and rat mesangial cell receptors for glucose-modifiedproteins: Potential role in kidney tissue remodelling and diabeticnephropathy. J Exp Med 174:931—939, 1991

69. STEWART J, WEIR DM: Carbohydrates as recognition molecules inmacrophage activities. J Clin Lab Immunol 28:103—108, 1989

70. ROLLINS BJ, YOSHIMURA T, LEONARD EJ, PROBER SJ: Cytokine-activated human endothelial cells synthesize and secrete a monoche-moattractant, MCP-1/JE. Am J Pathol 136:1229—1233, 1990

71. WEISER WY, GREINEDER DK, REMOLD HG, DAvID JR: Studies onhuman migration inhibitory factor: Characterization of three molec-ular species. Jlmmunol 126:1958—1963, 1981

72. NATHAN CF: Secretory products of macrophages. J Clin Invest 79:3 19—326, 1987

73. OKUDA 5, LANGUINO LR, RUOSLAHTI E, BORDER WA: Elevatedexpression of transforming growth factor-p and proteoglycan produc-tion in experimental glomerulonephritis. Possible role in expansion ofthe expansion of the mesangial extracellular matrix. J Clin Invest86:453—462, 1990

74. MACKAY K, STRIKER U, STAUFFER JW, Dot T, AGODOA LY, STRIKER

GE: Transforming growth factor-/3. Murine glomerular receptors andresponses of isolated glomerular cells. J Clin Invest 83:1160—1167,1989

75. BORDER WA, OKUDA 5, LANGUINO LR, RUOSLAHTI E: Transforming

growth factor-p regulates production of proteoglycans by mesangialcells. Kidney Int 37:689—695, 1990

76. BOLLINENI JS, REDDI AS: Transforming growth factor-131 enhancesglomerular collagen synthesis in diabetic rats. Diabetes 42:1673—1677,1993

77. BORDER WA, NOBLE NA, YAMAMOTO T, HARPER JR, YAMAGUCHI Y,PIERSCHBACHER MD, RuoSIJ1TI E: Natural inhibitor of transforminggrowth factor-13 protects against scarring in experimental kidneydisease. Nature 26:362—364, 1992

78. MCCLAIN DA, PATERSON AJ, Roos MD, WEI X, KUDLOW JE:Glucose and glucosamine regulate growth factor gene expression invascular smooth muscle cells. Proc NatlAcad Sci USA 89:8150—8154,1992

79. WOLF G, SHARMA K, CHEN Y, ERICKSEN M, ZIYADEH FN: Highglucose-induced proliferation in mesangial cells in reversed by auto-crine TGF-j3. Kidney mt 42:647—656, 1992

80. POULSOM R, KURKINEN M, PROCKOP DJ, BOOT-HANFORD RP: In-creased steady-state levels of laminin Bi mRNA in kidneys oflong-term streptozotocin-diabetic rats. J Biol Chem 263:10072—10076,1988

81. FuKUI M, NAKAMURA T, EBIHARA I, SHIRATO I, ToanNo Y, KOIDE H:ECM gene expression and its modulation by insulin in diabetic rats.Diabetes 41:1520—1527, 1992