Kidney International, Vol. 52 (1997), pp. 393—403

Podocytes are the major source of IL-la and IL-113 in human

glomerulonephritidesZOFIA I. NIEMIR, HENNING STEIN, GRZEGORZ DwoRAciu, PETER MUNDEL, NADJA KOEHL, Boius KOCH,FRANK AUTSCHBACH, KONRAD ANDRASSY, EBERHARD RITZ, RUEDIGER WALDHERR, and HERWART F. Orro

Departments of Pathology, Nephrology and Anatomy, Ruperto-Carola University, Heidelberg, Germany, and Department of Pathology,Poznan Medical Schoo4 Poland

Podocytes are the major source of IL-la and IL-lf.3 in human glomer-ulonephritides. To address the question of in situ production of IL-la andIL-l/3 in proliferative and non-proliferative forms of human glomerulo-nephritis (UN), we performed immunocytochemical and in situ hybridiza-tion studies on renal biopsies from patients with mesangial IgA-GN (N38), idiopathic membranous GN (MGN; N 12), minimal change disease(MCD; N = 9), focal segmental glomerulosclerosis (FSGS; N = 5) andacute endocapillary UN (AGN; N = 3). Normal kidneys (N = 10) servedas controls. Concomitantly, the expression of IL-I receptor type I (IL-iRI), IL-I receptor type II (IL-i Ru) and of IL-i receptor antagonist (IL-IRA) was analyzed. Antibodies against antigens expressed on podocytes(PP-44), endothelial cells (CD3I) and monocytes/macrophages (CDilb,CD 14, CD68) were applied to attribute the expression of IL-i/IL-I relatedpeptides to intrinsic glomerular and/or blood-derived infiltrating cells. Ourresults demonstrate that IL-i Rh is constitutively expressed on endothe-hal cells, and its expression can be induced in proximal tubular cells andin the interstitium. In diseased glomeruli podocytes are capable ofproducing IL-la/13. In MGN and MCD/FSGS, the expression of both IL-iforms is particularly noted in early stages of the disease and is not onlyaccompanied by a marked reactivity for IL-i RI, but also for IL-I RA. Insegmental sclerosing lesions in FSGS and in IgA-GN with markedglomerular proliferation and/or sclerosis, a reduced expression of thePP-44 antigen and a diminished ability of podocytes to produce IL-i/IL-irelated peptides are noted. These results suggest that intrinsic glomerularproduction of IL-i may be of relevance for the protection of glomerulifrom continuing injuly.

Among inflammatory cytokines IL-i represents a central medi-ator of inflammation [1]. IL-I has been shown in glomeruli insevere forms of human and animal UN [2—6]. In vitro, mesangialcells stimulated with immune complexes produce this cytokine [7].Recently, it has also been suggested that visceral epithelial cellsare capable of producing IL-lp in glomerular inflammation [5].The production of IL-i in chronic proliferative forms of humanUN, however, has mainly been attributed to infiltrating blood-derived monocytes/macrophages [8—101. On the other hand, anincreased excretion of IL-i in the urine has been described inpatients with idiopathic membranous UN, where proliferation of

glomerular cells is absent or minimal and a significant number ofinfiltrating cells is not observed in glomeruli [ii].

Restriction of the IL-i action in accelerated nephrotoxic ne-phritis in the rat by infusion of recombinant IL-i receptorantagonist (IL-i RA) reduced glomerular injury [12]. ExogenousIL-I RA, however, was without effect in the established murinelupus nephritis and a modulating role of endogenously producedIL-i RA for IL-i expression has been suggested [131. Indeed, acoordinated expression of IL-1/3 and IL-RA genes has beenshown in glomeruli after the induction of antibody-mediated UNin rats [6]. We examined the in situ production of IL-la and 1L-1f3in renal biopsies from patients with various forms of proliferativeand non-proliferative UN by means of immunocytochemistry andin situ hybridization. The expression of peptides blocking IL-ibinding to the IL-i signal IL-i receptor type I (IL-i RI) [14], orthe IL-i receptor type II (IL-i RII) [15, 16] and IL-i RA [17] wasanalyzed at the protein level.

In an attempt to attribute the expression of these peptides tointrinsic glomerular and/or blood-derived infiltrating cells, immu-nocytochemical staining with monoclonal antibodies reacting withspecific glomerular and infiltrating cells' antigens was performed.Recently, two glycoproteins, that is, CD3 1 and PP-44, have beenshown to be exclusively expressed by endothelial and visceralepithelial cells, respectively [18—20]. CD31, platelet-endothelialcell adhesion molecule (PECAM), represents an adhesion mole-cule involved in the formation of intercellular junctions that isconstitutively expressed on endothelial cells [18, 19]. PP-44 hasbeen localized to podocyte foot processes and is assumed to beinvolved in the podocyte contractile apparatus [20]. Concomi-tantly, the expression of cell surface molecules characteristic ofmonocytes/macrophages, particularly those implicated in leuko-cyte-mesangial and endothelial-leukocyte adhesion, was studied.

Our results clearly demonstrate that in glomerular diseaseIL-la and IL-f3 are produced in glomeruli and their production istightly regulated.

METHODS

Patients

Sixty-seven patients were enrolled in this study: 38 with mesan-gial IgA-glomerulonephritis (IgA-UN; 10 females, 28 males,median age 37.6, range 8 to 71 years), 9 with minimal changedisease (MCD; 4 females, 5 males, median age 19.2, range 5 to 47

393

Key words: interleukin, podocytes, glomerulonephritis, lesion, sclerosis.

Received for publication December 3, 1996and in revised form March 5, 1997Accepted for publication March 6, 1997

© 1997 by the International Society of Nephrology

394 Niemir et al: IL-I and glomerular protection

Table 1. Histological and clinical data of patients at the time of renal biopsy

Glomerularobsolescence!

