ARTHRITIS & RHEUMATOLOGY
Vol. 66, No. 2, February 2014, pp 397–406
DOI 10.1002/art.38250
© 2014, American College of Rheumatology
Pathogenic Mechanisms in Lupus Nephritis
Nucleosomes Bind Aberrant Laminin ␤1 With High Affinity and
Colocalize in the Electron-Dense Deposits
Anders I. Olin,1 Matthias Mörgelin,2 Lennart Truedsson,1
Gunnar Sturfelt,1 and Anders A. Bengtsson1
Objective. Apoptotic nucleosomes are structurally
and immunologically involved in lupus nephritis. The
purpose of this study was to examine the expression and
function of laminins and their interactions with nucleo-
somes in the kidneys of patients with lupus nephritis,
using surface plasmon resonance (SPR) analysis.
Methods. SPR interaction analysis was used to
quantify the strength of laminin–nucleosome interac-
tions. Electron microscopy techniques were used to
determine in vivo colocalization of IgG, chromatin, and
laminin ␤1, as well as to characterize nucleosome–
laminin interactions in vitro.
Results. Nucleosomes were found to possess high
affinity for laminin ␤1–containing laminins and to have
the potential to form stable molecular complexes with
these structures. In vivo, laminin ␤1 was aberrantly
expressed in the glomerular basement membrane
(GMB) of lupus nephritis patients, and in situ, it acted
as a nucleosome ligand, selectively colocalizing with
nucleosomes within electron-dense deposits (EDDs).
Normal adult laminin 11, which contains laminin ␤2,
did not bind nucleosomes, and it did not colocalize in
Supported by grants from the Swedish Research Council, the
Medical Faculty at Lund University, the Swedish National Association
Against Rheumatism, the Alfred Österlund Foundation, the Crafoord
Foundation, the Greta and Johan Kock Foundation, King Gustaf V’s
80th Birthday Foundation, the Swedish Society of Medicine, and Lund
University Hospital.
1
Anders I. Olin, MD, PhD, Lennart Truedsson, MD, PhD,
Gunnar Sturfelt, MD, PhD, Anders A. Bengtsson, MD, PhD: Lund
University and Lund University Hospital, Lund, Sweden; 2Matthias
Mörgelin, PhD: Lund University, Lund, Sweden.
Dr. Truedsson has received consulting fees from Euro Diag-
nostica (less than $10,000).
Address correspondence to Anders I. Olin, MD, PhD, Lund
University, Division of Infection Medicine, BMC B14, SE-221 84
Lund, Sweden. E-mail: anders.olin@med.lu.se.
Submitted for publication April 4, 2012; accepted in revised
form October 22, 2013.
397
vivo with the nucleosomes in the nephritic GBM. In
addition, TGF␤1 was expressed by the glomerular mes-
angium, glomerular endothelial cells, and by podocytes
in patients with lupus nephritis. It was trapped in situ
within EDDs by an as-yet-unknown ligand. As was
recently described in a transgenic mouse model, para-
crine kidney glomerular TGF␤1 may thereby contribute
to the development of glomerulopathy via the induction
of laminin ␤1 incorporation in the GBM, whereas
systemic blood vessel–derived TGF␤1 could be trapped
during filtration.
Conclusion. Our findings of the specific high-
affinity binding of nucleosomes to aberrantly expressed
laminin ␤1 in the GBM and their colocalization in the
GBM may explain important features of the initial steps
in the pathogenesis of lupus nephritis, the planted
antigen hypothesis.
It has recently become clear that the nucleosome
particle has a particular pathogenic impact on lupus
nephritis. Aside from their being a central target struc-
ture for nephritogenic autoantibodies, nucleosomes are
proven major autoantigens that drive T cell–dependent
autoimmune responses (1). It has been suggested that
the glomerular basement membrane (GBM) itself does
not bind lupus autoantibodies in vivo; instead, a planted
antigen hypothesis, in which nucleosomes have associ-
ated with the GBMs or with the mesangial matrix, has
been proposed as an initial step in the pathogenesis of
lupus nephritis (2,3). These nucleosomes may originate
from circulating or intraglomerular apoptotic cells, re-
sulting in both cases in the formation of subendothelial
or subepithelial electron-dense deposits (EDDs) (4).
The nephritogenic autoantibodies could, in turn, bind to
the GBM-associated nucleosomes if the nucleosomes
398
are not already trapped in complexes with anti–double-
stranded DNA (anti-dsDNA) antibodies in the circula-
tion (3,5), which may also be the case.
The laminins represent a family of large glycosy-
lated proteins with a multidomain structure, largely
confined to the basement membranes. They are hetero-
trimeric molecules composed of ␣, ␤, and ␥ chains,
which are linked by disulfide bonds and named accord-
ingly. In the kidney, laminin 11 (␣5␤2␥1 [or, 521]) is the
apparent constituent of fully mature normal GBM,
replacing laminin 1 (␣1␤1␥1 [or, 111]) during develop-
ment (6). The kidney mesangial matrix contains laminin
1, laminin 2 (␣2␤1␥1 [or, 211]), laminin 8 (␣4␤1␥1 [or,
411]), and laminin 10 (␣5␤1␥1 [or, 511]) (7), while in
tubular basement membranes, there are laminin 1 (prox-
imal tubule), laminin 5 (␣3A␤3␥1 [or, 331]) (collecting
ducts), and laminin 10 (distal tubule and collecting
ducts).
In the kidney, glomerular endothelial cells and
podocytes jointly synthesize both laminin 1 chains during
development and laminin 11 chains during maturation
(8).