Crescents sclerosisAge

No. years/sex %

Duration ofsymptoms

years

Serumcreatinine

mg!dl

Urinaryproteing/24 hr

Urinaryerythrocytes

E/mm3

IGA-GNMGA

1 15/Ma 4/12 0.9 0.5 30,0002 21/M 2 0.9 0.3 gross hematuria3 27/Me 6/12 1.2 0.3 317,0004 20!M 3/12 0.9 0.4 gross hematuria

MesPGN5 49/Ms — — unknown 1.0 2.4 50,0006 43/F — — unknown 0.8 NS 15,0007 36/F 11 22 2 1.1 0.5 5,0008 19/M — — unknown 0.8 0.9 gross hematuria9 20/Ma — — 1.5 0.9 3.0 30,000

10 71/M — — unknown 2.1 1.0 735,00011 52/F 5 5 1.5 0.9 1.0 gross hematuria12 11/M — — 1.5 0.8 2.0 22,00013 44/F — — 2 0.8 0.5 gross hematuria14 54/F — — 10 1.7 1.2 gross hematuria

CrGN15 71/M 33 33 unknown 4.4 NS 30,00016 44!M 50 — 2/12 1.7 NS gross hematuria17 47/F 55 11 unknown 1.0 NS 10,00018 36!M 33 — 1 1.0 3.0 gross hematuria19 25!M 25 25 2 1.2 2.0 gross hematuria20 12!M 50 50 2 1.0 NS gross hematuria21 8/M 40 — 1 0.7 NS gross hematuria22 40/Ms 35 14 1 0.9 2.0 30,00023 33/Ms 36 18 1 1.0 3.5 20,000

ScGN24 32/Ma — 88 unknown 1.8 2.7 100,00025 49/Me — 50 15 2.0 4.1 gross hematuria26 40/F 11 44 unknown 2.2 2.3 113,00027 28/Fe — 44 1 1.5 NS 100,00028 58/Ma 21 50 unknown 1.8 2.2 40,00029 47/M — 60 1 1.6 3.0 100,00030 57/M 20 50 unknown 9.0 2.5 gross hematuria31 51/M 20 45 unknown 2.6 2.0 100,00032 37/M 7 77 unknown 2.2 NS 260,00033 56/M — 50 2 2.2 1.7 gross hematuria34 68/Ms — 62 unknown 1.9 1.8 25,00035 40/F — 43 unknown 1.8 4.0 gross hematuria36 41/Fe — 40 20 1.0 3.5 20,00037 21/Ms — 30 3 0.9 2.1 30,00038 35/Jyfa — 30 7 1.0 1.5 50,000

MCD39 6/Mb — 1 0.5 NS 10,00040 10/F1' — 2 0.9 NS —41 32/F — 1 0.8 NS 10,00042 25/F — 1/12 0.6 NS 15,00043 47/Me — 2 0.9 NS 10,00044 71ph 3/12 0.9 NS —45 31/Mb 5 0.9 NS —46 5/Mb — 2 0.6 NS —47 10/Wb 5 3/12 0.5 NS —

FSGS48 9/Mb 15 (segm.) 8 0.9 NS 10,00049 35/M 13 (segm.) 4 0.9 NS —50 42/F 100 (scgm.) unknown 1.9 3.0 —51 33/MU 77(segm.) 6/12 1.0 2.0 15,00052 12/MU 89 (segm.) unknown 1.6 1.5 100,000

AGN53 33/F — 2 weeks 2.4 0.5 gross hematuria54 71/M 25 3 weeks 3.4 NS gross hematuria55 22/MU — 1/12 1.0 5,000

Niemir et at: IL-I and glomerular protection 395

Table 1. Continued

Glomerular

No.Age

years/sex

Crescentsobsolescence!

sclerosis.

Duration ofsymptoms

years

Serumcreatinine

mg!dl

.Urinaryproteing/24 hr

.

Urinaryerythrocytes

E/mm3%

IMGN

56 7/MStageI/I! — 4/12 0.6 NS 10,000

57 61/M II 12 3/12 1.0 NS 10,00058 55/Ma II — 4/12 1.6 NS 100,00059 40/Ms H — 3/12 0.9 NS60 64/M lI/Ill 40 2 1.0 NS —61 71/F" Il/Ill 33 7/12 1.6 NS 15,00062 53/MC Il/Ill — 6/12 0.9 2.8 10,00063 40/F Il/Ill — 1 0.7 NS 15,00064 53/M Il/Ill — 6/12 0.9 2.8 —65 77/M III 25 unknown 1.7 3.5 —66 35/M III 83 3 9.0 NS 20,00067 26/M III — 1.5 0.8 NS 10,000

Abbreviations are: MGA, minimal glomerular abnormalities; MesPGN, mesangial-proliferative GN; CrGN, mesangial-proliferative GN withsegmental crescents; ScGN, sclerosing GN; MCD, minimal change disease; FSGS, focal-segmental glomerulosclerosis; AGN, acute endocapillary GN;IMGN, idiopathic membranous GN; NS, nephrotic syndrome.

a Patients/biopsies in whom in situ hybridization studies were performedb

Steroid-dependent NSC Steroid and cyclophosphamide resistant NS"Steroid-resistant NS

years), 5 with focal-segmental glomeruloscierosis (FSGS; I fe-male, 4 males, median age 26.5, range 9 to 42 years), 3 with acuteendocapillary GN (AGN; 1 female, 2 males, median age 42, range22 to 71 years) and 12 patients with idiopathic membranous GN(IMGN; 2 females, 10 males, median age 56.2, range 7 to 71years).

The diagnosis was based on conventional light and immunoflu-orescence microscopy. Morphologically, patients with IgA-GNpresented with different degrees of mesangial expansion/prolifer-ation and/or sclerosis with variable severity of tubulointerstitiallesions. They were categorized as: minimal glomerular abnormal-ities (MGA; N = 4), mesangial-proliferative GN (MesPGN; N =10), mesangial-proliferative GN with segmental crescents (CrGN;N = 9), and sclerosing GN (ScGN; N = 15) [21]. Histologicalclassification of IMGN was made according to the extent ofglomerular membrane alterations revealed by silver staining.

Systemic diseases were excluded by detailed clinical history,examination and laboratory tests. Morphological, clinical andlaboratory data of the patients are summarized in Table 1.Clinically, all patients exhibited a variable degree of proteinuriaand erythrocyturia, and normal or impaired renal function. Wherenephrotic range proteinuria (above 3.5 g/1.72 m2/24 hr) wasaccompanied by hypoalbuminemia, dysproteinemia and lipid ab-normalities, the designation of nephrotic syndrome (NS) wasused. Identified are patients biopsied for a revision of thediagnosis due to ineffectiveness of the administered steroid and/orsteroid-cytotoxic treatment (Table 1).

In patients with AGN, a detailed clinical history revealed anupper respiratory tract infection preceding two weeks to onemonth the occurrence of urinary abnormalities. In two patients ofthis group, the biopsy was performed during an early acute phaseof GN associated with renal function impairment. In the thirdpatient proteinuria was discovered accidentally, and the biopsywas prescribed due to the presence of an upper respiratory tract

infection in his anamnesis and negative urine tests two monthsprior to the biopsy.

Renal biopsies

Tissue preparations for conventional immunofluorescence mi-croscopy, immunocytochemistiy and in situ hybridization wereperformed as described elsewhere [21]. The immunocytochemicalstudies were conducted on renal sections from all patients. Tissuefor in situ hybridization was available from 30 patients. Normalappearing kidney tissue from patients undergoing tumor nephrec-tomies (N = 10) and from kidneys refused for transplantation(due to arterial connection problems; N = 2) served as controlsfor both methods employed.

Immunocytochemistry

The following primary antibodies were applied in this study: amurine monoclonal antibody to CD31 (0672; Immunotech, Mar-seille, France), an anti-PP-44 monoclonal antibody (kindly pro-vided by Dr. P. Mundel, Ruperto-Carola University, Heidelberg,Germany) [20]), a murine anti-IL-la antibody (0795; Immuno-tech), an anti-IL-113 monoclonal antibody (M 400; Immunotech),a monoclonal mouse anti-IL-i receptor type I (CDwl2la) anti-body (1592-01; Genzyme, Cambridge, MA, USA), an anti-IL-ireceptor type II mouse monoclonal antibody (1994-01; Genzyme),a polyclonal anti-IL-i receptor antagonist antibody raised in goat(BDA 29; R&D Systems, Minneapolis, MN, USA), amonoclonalmouse anti-human monocyte CD14 (M 825, TUK4; Dako,Glostrup, Denmark), a monoclonal antibody against CD lib (M741; Dako) and a murine monoclonal antibody to CD68 (M 718,EMB11; Dako). Acetone-fixed cryostat sections were stained withall antibodies. The immunostaining procedure was performed on

- r

e--

• It

—

TW

4

set

S

•,E

. -

,s.f•

.$

r%__

$ %

"

S —

c'

;.

•1'

174

a C

4

'—S

e.,

C

—?.

a.' a

/

¼

:s

.

'-'p.

S

-I —

,

a a

I P

I.

t I.

• t

ED

e

• ¼

t -

II I I

V t

.1'

396 Niemiret al: IL-i and glomerular protection

Niemir et al: IL-i and glomerular protection 397

Fig. 1. The immunoreactivity for PP-44 in a normal glomerulus. The reaction is confined to the outer aspects of glomerular capillaries (X250).

Fig. 2. Biopsy from a patient with crescentic IgA-GN. Note the segmental loss of the PP-44 immunoreactivity on podocytes (X250).

Fig. 3. IL-i R II immunoreactivity in a normal glomerulus (control). A positive reaction is confined to endothelial cells of glomeruli, interstitialcapillaries and the afferent arteriole (X250).

Fig. 4. In situ hybridization using a digoxigenin-labeled IL-la antisense riboprobe in a normal glomerulus. IL-la transcripts are present in cellslocalized in glomerular tufts and in capsular epithelial cells. Single periglomerular cells are also positive for IL-la (X250).

Fig. 5. IL-lp immunoreactivity in a biopsy specimen from a patient with IgA-GN and mild glomerular abnormalities. A positive reaction is mainlyconfined to podocytes, outside glomerular capillaries. Tubular proximal epithelial cells are also positive (X250).

Fig. 6. In situ hybridization using an IL-i13 antisense riboprobe in a biopsy from a patient with IgA-GN and mild mesangial proliferation. Transcriptsfor IL-113 are primarily localized outside mesangial areas, and furthermore, in tubular epithelial cells and some interstitial cells (X250).

Fig. 7. IL.i13 in membranous GN (stage 1111). A variable immunoreactivity is observed in two neighboring glomeruli. Positivity is limited to podocytes(x250).Fig. 8. IL-la immunoreactivity in a biopsy from a patient with membranous GN stage III and segmental sclerosis. Remaining podocytes disclose apositive reaction (x250).

acetone-fixed cryostat sections using the APAAP (alkaline phos-phatase anti-alkaline phosphatase) method as described previ-ously [211. The only modification concerned the overnight incu-bation of sections with the primary antibodies in the presence of0.05% Tween-20 at —4°C.

Control experiments were conducted by omitting the incuba-tion with the primary antibody and by substituting for the primaryantibody a non-immune murine or goat serum. In the case ofIL-113, an additional set of experiments after preincubation of theantibody with a human recombinant IL-1f3 (G551A; PromegaBiotech, Madison, WI, USA) was performed. At the workingdilution of the antibody (1:200), the addition of 100 nglml ofrhIL-1f3 resulted in a negative staining even in biopsies with thepredominant immunoreactivity for IL-113.

In situ hybridization

Probe preparation. To generate riboprobes, respective cDNAfragments were subcloned into pGEM (Promega Biotech, Madi-son, WI, USA) and Bluescript II SK (Stratagene Ltd., La Jolla,CA, USA) transcription vectors, A 460 bp HindIII-Avr II frag-ment of the IL-113/YEpsecl plasmid (67024; ATCC, Rockville,MD, USA) [221 was subcloned into HindIII-XbaI sites nf pGEM-3z. A 396 bp fragment of coding frame for IL-la was subclonedbetween the BamHI and EcoR I sites of BS II SK(+) [23]. Thetemplates were linearized with the appropriate restriction en-zymes. Labeled antisense and sense cRNA probes were tran-scribed in vitro using SP-6, T7 or T3 polymerases and digoxigenin-labeled uridine-triphosphate (DIG-UTP) as substrate accordingto the manufacturer's instructions (DIG RNA labeling kit; Boe-hringer-Mannheim, Mannheim, Germany) [24].