Transforming growth factor ␤ (TGF␤) is a key
component in the control of wound repair and in the
development of fibrosis (9,10). The TGF␤1 isoform is
specifically found in the developing duct, in mesenchy-
mal cells, and in the immature glomeruli. During subse-
quent stages of nephron development, TGF␤1 is not
expressed and is not found in the healthy matured
glomeruli or proximal tubules (11).
In the present investigation, we demonstrated
the presence of TGF␤1, as expressed by podocytes and
the mesangium, during lupus nephritis, thus, in a sense
recapitulating glomerulogenesis in the disease. In-
terestingly, during infection and inflammation, TGF␤1
has specifically been shown to induce an aberrant
laminin ␤1 chain expression in the kidney (12). We also
found that nucleosomes associated only with laminins
containing the ␤1 chain in the GBM and in the mesan-
gial matrix of patients with lupus nephritis, which led
us to hypothesize that local TGF␤1 may play an im-
portant role in the pathogenesis of lupus nephritis.
We showed that nucleosomes were bound to the aber-
rant laminin ␤1–containing laminin and were therefore
trapped, where they became available for binding to
anti-dsDNA and antinucleosome autoantibodies, which
are lupus nephritis–specific antibodies. Our findings may
contribute to an explanation of the planted antigen
hypothesis.
OLIN ET AL
MATERIALS AND METHODS
Surface plasmon resonance (SPR) interaction analy-
sis. Nucleosomes and laminin isoforms and chains were puri-
fied as described elsewhere (13–15). Briefly, nucleosomes were
prepared from BALB/c mouse fibroblasts, disrupted by freez-
ing and thawing, and then sheared by aspiration through a thin
needle. Nucleosomes were solubilized by digestion of nuclei
for 20 minutes at 37°C with 2 units of micrococcal nuclease per
50 ␮g/ml of DNA in a 0.3M sucrose and 10 mM Tris buffer.
Digested nuclei were sedimented at 12,000g, and chromatin
was extracted overnight in 0.2 mM EDTA, pH 7.0. Extracted
nuclear debris was pelleted at 3,000g. Supernatants from the
fractionations were further fractionated by the addition of
NaCl to a concentration of 0.15M to solubilize nucleosomes.
The nucleosomes were dialyzed against 0.2 mM EDTA, pH
7.0, and stored frozen until used.
Electrophoresis of nucleosomes in sodium dodecyl
sulfate–17% polyacrylamide showed that the nucleosomes
contain all of the histone classes (H1, H2A, H2B, H3, H4, and
H5) (for a description of the electrophoresis result, see Figure
3, lane 1, in ref. 16). In the present study, the purity of the
nucleosome preparation was additionally validated by negative
staining and electron microscopy of nucleosome complexes, as
described below.
The laminins (a kind gift from Dr. Lydia Sorokin,
Institute of Physiological Chemistry and Pathobiochemistry,
University of Munster, Munster, Germany) were diluted in 10
mM sodium acetate, pH 4, and immobilized via amine coupling
to flow cells of CM5 sensor chips (Biacore 2000). Immobiliza-
tion levels were optimized to ⬃600–4,000 response units. A
flow cell subjected to the immobilization protocol but without
any addition of protein was used as a control.
For affinity measurements, the binding and dissocia-
tion phases were monitored with a Biacore 2000 instrument. In
control experiments for possible mass transfer limitations,
rabbit anti-human laminin ␤1 (HPA004056; Atlas Antibodies)
and nucleosomes were injected over the laminin surface at
different flow rates. No differences in initial binding were
observed at a rate of 5 ␮l/minute or above, indicating no
limitations to any combinations.
Antilaminin antibodies or nucleosomes were injected
at different concentrations (typically 8–500 nM; flow rate 15
␮l/minute at 25°C) over the flow cells in running buffer (10 mM
HEPES, pH 7.5, 150 mM NaCl, 0.005% Biacore Surfactant
P20, and 3.4 mM EDTA).
Between experiments, the surfaces were strictly regen-
erated with pulses of 0.1M NaHCO3, pH 12, and with running
buffer containing 2M NaCl, followed by an extensive wash
procedure after reaching baseline. After x-axis and y-axis
normalization of data (injection starting points set at zero), the
blank curves from the control flow cells of each injected
concentration were subtracted. The association and dissocia-
tion rate constants were determined simultaneously using the
equation for 1:1 Langmuir binding in the BIAevaluation 4.1
software (Biacore). The binding curves were fitted locally, and
the equilibration dissociation constants (KD) were calculated
from mean values of the rate constants obtained.
Negative staining and electron microscopy. Typically,
5-␮l samples of complexes between nucleosomes and the
different laminins (20–30 nM) were adsorbed to 400-mesh
NUCLEOSOME BINDING OF ABERRANT LAMININ ␤1 IN LUPUS NEPHRITIS 399

carbon-coated copper grids, washed briefly in water, and

stained with 0.75% uranyl formate. The nucleosomes and
laminins were also treated similarly for individual visualization.
Specimens were observed with a JEOL 1230 EX transmission
electron microscope operated at an accelerating voltage of 60
kV. Images were recorded with a Gatan Multiscan 791 CCD
camera. From the photos obtained, typical monomers and
complexes were selected.
Analysis of proteins applied to PVDF membranes.
Nucleosomes were radiolabeled with 125I using Iodo-Beads
iodination reagent (Pierce). Laminin ␤1 and laminin ␤2 (Neo-
markers; Thermo Fisher Scientific) or laminin ␣1 and laminin
␣2 (a kind gift of Rupert Timpl’s laboratory, Max-Planck-
Institute for Biochemistry, Martinsried, Germany) were di-
rectly applied in equal amounts to a PVDF membrane using a
Milliblot-D system device (Millipore). The membrane was
blocked for 1 hour with 3% bovine serum albumin (BSA) in
phosphate buffered saline/0.05% Tween 20 (PBST) and incu-
bated for 3 hours in room temperature with 4 ⫻ 105 counts per
minute/ml of 125I-labeled nucleosomes. After washing with
PBST (3 times for 20 minutes each), bound ligands were
detected using a Fuji Imaging System (Fuji Photo Film Com-
pany).
Patients. Five patients with systemic lupus erythema-
tosus (SLE) attending the Department of Rheumatology at
Lund University Hospital who fulfilled the American College
of Rheumatology (ACR) 1982 classification criteria for SLE
(17) were selected for study based on the following features: 1)
an episode of active renal disease with proteinuria (⬎0.5 gm of Figure 1. Typical Biacore sensorgrams showing the binding of anti-
albumin per 24 hours), hematuria, and/or cellular casts plus 2) laminin antibodies (A) and nucleosomes (B) to immobilized laminin
histopathologic findings compatible with diffuse proliferative (laminin 1 in these images). Typically, an 8–500 nM analyte was
lupus nephritis that included active lesions. Disease activity injected and, afterwards, replaced by buffer, for monitoring the
was assessed by the SLE Disease Activity Index (18), and was dissociation phase (arrowheads). For laminin 1, the equilibration
also calculated using solely the renal items from the instru- dissociation constant (KD) was calculated to be 0.5 nM for the antibody
ment. and 4.3 nM for nucleosome binding. RU ⫽ response units.

Clinical characteristics of the patients are available

Table 1. The laminins investigated, their renal expression within upon request from the corresponding author. The patients met
membranes, their laminin ␤1 chain content, and their nucleosomal a median of 7 ACR criteria (range 5–7). Their mean age at the
binding, as determined by surface plasmon resonance* time of biopsy was 39 years (range 24–69 years). Four of the 5
patients were women. Renal biopsies were performed by
Laminin, renal Nucleosomal percutaneous ultrasonography–guided puncture in accordance
compartment expression ␤1 chain binding (Kd, nM) with a standard protocol. The specimens obtained were clas-
Laminin 1 (␣1␤1␥1) Yes 4.3 sified by an experienced, independent renal pathologist ac-
GBM (immature) cording to the International Society of Nephrology/Renal
GBM (lupus nephritis) Pathology Society 2003 guidelines for lupus nephritis (19). All
TBM (proximal) biopsy samples were evaluated by both light microscopy and
MM immunofluorescence.
Laminin 2 (␣2␤1␥1) Yes 5.8 Control specimens were obtained from healthy renal
MM
tissue that originated from a patient undergoing nephrectomy
Laminin 5 (␣3␤3␥1) No ND
TBM (collecting ducts) because of renal cancer.
Laminin 8 (␣4␤1␥1) Yes 7.2 Immunoelectron microscopy. Biopsy samples were
MM fixed with a modified version of Karnovsky’s fixative containing
Laminin 10 (␣5␤1␥1) Yes 7.7 2% paraformaldehyde and 0.5% glutaraldehyde in 0.1M So-
MM rensen’s phosphate buffer. After blocking with 0.2% BSA
TBM (proximal) (Aurion) and 5% goat serum, grids were incubated overnight
TBM (collecting ducts) with rabbit anti-human laminin ␤1 antibody (HPA004056) or
Laminin 11 (␣5␤2␥1) No No rabbit anti-human laminin ␤2 antibody (HPA001895; both
GBM (mature/healthy)
from Atlas Antibodies) and with a murine anti-human dsDNA
* GBM ⫽ glomerular basement membrane; TBM ⫽ tubular basement monoclonal antibody (163p77; obtained from Dr. T. Marion,
membrane; MM ⫽ mesangial matrix; ND ⫽ not determined. University of Tennessee, Memphis, TN) in 0.2% BSA.
400 OLIN ET AL

Figure 2. Molecular complexes of nucleosomes and laminin isoforms, as visualized using negative staining and electron microscopy. A, Overview
of the complex network formation, showing laminin (arrows) and nucleosomes (arrowheads). Bar ⫽ 100 nm (75 nm for B, C, D, F, and H). B, The
diversity and purity of nucleosomes in 3 different visual fields. C, Typical shapes of laminin isoforms, showing laminin 1 (cruciform), laminin 8
(T-shaped), and laminin 2 (cruciform), respectively. D–I, The 3 laminin isoforms in C are each shown in complex with nucleosomes (arrowheads)
together with an illustrative drawing, demonstrating different binding sites of laminin 1 (D and E), laminin 8 (F and G), and laminin 2 (H and I).
In the drawings, the black areas indicate nucleosomes, and the gray areas indicate laminin molecules.

The 163p77 experimental antibody was chosen as a restricted to the EDDs, and shown not to cross-react with
molecular tool for recognizing nucleosomes since it has previ- regular structures of the GBM (3). In our experience, the
ously been shown to colocalize with in vivo–bound autoanti- antibody was indeed specific for EDD nucleosomes and did
body deposits (as detected with gold-conjugated protein A) not stain any GBM components (including laminins) in tissue
NUCLEOSOME BINDING OF ABERRANT LAMININ ␤1 IN LUPUS NEPHRITIS
Figure 3. Binding of 125I-labeled nucleosomes to laminin ␤1 chains
following incubation with laminin ␤1 (left lane) and laminin ␤2 (right
lane). Both laminins were immobilized (slot-blot) in the indicated
amounts (in ␮g) onto a PVDF membrane.
sections from healthy controls or in regions outside the
EDDs in sections from lupus nephritis patients. With this
technique, colocalization studies could be performed as simul-
taneously detecting trapped nucleosomes and their GBM
ligand (laminin ␤1) without concomitant staining of the in
vivo–deposited autoantibodies (human IgG) in the EDDs.