In situ hybridization. The in situ hybridization studies wereperformed as described previously [21] with some additionalsteps. After acetylation, slides were rinsed in PBS (phosphatebuffered saline), immersed in 70% formamide/l >< SSC (sodiumsaline citrate, 1 x SSC = 0.15 M NaC1 + 0.015 M sodium citrate,pH 7.0), incubated at 75°C for five minutes, washed two times inPBS and dehydrated in graded alcohol. The hybridization mixturecontaining the respective cRNA probes was denatured at 65°C for10 minutes. Immediately after application of this mixture onto theslides, overnight hybridization in a humid chamber was per-formed. Following hybridization, posthybridization washes in 1 xSSC at the hybridization temperature and in 1 X SSC at roomtemperature were done. Subsequent colorimetric detection, using

the DIG-Nonradioactive Nucleic Acid Detection Kit, was per-formed as described previously [21].

Negative controls consisted of matched serial sections hybrid-ized with sense probes and sections hybridized with unlabeledantisense probes.

Evaluation of IL-i/IL-i related peptide expression

The expression of IL-la/ and IL-i related peptides in renalbiopsies at the protein versus mRNA level was scored by compar-ing the localization and intensity of staining and/or number ofpositive cells with their expression in normal renal tissue. Theexpression of examined proteins/mRNAs was graded from 0 to 3points according to the following scale: 0, no immunoreactivity/nopositive signals; 1, faint immunoreactivity/single positive cells; 2,scattered moderately intense reactivity/numerous positive cells; 3,dense intense immunoreactivity/clusters of positive cells.

RESULTS

Normal kidney

No differences were detected between the expression of exam-ined peptides in normal appearing renal tissue from tumornephrectomies and donor kidneys. Thus, the designation "normalkidney" comprises both sources.

The expression of PP-44 was exclusively confined to podocytes(Fig. i). Endothelial cells of glomerular and interstitial capillarieswere positive for CD31.

A weak immunoreactivity for IL-1a/ was observed in glomerulartufts and cells scattered throughout the interstitium. By in situhybridization, transcripts for IL-la//3 (Fig. 4) were detected in cellslocalized to the glomerular tufts, in parietal epithelial cells, inepithelial cells of distal tubules and in some cells in the interstitium.

The pattern of positivity for IL-i R II was identical to theexpression of CD31 (Fig. 3). A weak staining was additionallyobserved in proximal tubular epithelial cells.

Only single cells scattered throughout the interstitium discloseda positive reaction for IL-i R type I.

Single cells positive for IL-i RA were seen in glomeruli, inperivascular sites of interstitial vessels and in epithelial cells of distaltubules.

CDiib, CD14 and CD68 positive cells (monocytes/macro-phages) were only sporadically observed in glomerular tufts and inthe interstitium.

'0

S f.

398 Niemiret al: IL-i and glomerular protection

Fig. 9. Biopsy from a patient with membranous GN (stage II). An intensereactivity for IL-i RA is confined to the outer sites of glomerularcapillaries (x250).

LgA-GN

Compared to normal kidneys the most striking finding at theprotein level was a strong immunoreactivity for IL-1f3, confined toouter aspects of glomerular capillaries (corresponding to theimmunoreactivity for PP-44), in biopsies with only minor glomer-ular abnormalities (Fig. 5). Marked mesangial proliferation inMesPGN and particularly in CrGN was accompanied by a re-duced or even loss of PP-44 staining (Fig. 2) and a negligiblereaction for IL-113. In biopsies from patients with ScGN theexpression of IL-1/3 and PP-44 varied from glomerulus to glomer-ulus. Only a trace reactivity for both peptides was observed insclerotic glomeruli, whereas in the remaining non-sclerotic gb-meruli, the expression of IL-1f3 and PP-44 decreased dependenton the extent of glomerular proliferative lesions.

In MGA, single cells in the interstitium were positive for IL-1f3.In MesPGN and CrGN, positive cells were seen in clusters inperiglomerular infiltrates close to glomeruli with adhesions andcrescents. In MGA and MesPGN, staining for IL-i /3 was noted atthe apical membranes of proximal tubules (Fig. 5). Atrophictubules showed reduced staining or were completely negative.

An increased number of transcripts for IL-1f3 was noted inglomeruli with mild mesangial proliferation (Fig. 6). In theinterstitium, IL-1/3 mRNA localized to areas of periglomerularand interstitial mononuclear infiltration. Numerous transcripts forIL-1f3 were found in epithelial cells of distal tubules irrespective ofthe stage of the disease.

In contrast to IL-113, IL-la was nearly exclusively expressed inglomerular cells and hardly noted in infiltrating cells in theinterstitium. The glomerular expression of IL-la corresponded tothat of IL-1f3. By in situ hybridization the expression of mRNA forIL-la was similar to that observed for IL-113. Scans visualize thedifference between the number and intensity of signals in thenormal and diseased glomerulus (Fig. 11).

In glomenili, the immunoreactivity for IL-i R type II wasidentical to that of CD31, and loss of positivity was related to thedegree of mesangial proliferation. Interstitial infiltrates discloseda positive reaction for IL-i R II. In MGA and MesPGN, basalsites of tubular epithelial cells were positive.

A weak expression of IL-i R type I was observed in gbomerulartufts in MGA and MesPGN. A faint interstitial staining wasdetected when pronounced interstitial infiltration was present.

Fig. 10. Same biopsy as in Figure 9. Podocytes disclose a positivereaction for IL-i R I (X250).

A positive glomerular reaction for IL-RA was noted in MGA,MesPGN and to a lesser extent in CrGN. Infiltrates in the intersti-tium were positive. In MGA and MesPGN, a diffuse staining for IL-iRA was noted at the luminal sites of proximal tubular epithelial cellswhereas positivity varied in distal tubular cells.

Single cells positive for CD11b, CD14 and CD68 were found inglomerular tufts in MGA and MesPGN. In MesPGN such cellswere often seen in small adhesions. In CrGN, cells positive forCD1 lb were more commonly observed in glomerular tufts. Thesecells were localized to glomeruli with pronounced mesangialproliferation and seemed to be attached to podocytes surroundingareas of proliferating mesangial cells. Monocytes/macrophageswere only occasionally detected in sclerotic glomeruli.

Some monocytes/macrophages were observed in the intersti-tium in MGA. In MesPGN, focal infiltrates were additionallynoted in the vicinity of glomerular adhesions. This finding wasfrequently observed in patients with CrGN and ScGN.

IMGN

The intensity of IL-i expression was related to the stage of thedisease. In biopsies with stage I and II a mild to pronouncedexpression of IL-113 was observed in glomeruli from the sametissue sample (Fig. 7). A scan of the last picture confirmed that theouter aspects of glomerular capillaries were positive (Fig. 12; thesame analysis of PP-44 staining in the normal kidney is given forcomparison). In stages I/Il of IMGN, all gbomeruli exhibited asignificant reactivity for IL-ia, IL-i R type I (Fig. 10) and IL-iRA (Fig. 9), corresponding to the pattern of podocyte staining.Scanning exposed a much stronger reactivity for IL-i RA thanIL-i RI when the intensity of these reactions in the same tissue(serial sections) was compared (Fig. 13). In biopsies from patientswith stage III disease, the immunoreactivity for PP-44 diminishedand concomitantly, the expression of IL-la//3 (Fig. 8), IL-i RI andIL-i RA was also reduced.

The pattern of IL-i R II expression resembled that for CD31.In proximal tubules, in addition to the positivity noted at theluminal sites, basal sites of proximal epithelial cells exhibited apositive reaction for IL-i R II.

In patients with IMGN stage Il/Ill infiltrates scattered through-out the interstitium showed a positive reaction for IL-i/3 andconcomitantly, for IL-i RA and IL-I R II.

p

Ar

SI.

a. —'t t,se/ar _'... J.e.s i_.

• '4 9_.a.

,.• 3k' •'..S • •

Fig. 12. Reactions for PP-44 and IL-Ifi are confined to the outer aspects of glomerular capillaries (scans of Figs. land 7).

II

r

4' S. * ..4 .s e' s'.5S

• '7• . .,aSe — 'I'.

.P ,'a.

I .. ;. _.• —

Ii. Signals For IL-la are detected in a normal glomenilus. The number of transcripts for IL-1J3 (and for IL-Icr) increases significantly in anof glomerular disease (scans of Figs. 4 and 6).

Niemir et at: IL-I and glomerular protection 399

Fig. 11. Signals for IL-la are detected in a normal glomerulus. The number of transcripts for IL-113 (and for IL-la) increases significantly in an earlystage of glomerular disease (scans of Figs. 4 and 6).

Fig. 12. Reactions for PP-44 and IL.lp are confined to the outer aspects of glomerular capillaries (scans of Figs. 1 and 7).

Fig. 13. The reactivity for IL-i RA significantly exceeds that for IL-i R I (serial sections from the same tissue; scans of Figs. 9 and 10).

MCD FSGS

The expression of all peptides examined was comparable to that Biopsies with FSGS were heterogeneous with respect to theobserved in IMGN stage I/Il. Glomeruli stained moderately for extent of sclerotic glomerular changes and tubulointerstitial alter-IL-la/f3, IL-i R I, IL-i RA (podocytes) and for IL-i R II. ations. The expression of IL-la/f3, IL-i R I and IL-i RA in

400 Niemir et al: IL-i and glomerular protection

non-sclerotic glomeruli was similar to that observed in MCD. Inareas of segmental sclerosis, marked loss of PP-44 and CD31 wasnoted and accompanied by only trace expression of IL-113 andIL-i R II. Interstitial and proximal tubular epithelial staining forIL-i R II and IL-i RA was strong in biopsies/areas with mild tomoderate alterations, whereas only faint positivity was observed inareas with purely fibrotic interstitial changes. Interstitial infiltratesdisclosed a positive reaction for IL-1a/, IL-i RA and IL-i R II.

AGN

The expression of IL-i, IL-i R I and IL-i RA was largelydependent on the degree of the proliferative response in glomer-uli. In biopsies performed during an early stage of the disease, areduced staining for IL-1/3 and PP.44 on glomerular podocyteswas observed. In a patient subjected to the biopsy later in thecourse of the disease an increase in the positivity for PP-44 alongwith that for IL-1f3 was noted. The expression of IL-la, IL-i R Iand IL-i RA was very weak in the early stage of the disease andincreased slightly in a biopsy from a patient with a resolving stageof AGN. A strong expression of IL-i R II was found in interstitial

Fig. 14. Semiquantitative evaluation of theglomerular expression of IL-1o/I3 and IL-irelated peptides in all tissue specimensexamined. The score relates to the intensity ofimmunocytochemical staining and/or thenumber of positive cells. Asterisk (*) denotesthe expression of mRNAs, whereas (0)identifies the respective proteins. Abbreviationsare: MGA, minimal glomerular abnormalities;MesPGN, mesangial proliferative GN; CrGN,mesangial proliferative ON with segmental

NK crescents; MCD, minimal change disease; AGN,acute endocapillary GN; IMGN, idiopathicmembranous GN; NK, normal kidney.

areas. Interstitial infiltrates stained for IL-113, IL-i R II and IL-iRA.

An increase in cells of monocyte/macrophage lineage wasobserved in glomerular tufts in the early stage of the disease. Inthe biopsy performed later, the number of cells positive formonocyte/macrophage markers was comparable to the normalkidney. These cells were relatively frequently noted in the inter-stitium.

A semiquantitative evaluation of IL-la/13 and IL-i relatedpeptide expression in renal biopsies of all patients examined incomparison to the expression noted in the normal kidney (accord-ing to the grading system given in the Methods section) ispresented in Figure 14.

DISCUSSION

Our results demonstrating a constitutive expression of IL-i R IIon renal endothelial cells suggest an involvement of the kidney inthe protection against immunoreactive IL-i that may be presentin the circulation under any inflammatory conditions. IL-i recep-tor type II expressed on the cell membrane has been shown to

0000

88888

0

* **000 000*00000

000

*000

3.2

1

0

3

-J

0

3

2

0

3

=,-. 2O

0

0

.

.