Due to a species-specific secondary antibody (anti-mouse),
staining will thus display the highest specificity possible.
Colocalization was defined as molecular probes within a
proximity of 30 nm, according to the method of Philimonenko
et al (20).
Rabbit anti-human TGF␤1 antibody was obtained
from BioVision. The binding signal was detected with 10 nm
gold-conjugated goat anti-rabbit and 5 nm gold-conjugated
goat anti-mouse (Agar Scientific) secondary antibodies, re-
spectively. The specificity of the secondary antibodies was
tested by omission of the primary antibody. Samples were
finally stained with uranyl acetate and lead citrate before
examination by electron microscopy and evaluation as de-
scribed above. In all experiments, molecules were deemed to
be colocalized when 80–90% of the different sized gold-
conjugated antibodies met the 30 nm proximity criterion
(20,21).
RESULTS
High affinity of nucleosomes for laminins that
contain ␤1 chains. Screening for the binding of nucleo-
somes to GBM and mesangial matrix components was
undertaken using Biacore SPR technology. Nucleo-
somes were clearly found to have affinity for various and
distinctive laminin constructs (Table 1 and Figure 1B).
401
The laminin 1 surface was validated by the binding of an
antilaminin antibody (Figure 1A). Affinities were calcu-
lated to be KD ⫽ 0.5 nM for the antibody and KD ⫽ 4.3
nM for the nucleosomes.
Electron microscopic visualization of the binding
of nucleosomes to laminins. Electron microscopy al-
lowed the visualization of the large tentacled laminin
molecules (Figure 2). Representative images from mul-
tiple experiments provide an overview of the complexity
of the laminin network formation with nucleosomes
(Figure 2A), as well as the diversity and purity of the
nucleosomes (Figure 2B). The shapes of individual
laminins are demonstrated in Figure 2C, which shows
laminin 1 (cruciform), laminin 8 (T-shaped), and
laminin 2 (cruciform). A detailed view of laminin 1 in
complex with nucleosomes is shown in Figure 2D, with
the multimericity of the interaction emphasized in the
corresponding drawing in Figure 2E. Noteworthy is the
binding of nucleosomes to different globular domains of
the laminin 1 molecule, which allows the laminin 1
molecule to concentrate nucleosomes around itself.
Figure 2F and its corresponding drawing in Fig-
ure 2G demonstrate the binding of T-shaped laminin 8
to nucleosomes, both at globular and stretched domains.
The multimericity in the context of the binding of
different ends of nucleosome aggregates to different
laminin domains can also be seen. The 2 concepts of
multiple points of interaction are again shown in Figures
2H and I, in this case with laminin 2. Thus, using
electron microscopy we could demonstrate and visualize
the binding of nucleosomes to different laminins and to
different domains of the laminin molecules.
Binding of nucleosomes to ␤1 chains, but not ␤2
chains, of laminin. We investigated differences in
nucleosome binding to the individual laminin ␤1 and ␤2
chains, since ␤1 represents an immature form of laminin,
which we found to be expressed in vivo during lupus
nephritis (see below) and which has been reported to be
seen during inflammatory conditions, such as viral infec-
tions (12). The laminins investigated above that were
found to bind nucleosomes all have the laminin ␤1 chain
in common. In these experiments, we used a slot-blot
assay and clearly showed laminin ␤1 to be the preferred
ligand for interaction (Figure 3). Laminin ␣1 and
laminin ␣2 did not bind nucleosomes (results not
shown).
Colocalization of nucleosomes and laminin ␤1–
containing laminins within EDDs. A colocalization im-
munoelectron microscopic investigation using anti–
laminin ␤1 or anti–laminin ␤2 antibodies was performed
on GBMs derived from kidney biopsy tissues obtained
from SLE patients with glomerulonephritis (Figures
402 OLIN ET AL

Figure 4. Colocalization immunoelectron microscopy of glomeruli (glomerular basement membrane [GBM]) in kidney sections from patients with
lupus nephritis (A–C and G–P) and sections of a healthy area of the kidney from a patient undergoing nephrectomy because of cancer (D–F).
Anti–double-stranded DNA (anti-dsDNA; i.e., nucleosomes) tagged with a 5-nm gold tracer and anti–laminin ␤1 (i.e., laminin ␤1) tagged with a
10-nm gold tracer are colocalized (80–90% of the gold-labeled antibodies are within 30 nm of each other) within electron-dense deposits (EDDs)
(B) and within adjacent membranes (C) in kidney sections from a patient with lupus nephritis. Boxed areas in A (top and bottom, respectively) are
shown at higher magnification in B and C. Boxed areas in B and C indicate representative areas of colocalization that were investigated further.
Detailed colocalization in kidney sections from each of the 5 lupus nephritis patients (G–K) is also shown, demonstrating nucleosomes/anti-dsDNA
antibodies (arrowheads) and laminin (arrows). Laminin ␤1 is also shown in GBMs that do not have EDDs (nucleosomes). Only very sparse
background detection of gold-labeled antibodies is observed in the healthy kidney section. Boxed areas in D are shown at higher magnification in
E and F (top and bottom boxes, respectively). Anti-dsDNA antibodies (arrowheads) tagged with a 5-nm gold tracer and anti–laminin ␤2 antibodies
(arrows) tagged with a 10-nm gold tracer are not colocalized within comparable sections of the EDDs from each of the 5 patients with lupus nephritis
(L–P). Laminin ␤2, however, is widely distributed throughout the GBMs of all 6 individuals (results not shown), including the EDDs of the lupus
nephritis patients (L–P). Colocalization is indicated by the proximity of the 2 different gold conjugates (within 30 nm). Bars ⫽ 0.5 ␮m in D (applies
also to A); 0.2 ␮m in F (applies also to B, C, and E); and 25 nm in P (applies also to G–O).

4A–C and G–P) and from healthy kidney tissue obtained
from a control patient who had undergone nephrectomy
because of a localized tumor in a different area of the
kidney (Figures 4D–F).
Nucleosomes/anti-dsDNA monoclonal anti-
bodies (tagged with a 5-nm gold tracer) and anti–laminin
␤1 antibodies (tagged with a 10-nm gold tracer) were
shown to colocalize in distinct EDDs located adjacent to
glomerular capillary membranes (Figure 4A), as well as
in more diffuse and smaller EDDs dispersed within the
NUCLEOSOME BINDING OF ABERRANT LAMININ ␤1 IN LUPUS NEPHRITIS
Figure 5. A, Kidney section from a patient with lupus nephritis, showing
the characteristic thickening of the glomerular basement membrane
(GBM) and overall distorted morphology typical of the disease. Boxed
areas in A (labeled B–G) are shown at higher magnification in the
accompanying images. B–G, Immunoelectron microscopy, demonstrating
the localization of transforming growth factor ␤1 (TGF␤1) (tagged with
10-nm gold tracer) in electron-dense deposits (EDDs) (B–D) and in
podocytes (P) and the mesangium (E–G). TGF␤1 is shown in E and F as
a pattern of accumulation in the podocyte cell body (4 vesicles [V] shown
at upper right in E), which implements vesicular intracellular transport
(exocytosis), and as accumulations (arrowheads) in the extracellular space
(ES) in the vicinity of the foot processes (lighter areas). The typical EDDs
of the GBM (BM) in E and F display their TGF␤1 content. Within the
mesangial cells and mesangial compartment, the expression and activity of
TGF␤1 are also abundant. Arrowheads in G indicate vesicles. H, Glomer-
ular endothelial cells (E) and podocytes, demonstrating their joint syn-
thesis of aberrant laminin ␤1–containing laminin (10-nm gold tracer) in
lupus nephritis, which was not produced in the healthy kidney (results not
shown). Bars ⫽ 5 ␮m in A; 100 nm in D (applies also to B and C); and 0.5
␮m in G and H (bar in G also applies to E and F).
403
membranes (Figures 4A–C). Figures 4G–K show the
details of the tissues from each of the 5 typical SLE
patients examined, demonstrating the colocalization of
nucleosomes with laminin ␤1, as determined by the
finding that 80–90% of different-sized gold particles lay
within 30 nm of each other. Nucleosomes (traced with
5-nm gold) did not colocalize with anti–laminin ␤2
(traced with 10-nm gold) in samples from any of the 5
SLE patients (Figures 4L–P), since they were not found
within 30 nm of each other (20,21).
Aside from the normal appearance of the healthy
kidney sample from the control patient (Figure 4D),
only very sparse background laminin ␤1 and dsDNA
signals were detected in the GBM (Figures 4E and F).
Normal laminin ␤2 (10-nm gold tracer) was widely
codistributed with laminin ␤1 in GBMs from the SLE
patients (with EDDs included) (Figures 4L–P), as well
as being expressed alone in the healthy control tissue, as
was expected (results not shown).
Immunoelectron microscopy of local TGF␤1 pro-
duction, paracrine activities, laminin ␤1 production,
and EDD entrapment. Transgenic mice producing
TGF␤1 locally in their kidneys undergo an isotype
switch from normal adult laminin ␤2 to aberrant laminin
␤1 (22). The production of both laminin 1 during
development or lupus (Figure 5H) and laminin 11 in the
healthy adult takes place in endothelial cells and podo-
cytes (8). Since we have demonstrated intense staining of
laminin ␤1 in the kidneys of patients with lupus nephri-
tis, we also investigated whether TGF␤1, which is known
to drive laminin chain conversion, was present in the
same nephritic glomeruli.
Indeed, TGF␤1 was found in glomerular struc-
tures and cells in direct comparison with its distribution
in the kidneys of transgenic animals (23). Accumulations
of TGF␤1 were present in this pattern in podocytes
(Figures 5E and F) and the mesangium (Figure 5G), as
well as in the extracellular spaces (Figures 5E and F),
which implies exocytosis.
We were thus able to demonstrate local up-
regulation of TGF␤1 and paracrine TGF␤1 activities
(e.g., through exocytosis) in kidneys from patients with
lupus nephritis. During lupus, aberrant laminin ␤1 is
produced by the endothelium and by podocytes (Figure
5H). We also found that the TGF␤1 content was con-
fined to the EDDs (Figures 5A–D), as previously de-
scribed by other investigators using lower-resolution
light microscopy (23,24). The specific ligand for TGF␤1
trapping within the EDDs, however, was not investi-
gated in the present study.