,

***0000

* * * * *

88888

MM

8880*000

**000

*00000

000

*0

00*****

88888

0

0000 88880 000 000

88888 8888° 00000 000 000 00 88888

000088880 0000 000 88888

88888 000 00000

8888° 000

.§ 00

'MGAMesPGNCrGNScGN MCD

IgAGNFSGS I II

LIMGNJIII AGN

Niemir et a!: IL-i and glomen4lar protection 401

represent a "decoy" receptor that particularly binds IL-113 withouttransducing a signal [15]. IL-lp represents an extracellular work-ing form of IL-i [1, 25]. IL-1f3 produced as a 31 kDa peptiderequires activation by an intracellular cysteine protease processingthis peptide into an active 17 kDa molecule [25]. After intrave-nous administration of radiolabeled IL-i/3, a rapid clearance ofthis cytokine from the circulation is observed [261. Kidney, smallintestine and liver have been found to accumulate the highestcontent of radioactivity [26]. The expression of IL-i R II onendothelial cells in the kidney allows the assumption that thismolecule is also present on endothelial cells in other organs. Thus,binding to the IL-i type II R may represent a physiologicalmechanism of IL-113 inactivation and host protection.

Intriguingly, in phases of glomerular injury glomerular cellsthemselves are capable of producing IL-la and IL-if3. Thepattern of immunoreactivity, identical with that of PP-44, pointsto visceral epithelial cells as the main source of these cytokines.The visualization of mRNA for IL-1a/ using in situ hybridizationconfirmed their production in situ in the glomerulus. Although insitu hybridization does not allow characterization of positive cells,the localization and number of these cells (largely exceeding thatof cells of monocyte/macrophage lineage) permit the assumptionof intrinsic glomerular cell capability of producing IL-ia andIL-113. The most pronounced expression of these cytokines isassociated with a generally well preserved glomerular structureand only minor interstitial alterations present that point to theearly stages of glomerular diseases.

IL-la and IL-i/3 are devoid of signal propeptide allowing acontrol of protein secretion [1, 23]. Regarding IL-lp, it has beenreported that it can only be released from cells under stressconditions [27] or cells undergoing apoptosis [28]. In the light ofthese data, podocyte staining for IL-1f3 might point to visceralepithelial cell injury.

In IMGN stage I/Il and MCD an increased podocyte stainingfor IL-1/3 is accompanied by an increased immunoreactivity forIL-la, IL-i R type I and IL-i RA. IL-ia acts mainly intracellu-larly and/or as a membrane-bound form [1, 29]. In contrast toIL-Ip, IL-la does not require previous processing to the 17 kDamolecule to be bound to IL-i R I for the signal transduction [301.The presence of both forms of IL-i along with an increasedexpression of IL-i R I in podocytes could identit' an autocrinestimulation of IL-i production. On the other hand, however, aconcomitant expression of IL-i RA, which represents a restrictivefactor for the IL-i action [17], is noted in podocytes in thesebiopsies. IL-i RA functions as a soluble molecule released fromcells and binding IL-ia/13 [17] or as an intracellular form thatinhibits IL-lp mediated responses [31, 32]. Both forms of IL-i RAhave been shown to be produced in epithelial cell lines [31, 32],and they are indistinguishable by means of immunocytochemistry[31]. In IMGN and MCD, an intense staining of podocytes forIL-i RA, in addition to a diffuse positivity in the lumen ofproximal tubular epithelial cells, was observed. Therefore, theproduction of both forms of IL-i RA in podocytes can besupposed.

Recent results in mice overexpressing or lacking IL-i RA genedemonstrate that endogenously produced IL-i RA was critical forsurvival of Salmonella typhimurium-LPS induced endotoxemia,but it impaired the host response to the infection with intracellu-lar bacteria Listeria monocytogenes [33]. Intriguingly, it has been

found that IL-i RA acts as a positive regulator of IL-i in serumduring endotoxemia [33]. In the light of these results, podocyteexpression of IL-i RA may account for the modulation of IL-iaction, whereas binding of IL-i released from cells by a solubleform of IL-i RA will limit a potentially toxic influence of IL-i onneighboring cells.

In the latter context, concomitantly with the pronounced ex-pression of IL-la/f3 in glomeruli in all patients with no or onlyminor interstitial alterations, an immunoreactivity for IL-i R II inproximal epithelial cells and interstitial cells was observed. Anincrease in the IL-i R II reactivity was noted at the luminal andbasal sites of proximal tubular epithelial cells. The latter obser-vation may point to in situ production of IL-i R II in proximalepithelial cells, whereas positivity noted in the lumen of tubulesindicates a reabsorption of a soluble form of the receptor [16].According to recent results, this particular molecule may be shedfrom cells and can bind, as a soluble receptor, an active form ofIL-i/3 by blocking its access not only to the type I R, but also to theIL-i13 precursor, thus preventing processing of IL-i/3 into itsmature form [16]. Moreover, in a soluble form it does notinterfere with inhibition of IL-1f3 action mediated by IL-i RA[16].

The association of IL-1a/ expression with the presence ofpeptides preventing IL-i action indicates the involvement of thesecytokines in mechanisms that allow the withstanding of injury.Results demonstrating that both forms of IL-i can induce mem-brane component production [34, 351 support this hypothesis.IL-113 has been shown to induce the production of laminin B2chain in cultured visceral epithelial cells [34]. Laminin is a majornon-collageneous component of the normal GBM that is involvedin cellular binding [36, 37]. IL-ia is a potent stimulator of collagentype IV production [35], which is the basal form of collagen in thenormal GBM. Since IL-la represents mainly an intracellularworking molecule of IL-i [29], its increased expression mayparticularly reflect mobilization of metabolic processes in podo-cytes to limit glomerular basement membrane injury.

Interestingly, in IMGN stage I/Il and MCD, in parallel with anincreased expression of IL-ia, IL-i R I and IL-i RA, an increasedimmunoreactivity for PP-44, a structural component of podocytecontractile apparatus [20], is observed when compared to normalcontrol kidneys. The expression of PP-44, IL-I and IL-i relatedpeptides decreases along with the disruption of glomerular archi-tecture observed in advanced stages of glomerular disease. Studiesin experimental glomerular diseases indicate a crucial role ofpodocytes for the preservation of glomerular architecture andfunction [38, 39]. Podocytes are postmitotic cells with only alimited potential for cell division [38, 40]. The inability to replacediseased podocytes is thought to be decisive for the occurrence ofGBM denudation leading subsequently to synechia formationinitiating sclerotic alterations [38, 39]. Although further studiesaddressing this issue are required, it seems conceivable that IL-Imediated mobilization of podocyte metabolic processes is aimedat preservation of podocyte contractile functions to maintain theappropriate capillary wall tension [411.