Summary of the results. Findings of the present
study confirm that in vivo–bound dsDNA-containing
structures (nucleosomes) are present in the glomeruli of
404
nephritic kidneys from patients with SLE and that their
presence is confined to small and large EDDs within
capillary membranes (GBM) (2). Our findings add to
what was already known: that TGF␤1 and laminin ␤1 are
aberrantly expressed during lupus nephritis and that the
nucleosomes are colocalized only with that specific
laminin chain within the GBM. Nucleosomes were not
detected outside the EDDs and were not colocalized
with normal laminin ␤2 coexpressed within the EDDs,
which demonstrated the binding specificity for laminin
␤1. Based on our novel TGF␤1 findings in the kidneys of
patients with lupus nephritis, we suggest that the ob-
served ␤1 conversion of laminin is driven by TGF␤1.
DISCUSSION
The exact pathogenic processes accounting for
development of glomerulonephritis in SLE have re-
mained elusive. The extent to which immune complex
deposits are derived from the circulation or are formed
locally by the cross-reaction between anti-dsDNA and
␣-actinin or other glomerular constituents has been
examined (25). Nucleosomes have been suggested as a
major immunogen responsible for the induction of po-
tentially pathogenic antinuclear antibodies (1), and they
also seem to serve as a target antigen complex for
autoantibody-mediated tissue lesions in the kidneys of
patients with lupus nephritis (26,27). The possibility that
nucleosomes are trapped in the glomeruli by glomerular
constituents such as type IV collagen, heparan sulfate,
or other negatively charged residues has been suggested
previously (5,25,26). Impaired clearance of apoptotic
cells has been proposed as a defect that may explain a
systemic and/or intraglomerular release of nucleosomes
as a significant source of autoantigens (4,5,28,29).
It has been shown that nucleosomes have the
capacity to associate with the GBM or the mesangial
matrix in murine and human variants of lupus nephritis
(2,26). In glomerular membranes, they are consistently
observed as small and larger EDDs, which are found
both subendothelially and subepithelially. We show in
the present study that at least one reason for the
trapping could be that they have intrinsic affinity for
specific GBM structures, as previously reported by
Mjelle et al (16), who used SPR to demonstrate binding
to a mixture of laminins and to isolated type IV collagen.
The laminins form cruciform or T-like structures de-
pending on their chain content and are held together
with the type IV collagen meshwork by nidogen and
heparan sulfate proteoglycans. Laminins self-assemble,
and they fully traverse the GBM. All of these character-
istics, together with the known up-regulation of laminin
OLIN ET AL
during infection and inflammation, make them ideal
ligands for the attachment of nucleosomes, which is
relevant to the development of nephritis in SLE patients.
However, the normal adult laminin of the GBM, laminin
11, does not bind nucleosomes.
In the present investigation, we were able to
visualize the interactions between nucleosomes and
laminin ␤1–containing laminins by electron microscopy.
In addition, the binding strength of these interactions
was found to be substantial, as measured by Biacore
SPR, which is consistent with previous results (16).
Individual ␣1 chains did not bind to nucleosomes when
immobilized on PVDF membranes (results not shown),
nor did the ␣2 chain of laminin 2 bind to nucleosomes
(results not shown), whereas full-length laminin 2 did
bind to nucleosomes. Since normal adult laminin 11 did
not bind nucleosomes, we conclude that like the ␤2
chain, the ␥1 chain could not be a determinant for the
nucleosome binding.
The laminin ␤1 chain with affinity for nucleo-
somes was found to be aberrantly expressed in the
kidneys of our patients with lupus nephritis, which is a
novel finding in this disease, since laminin ␤1 is not
normally expressed in adult and mature GBM. Nucleo-
somes and laminin ␤1–containing laminin colocalize in
large and smaller EDDs distributed along the mem-
branes of the nephritic glomeruli in SLE patients. Nor-
mally expressed laminin ␤2 has been reported not to
colocalize with nucleosomes (15). We also confirmed
these laminin ␤2 results in our SLE patients.
In the same biopsy samples, we detected specific
staining for TGF␤1, and we suggest that local produc-
tion of TGF␤1 by podocytes and mesangial cells drives
the production of immature laminin chains. This feature
has also been demonstrated experimentally using
TGF␤1-transgenic animals (12,22). Bioactive TGF may
derive from the circulation, the juxtaglomerular appara-
tus, or the glomerular cells themselves during infections
and inflammation (30,31). In the latter cases, the latent
pool of TGF␤ constitutively secreted as inactive forms is
attenuated by the production of a postulated “inducer”
from the glomerular cells (32,33). TGF␤1 induces the
expression, and thus the deposition, of type IV collagen
as well as aberrant fetal laminin ␣1, ␣2, and ␤1 chains,
into the GBM (12). Simultaneously, TGF␤1 blocks
matrix degradation by decreasing the synthesis of pro-
teases and by increasing the levels of protease inhibitors
(34). The aberrant and thickened GBM in the kidneys of
SLE patients will then serve as a target for immune
complex deposition.
Interestingly, TGF␤1 has also been found by us
and by other investigators to be confined to the EDDs in
NUCLEOSOME BINDING OF ABERRANT LAMININ ␤1 IN LUPUS NEPHRITIS
lupus nephritis (23,24). Consumption in the kidney may
reflect lower circulating levels of TGF␤1, which seem to
be associated with severe disease, kidney involvement,
and worse outcome in SLE (35,36). Additionally, an
elevated total plasma level of TGF␤1 in SLE has been
shown in some studies, and TGF␤1-transgenic mice
develop a glomerulonephritis resembling immune
complex–mediated glomerular injury (22). These find-
ings are all compatible with an important role of TGF␤1
in the pathogenesis of SLE.
Most likely, local production of TGF␤1 can be
seen in several different inflammatory kidney disorders,
and its biologic function is probably to moderate the
inflammatory reaction and recover damaged tissue.
However, in SLE, large amounts of nucleosomes due to
defective apoptotic cell clearance and abnormal apopto-
sis have been well described (5). In this situation, when
TGF␤1-induced aberrant laminin ␤1 is combined with
nucleosomes and autoantibodies that bind nucleosomes,
inflammation and organ damage, which are seen as the
planted antigen in SLE, is an unwanted consequence.
In conclusion, our findings are compatible with
the idea that nucleosomes derived from apoptotic cells
are trapped within glomerular basement membranes
during the development of lupus nephritis. Here, circu-
lating anti-dsDNA antibodies may cause nephritis
through binding the planted antigens in situ, although
the possibility that the nucleosomes also could bind
GBMs as part of preformed nephritogenic immune
complexes cannot be excluded. Based on the results of
our immunoelectron microscopy studies, the nucleo-
somes must, however, have different binding sites for the
laminin ␤1 and for the autoantibody in order to be able
to crosslink these proteins. We therefore conclude that
aberrant laminin 1 is an intrinsic ligand of the GBM,
with multiple sites having the potential to bind nucleo-
somes with high affinity. TGF␤1, which is dysregulated
in SLE, might drive the induction of an aberrant
laminin-chain expression (i.e., ␤1), which in turn, is
possibly related to certain viral infections. This feature
could be of a significant pathobiologic interest during
the development of lupus nephritis.
ACKNOWLEDGMENTS
Professor Lydia M. Sorokin (Institute of Physiological
Chemistry and Pathobiochemistry, Münster University, Mün-
ster, Germany), Dr. Ole Petter Rekvig (Molecular Pathology
Research Group, Faculty of Medicine, University of Tromsø,
Tromsø, Norway), and the late professor Rupert Timpl (Max-
Planck-Institute for Biochemistry, Martinsried, Germany) are
gratefully acknowledged for invaluable discussions and gifts of
405
proteins. Mrs. Maria Baumgarten provided skillful technical
advice and assistance.
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it
critically for important intellectual content, and all authors approved
the final version to be published. Dr. Olin had full access to all of the
data in the study and takes responsibility for the integrity of the data
and the accuracy of the data analysis.
Study conception and design. Olin, Mörgelin, Sturfelt, Bengtsson.
Acquisition of data. Olin, Mörgelin, Bengtsson.
Analysis and interpretation of data. Olin, Mörgelin, Truedsson,
Sturfelt, Bengtsson.
REFERENCES
1. Mohan C, Adams S, Stanik V, Datta SK. Nucleosome: a major
immunogen for pathogenic autoantibody-inducing T cells of lupus.
J Exp Med 1993;177:1367–81.
2. Van Bruggen MC, Kramers C, Walgreen B, Elema JD, Kallenberg
CG, van den Born J, et al. Nucleosomes and histones are present
in glomerular deposits in human lupus nephritis. Nephrol Dial
Transplant 1997;12:57–66.
3. Kalaaji M, Mortensen E, Jorgensen L, Olsen R, Rekvig OP.
Nephritogenic lupus antibodies recognize glomerular basement
membrane-associated chromatin fragments released from apopto-
tic intraglomerular cells. Am J Pathol 2006;168:1779–92.
4. Dieker JW, van der Vlag J, Berden JH. Deranged removal of
apoptotic cells: its role in the genesis of lupus. Nephrol Dial
Transplant 2004;19:282–5.
5. Berden JH, Grootscholten C, Jurgen WC, van der Vlag J. Lupus
nephritis: a nucleosome waste disposal defect? J Nephrol 2002;15
Suppl 6:S1–10.
6. Miner JH. Renal basement membrane components. Kidney Int
1999;56:2016–24.
7. Petajaniemi N, Korhonen M, Kortesmaa J, Tryggvason K, Seki-
guchi K, Fujiwara H, et al. Localization of laminin ␣4-chain in
developing and adult human tissues. J Histochem Cytochem
2002;50:1113–30.
8. St John PL, Abrahamson DR. Glomerular endothelial cells and
podocytes jointly synthesize laminin-1 and -11 chains. Kidney Int
2001;60:1037–46.
9. Blobe GC, Schiemann WP, Lodish HF. Role of transforming
growth factor ␤ in human disease. N Engl J Med 2000;342:1350–8.
10. Border WA, Noble NA. Transforming growth factor ␤ in tissue
fibrosis. N Engl J Med 1994;331:1286–92.
11. Clark AT, Young RJ, Bertram JF. In vitro studies on the roles of
transforming growth factor-␤1 in rat metanephric development.
Kidney Int 2001;59:1641–53.
12. Chai Q, Krag S, Miner JH, Nyengaard JR, Chai S, Wogensen L.
TGF␤1 induces aberrant laminin chain and collagen type IV
isotype expression in the glomerular basement membrane.
Nephron Exp Nephrol 2003;94:e123–36.
13. Andreassen K, Bredholt G, Moens U, Bendiksen S, Kauric G,
Rekvig OP. T cell lines specific for polyomavirus T-antigen
recognize T-antigen complexed with nucleosomes: a molecular
basis for anti-DNA antibody production. Eur J Immunol 1999;29:
2715–28.
14. Rekvig OP, Hannestad K. The specificity of human autoantibodies
that react with both cell nuclei and plasma membranes: the nuclear
antigen is present on core mononucleosomes. J Immunol 1979;
123:2673–81.