In conclusion, our results demonstrate that IL-la and IL-i/3 areproduced by podocytes in glomerular inflammation, in contrast toprevious results suggesting predominantly harmful effects of thesepeptides, we propose that IL-ia/f3 also represent modulators ofregulatory processes preventing lethal injury to podocytes.

402 Niemiret al: IL-I and glomendar protection

ACKNOWLEDGMENTS

This study was supported by grants from the Deutsche Forschungsge-meinschaft (WA 698/2-1), the German-Israeli Foundation for ScientificResearch & Development and the Transplantationsschwerpunkt Heidel-berg. We are grateful to Mrs. Sonja Steidel, Mrs. Angelika Mantar, Mrs.Jutta Scheuerer and Mrs. Malgorzata Majewska for their technicalassistance.

Reprint requests to Zofia I. Niemir, M.D., Department of Nephrology,Poznan School of Medicine, Al. Przybyszewskiego 49, 60-355 Poznan,Poland.

REFERENCES

1. DINARELLO CA: Biologic basis for IL-i in disease. Blood 87:2095—2147, 1996

2, N0RONHA IL, KRUGER C, ANDRASSY K, RITZ E, WALDHERR R: In situexpression of TNF-cx, IL-I /3 and IL-2R in ANCA-positive glomerulo-nephritis. Kidney mt 43:682—692, 1993

3. BOSWELL JM, Yui MA, BURT DW, KELLEY VE: Increased tumornecrosis factor and IL-1/3 gene expression in the kidney of mice withlupus nephritis. J Immunol 141:3050—3054, 1988

4. WERBER HI, EMANCIPATOR SN, TYKOCINSKI ML, SEnOR SJ: Theinterleukin 1 gene is expressed by rat glomerular mesangial cells andis augmented in immune complex glomerulonephritis. J Immunol138:3207—3212, 1987

5. COERS W, BROUWER E, Vos JTWM, CHAND A, HUITEMA S, HEER-INGA P, KALLENBERG CGM, WEENING JJ: Podocyte expression ofMHC class I and II and intercellular adhesion molecule-i (ICAM-1)in experimental pauci-immune crescentic glomerulonephritis. GunExp Immunol 98:279—286, 1994

6. TAM FWK, SMITH J, CASHMAN SJ, WANG Y, THOMPSON EM, REES AJ:Glomerular expression of interleukin-1 receptor antagonist and inter-leukin-1J3 genes in antibody-mediated glomerulonephritis. Am JPathol 145:126—136, 1994

7. CHEN A, CI-IEN W-P, SHEU L-F, LEN C-Y: Pathogenesis of IgA-nephropathy: in vitro activation of human mesangial cells by IgAimmune complex leads to cytokine secretion. J Pathol 173:119—126,1994

8. NOBLE B, REN K, TAVERNE J, DIPIRRO J, VAN LIEW J, DIJKSTRA C,JANOSSY G, POULTER LW: Mononuclear cells in glomeruli andcytokines in urine reflect the severity of experimental proliferativeimmune complex glomerulonephritis. Clin Exp Immunol 80:281—287,1990

9. TIPPING PG, LOWE MG, HOLDSWORTH SR: Glomerular interleukin 1production is dependent on macrophage infiltration in anti-GBMglomerulonephritis. Kidney mt 39:103—110, 1991

10. Y0sI0KA K, TAKEMURA T, MURAKAMI K, OKADA M, YAGY K,MIYAZATO H, MATSUSHIMA K, MAKI S: In situ expression of cytokinesin IgA nephritis. Kidney mt 44:825—833, 1993

11. HONKANEN E, TEPPO AM, MERI S, LETHO T, GRONHAGEN-RISKA C:Urinary excretion of cytokines and complement SC5b-9 in idiopathicmembranous glomerulonephritis. Nephrol Dial Transplant 9:1553—1559, 1994

12. LAN HY, NIKOLIC-PATERSON DJ, ZARAMA M, VANNICE JL, ATKINS

RC: Suppression of experimental crescentic glomerulonephritis byinterleukin-1 receptor antagonist. Kidney mt 43:479—485, 1993

13. KIBERD BA, STADNYK AW: Established murine lupus nephritis doesnot respond to exogenous interleukin-1 receptor antagonist; a role forthe endogenous molecule. Immunopharmacology 30:131—137, 1995

14. SIMS J, GAYLE M, SLACK J, ALDERSON M, BIRD T, GIRl J, COLLOTOFR, MANTOVANI A, SHANEBECK K, GRABSTEIN K, DOWER 5: Inter-leukin 1 signaling occurs exclusively via the type 1 receptor. Proc NatIAcad Sci USA 90:6155—6159, 1993

15. COLOrrA F, RE F, MUZIO M, BERTINI R, POLENTARUTTI N, SIR0NI M,GIRl JG, DOWER SK, SIMs JE, MANT0vANI A: Interleukin-1 type IIreceptor: A decoy target for IL-i that is regulated by IL-4. Science261:472—475, 1993

16. SYMONS JA, YOUNG PR, DUFF GW: Soluble type II interleukin 1(IL-i) receptor binds and blocks processing of IL-1/3 precursor andloses affinity for IL-i receptor antagonist. Proc Natl Acad Sci USA92:1714—1718, 1995

17. CARTER DB, DEIBEL MR, DUNN CJ, TOMICH CSC, LABORDE AL,SLIGHTOM JL, BERGER AE, BIENKOWSK1 MJ, SUN FF, MCEVAN RN,HARRIS PKW, YEM AW, WASZAK GA, CHOSAY GJ, SIEU LC, HARDEEMM, ZURCHER-NEELY HA, REARDON IM, HEINRIKSON RL, TRUES-DELL SE, SHELLY IA, EESSALU TE, TAYLOR BM, TRACEY DE:Purification, cloning, expression and biological characterization of aninterleukin-1 receptor antagonist protein. Nature 344:633—638, 1990

18. MULLER WA, RATFI CM, MCDONNEL SL, COHN ZA: A humanendothelial cell-restricted externally disposed plasmalemmal proteinenriched in inter-cellular junctions. J Exp Med 170:399—414, 1989

19. ALBELDA SM, OLIVER PD, ROMER LH, BUCH CA: EndoCAM: Anovel endothelial cell-cell adhesion molecule. J Cell Biol 110:1227—1237, 1990