15. Sixt M, Hallmann R, Wendler O, Scharffetter-Kochanek K, So-
rokin LM. Cell adhesion and migration properties of ␤2-integrin
negative polymorphonuclear granulocytes on defined extracellular
406
matrix molecules: relevance for leukocyte extravasation. J Biol
Chem 2001;276:18878–87.
16. Mjelle JE, Rekvig OP, Fenton KA. Nucleosomes possess a high
affinity for glomerular laminin and collagen IV and bind nephri-
togenic antibodies in murine lupus-like nephritis. Ann Rheum Dis
2007;66:1661–8.
17. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield
NF, et al. The 1982 revised criteria for the classification of systemic
lupus erythematosus. Arthritis Rheum 1982;25:1271–7.
18. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang DH,
and the Committee on Prognosis Studies in SLE. Derivation of the
SLEDAI: a disease activity index for lupus patients. Arthritis
Rheum 1992;35:630–40.
19. Weening JJ, D’Agati VD, Schwartz MM, Seshan SV, Alpers CE,
Appel GB, et al, on behalf of the International Society of
Nephrology and Renal Pathology Society Working Group on the
Classification of Lupus Nephritis. The classification of glomerulo-
nephritis in systemic lupus erythematosus revisited [published
erratum appears in J Am Soc Nephrol 2004;15:835–6]. J Am Soc
Nephrol 2004;15:241–50.
20. Philimonenko AA, Janacek J, Hozak P. Statistical evaluation of
colocalization patterns in immunogold labeling experiments. J
Struct Biol 2000;132:201–10.
21. Kim JH, Cramer L, Mueller H, Wilson B, Vilen BJ. Independent
trafficking of Ig-␣/Ig-␤ and ␮-heavy chain is facilitated by dissoci-
ation of the B cell antigen receptor complex. J Immunol 2005;175:
147–54.
22. Wogensen L, Nielsen CB, Hjorth P, Rasmussen LM, Nielsen AH,
Gross K, et al. Under control of the Ren-1c promoter, locally
produced transforming growth factor-␤1 induces accumulation of
glomerular extracellular matrix in transgenic mice. Diabetes 1999;
48:182–92.
23. Nakajima M, Kawahara S, Sakagami Y, Takagawa K, Akazawa H,
Kamitsuji H, et al. Immunogold labelling of cytokines in glomeruli
in children with various renal diseases. Nephron 1999;83:132–8.
24. Yoshioka K, Takemura T, Murakami K, Okada M, Hino S,
Miyamoto H, et al. Transforming growth factor-␤ protein and
mRNA in glomeruli in normal and diseased human kidneys. Lab
Invest 1993;68:154–63.
25. Mortensen ES, Rekvig OP. Nephritogenic potential of anti-DNA
antibodies against necrotic nucleosomes. J Am Soc Nephrol
2009;20:696–704.
OLIN ET AL
26. Kalaaji M, Sturfelt G, Mjelle JE, Nossent H, Rekvig OP. Critical
comparative analyses of anti–␣-actinin and glomerulus-bound
antibodies in human and murine lupus nephritis. Arthritis Rheum
2006;54:914–26.
27. Minota S, Yoshio T, Iwamoto M, Takeda A, Masuyama J, Mimori
A, et al. Selective accumulation of anti-histone antibodies in
glomeruli of lupus-prone lpr mice. Clin Immunol Immunopathol
1996;80:82–7.
28. Gaipl US, Voll RE, Sheriff A, Franz S, Kalden JR, Herrmann M.
Impaired clearance of dying cells in systemic lupus erythematosus.
Autoimmun Rev 2005;4:189–94.
29. Gaipl US, Sheriff A, Franz S, Munoz LE, Voll RE, Kalden JR, et
al. Inefficient clearance of dying cells and autoreactivity. Curr Top
Microbiol Immunol 2006;305:161–76.
30. Border WA, Yamamoto T, Noble NA. Transforming growth factor
␤ in diabetic nephropathy. Diabetes Metab Rev 1996;12:309–39.
31. Hong SW, Isono M, Chen S, Iglesias-De La Cruz MC, Han DC,
Ziyadeh FN. Increased glomerular and tubular expression of
transforming growth factor-␤1, its type II receptor, and activation
of the Smad signaling pathway in the db/db mouse. Am J Pathol
2001;158:1653–63.
32. Dobonici M, Podlech J, Steffens HP, Maiberger S, Reddehase MJ.
Evidence against a key role for transforming growth factor-␤1 in
cytomegalovirus-induced bone marrow aplasia. J Gen Virol 1998;
79:867–76.
33. Wallick SC, Figari IS, Morris RE, Levinson AD, Palladino MA.
Immunoregulatory role of transforming growth factor ␤ (TGF␤)
in development of killer cells: comparison of active and latent
TGF␤1. J Exp Med 1990;172:1777–84.
34. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS,
Wakefield LM, et al. Transforming growth factor type ␤: rapid
induction of fibrosis and angiogenesis in vivo and stimulation of
collagen formation in vitro. Proc Natl Acad Sci U S A 1986;83:
4167–71.
35. Jin T, Almehed K, Carlsten H, Forsblad-d’Elia H. Decreased
serum levels of TGF␤1 are associated with renal damage in female
patients with systemic lupus erythematosus. Lupus 201;21:310–8.
36. Becker-Merok A, Eilertsen GO, Nossent JC. Levels of transform-
ing growth factor-␤ are low in systemic lupus erythematosus
patients with active disease. J Rheumatol 2010;37:2039–45.