20. MUNDEL P, GILBERT P, KRIZ W: Podocytes in glomerulus of rat kidneyexpress a characteristic 44 kD protein.JHistochem Cytochem 39: 1047—1056, 1991

21. NIEMIR ZI, STEIN H, NORONFtA IL, KRUGER C, ANDRASSY K, RITZ E,WALDI-IERR R: PDGF and TGF-j3 contribute to the natural course ofhuman IgA glomerulonephritis. Kidney mt 48:1530—1541, 1995

22. BALDARI C, MURRAY JAM, GHIARA P, CESARENI G, GALEOTrI CL: Anovel leader peptide which allows efficient secretion of a fragment ofhuman interleukin 1f3 in Saccharomyces cerevisiae. EMBO J 6:229—234, 1987

23. MARCH CJ, MOSLEY B, LARSEN A, CERErrI DP, BRAEDT G, PRICE V,GILLIS 5, HENNEY CHS, KRONHEIM SR, GRABSTEIN K, CONLON PJ,HOPP TP, C0SMAN D: Cloning, sequence and expression of twodistinct interleukin-l complementary DNAs. Nature 315:641—647,1985

24. HEINO P, HUKKANEN V, ARSTILA P: Detection of human papillomavirus (HPV) DNA in genital biopsy specimens by in situ hybridizationwith digoxigenin-labeled probes. J Virol Meth 26:331—337, 1989

25. THORNBERRY NA, BULL HG, CALAYCAY JR, CHAPMAN KT, HOWARDAD, KOSTURA MI, MILLER DK, MOLINEAUX SM, WEIDNER JR,AUNINS J, ELLISTON KO, AYALA JM, CASANO FJ, CHIN J, DING GJ-F,EGGER LA, GAFFNEY EP, LIMJUCO G, PALYCHA OC, RAJu SM,RoLA.r4DO AM, SALLEY JP, YAMIN T-T, LEE TD, SHIVELY JE,MACCROSS M, MUMFORD RA, SCHMIDT IA, T0CCI MI: A novelheterodimeric cysteine protease is required for interleukin-113 pro-cessing in monocytes. Nature 356:768—774, 1992

26. KLAPPROTH J, CASTELL J, GEIGER T, ANDUs T, HEINRICH PC: Fateand biological action of human recombinant interleukin 1/3 in the ratin vivo. EurJlmmunol 19:1485—1490, 1989

27. RUBARTELLI A, CozzouNo F, TALIO M, SITIA R: A novel secretorypathway for interleukin-i/3, a protein lacking a signal sequence.EMBO J 9:1503—1510, 1990

28. HOGOQUIST KA, NETr MA, UNANUE ER, CHAPLIN DD: Interleukin 1is processed and released during apoptosis. Proc Natl Acad Sci USA88:8485—8489, 1991

29. CONLON P1, GRABSTEIN KH, ALPERT A, PRICKETr KS, HOPP TP,GILLIS S: Localization of human mononuclear interleukin 1. J Immu-nol 139:98—102, 1987

30. MOSLEY B, URDAL DL, PRICKETF KS, LARSEN A, COSMAN D, CONLONP1, GILLIS 5, DOWER SK: The interleukin-1 receptor binds the humaninterleukin-la precursor but not the interleukin-113 precursor. J BiolChem 262:2941—2944, 1987

31. HASKILL S, MARTIN G, VAN LE L, MORRIS J, PEACE A, BIGLER CK,JAFFE GJ, HAMMERBERG C, SPORN SA, FONG S, AREND WP, RALPH P:eDNA cloning of an intracellular form of the human interleukin 1receptor antagonist associated with epithelium. Proc Natl Acad SciUSA 88:3681—3685, 1991

32. KENNEDY MC, ROSENBAUM iT, BROWN I, PLANCK SR, HUANG X,ARMSTRONG CA, ANSEL JC: Novel production of interleukin-1 recep-tor antagonist peptides in normal human cornea. J Clin Invest95:82—88, 1995

Niemir et al: IL-i and glomerular protection 403

33. HIRSCH E, IRIKuIt VM, PAUL SM, HIRSH D: Functions of interleukin1 receptor antagonist in gene knockout and overproducing mice. ProcNatlAcad Sci USA 93:11008—11013, 1996

34. RICHARDSON CA, GORDON KL, COUSER WG, BOMSZTYK K: IL-1J3increases laminin B2 chain mRNA levels and activates NF-kB in ratglomerular epithelial cells. Am J Physiol 268(Renal Fluid ElectrolPhysiol 37):F273—F278, 1995

35. NAKAZATO Y, OKADA H, TAJIMA 5, HAYASHIDA T, KANNO Y, SUZUKIH, SARUTA T: Interleukin-4 modulates collagen synthesis by humanmesangial cells in a type-specific manner. Am J Physiol 270(RenalFluid Electrol Physiol 39):F447—F453, 1996

36. SMITH PS, FANNING JC, AARONS I: The structure of the normal humanglomerular basement membrane. Pathology 21:254—258, 1989

37. FUKATSU A, MATSUO 5, KILLEN PD, MARTIN GR, ANDRES GA,BRENTJENS JR: The glomerular distribution of type IV collagen and

laminin in human membranous glomerulonephritis. Hum Pathol19:64—68, 1988

38. KRIz W, KRETZLER M, NAGATA M, PRovoosT AP, SHIRATO I, UnS, SAKAI T, LEMLEY KV: A frequent pathway to glomerulosclerosis:Deterioration of tuft architecture—podocyte damage—segmentalsclerosis. Kidney Blood Press Res 19:245—253, 1996

39. KRETZLER M, KOEPPEN-HAGEMANN I, KRIz W: Podocyte damage is acritical step in the development of glomerulosclerosis in the unine-phrectomised-desoxycortisterone hypertensive rat. Virchows Arch 425:181—193, 1994

40. FRIES JW, SANDSTROM DJ, MEYER TW, RENNKE HG: Glomerularhypertrophy and epithelial cell injury modulate progressive glomeru-loscierosis in the rat. Lab Invest 60:205—218, 1989

41. KRIZ W, HACKENTHAL E, NOBILING R, SJcAI T, ELGER M: A role forpodocytes to counteract capillary wall distension. Kidney Jot 45:369—376, 1994