Scand J Rheumatol 2010;39:132–140 132

Immigration check for neutrophils in RA lining: laminin a5 low expressionregions act as exit points

P Poduval1,2, T Sillat1,2, I Virtanen2, M Dabagh3, YT Konttinen1,4,5

1Department of Medicine, Helsinki University Central Hospital, 2Department of Anatomy, University of Helsinki, 3Department of Energyand Environmental of Technology, Lappeenranta University of Technology, Lappeenranta, 4ORTON Orthopaedic Hospital of theORTON Foundation Helsinki, and 5COXA Hospital for Joint Replacement, Tampere, Finland

Objective: A correlation exists between the absence of a5-laminin and transit checkpoint fenestrations in vascular

basement membranes. We hypothesized that similar laminin a5 low expression regions might exist in synovial lining,

which, although lacking basement membrane, contains all basement membrane components in its interstitial matrix.

Methods: Laminin a4 and a5 chains and lactoferrin were stained using immunofluorescence and cathepsin G and

neutrophil elastase using immunoperoxidase. Quantitative real-time polymerase chain reaction (qRT-PCR) was used to

measure laminina4 anda5mRNA copy numbers in cultured synovial fibroblasts, without/with tumour necrosis factor-a(TNFa) and interleukin-1b (IL-1b).Results: Laminin a4 and a5 chains were found in the intercellular matrix in synovial lining samples of trauma and

revision total hip replacements. Laminin a5 was weaker in osteoarthritis (OA) and rheumatoid arthritis (RA), and RA

synovial lining also contained local low expression areas. Double staining disclosed convergence of lactoferrin-

degranulating neutrophils towards these laminina5 low expression regions. In culturedOA synovial fibroblasts, laminina5mRNA decreased (p < 0.05) at 1 ng/mL TNFa and was not found at all in cultured resting or cytokine-stimulated RA

fibroblasts. Degranulation of cathepsin G and neutrophil elastasewas seen in neutrophils passing through blood vessels or

synovial lining.

Conclusions:Migrating neutrophils in RA seem to use laminin a5 chain low expression regions to exit synovial tissue to

enter synovial fluid. Transmigrating neutrophils remodel the intercellular matrix by releasing their proteolytic granular

contents to enhance these low expression checkpoints and/or to produce chemotactic stimuli. In RA fibroblasts this is

facilitated by cytokine-mediated down-regulation or lack of laminin a5 synthesis.

Basement membranes are specialized sheets of extra-

cellular matrix in contact with the epithelia and endo-

thelia (1). Synovial lining is a unique structure that, in its

intercellular matrix, between macrophage-like type A

and fibroblast-like type B lining cells, contains most of

the laminins (2) and type IV collagen a chains (3), which

are elsewhere components of lamellar basement mem-

branes. This type of specialized extracellular matrix

forms the backbone that supports the tissue architecture.

Its other important functions include regulation of cell

behaviour by interaction with the cell surface matrix

receptors and storage of growth factors.

Laminins are large heterotrimeric T- or cross-shaped

molecules with one long arm and two or three short arms

(4, 5). The long arm or the foot of the cross is composed

of parts of each of the three chains (a, b, and g) organized

into an a-helical coiled-coil domain. The short arms are

composed of parts of one chain only, the handle of part of

the a chain and the left and right arms of the crossbar by

parts of the b and g chains, respectively. The laminin achains are considered to be the functionally important

portions of the heterotrimers because they exhibit tissue-

specific distribution and contain the major cell inter-

action sites (5). Human endothelial cells produce laminin

a4 and a5 chains. Laminin-411 (a4b1g1) and laminin-

511 (a5b1g1) are important constituents of the vascular

endothelial basement membranes (6). The endothelial

cells barely adhere to laminin-411 (7) because it lacks the

short arm composed of the a chain. By contrast, laminin-

511 carries two exposed arginine-glutamine-aspartic acid

(RGD) cell-binding sites in its NH2-terminal portion,

which might form a transit barrier.

In a mouse model of autoimmune encephalomyelitis,

transmigration of mononuclear cells through the base-

ment membrane was detected only at sites where laminin

a4 chains were present but laminin a5 chains wereYrjo T Konttinen, Department of Medicine, Biomedicum, PO Box 700,

FIN-00029 HUS, Finland.

E-mail: yrjo.konttinen@helsinki.fi

Accepted 22 July 2009

� 2010 Taylor & Francis on license from Scandinavian Rheumatology Research Foundation

DOI: 10.3109/03009740903198980 www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

lacking. This also indicates a role for laminins in the

regulation of transendothelial migration (8).

The basement membrane provides an effective barrier

to leucocytes and this barrier has to be breached for

effective migration of the leucocytes. The migration of

leucocytes from the vascular lumen to extravascular

tissues is a fundamental event in innate and adaptive

immunity. It has been shown that neutrophils migrate

through low expression areas in the basement membrane

of the alveolar capillaries and the venular basement

membrane (9, 10). Wang et al described regions within

the walls of the venules where expression of a5-laminins

was low (10). They suggested that the transmigrating

neutrophils might induce a transient remodelling of the

basementmembrane, possibly through the involvement of

leucocyte proteinases, such as neutrophil elastase, which

facilitates neutrophil transmigration. Cathepsin G is

another neutral proteinase found in azurophil granules

of polymorphonuclear neutrophilic leucocytes, which

could play a role in basement membrane remodelling.

We hypothesized that similar laminin a5 low expression

regions and remodelling might also be found in synovial

lining, which, although lacking the basement membrane,

contains all basement membrane components in its

interstitial matrix.

Materials and methods

Tissue samples

The study protocol was accepted by the Ethical Com-

mittee of the Helsinki and Uusimaa Hospital District.

Five trauma, five prosthesis loosening (synovialmembrane-

like interface tissue around aseptically loosening pros-

theses, with an ongoing foreign body synovitis), five

osteoarthritis (OA) and five rheumatoid arthritis (RA)

synovial tissue samples were used. Trauma synovial

membrane samples were obtained from injured knees,

which were examined using arthroscopy; biopsies taken

disclosed no or only slight synovial inflammation. Syno-

vial tissue samples were snap frozen in dry ice pre-cooled

isopentane, embedded in Optimal Cutting Temperature�(OCT) embedding medium (Sakura Finetek Europe B.V.,

Zoeterwoud, the Netherlands) and stored at –70�C.Samples were cut into 5-mm tissue sections, air dried

at +22�C for 1 h and stored at –20�C until used for

staining. Formalin-fixed, paraffin-embedded sections

were used to study neutrophil elastase and cathepsin G

in synovial tissues.

Immunofluoresence staining

After fixation in acetone at –20�C for 10 min, the cryostat

sections were rinsed thoroughly in 0.1% Triton-X in

10 mM phosphate-buffered saline (PBS; pH 7.4) and

then washed in PBS alone. The sections were incubated

serially at +22�C in a humidified chamber in (i) normal

goat serum (1:50, Dako, Glostrup, Denmark) for 1 h,

(ii) monoclonal mouse anti-human laminin a4 immuno-

globulin (Ig)G1 (1:50, 168FC10; 11) or monoclonal

mouse anti-human laminin a5 IgG2a (1:10, 4C7; 12, 13),

(iii) conjugated goat anti-mouse IgG (1:200, Alexa Fluor

568 for laminin a4 or Alexa Fluor 488 for the laminin a5antibodies, Molecular Probes, Eugene, OR, USA), all at

+22�C for 60 min. Between the steps, the sections were

washed in PBS, and after the last step they were also

washed once in distilled water before mounting in

Vectashield� (Vector Laboratories, Burlingame, CA,

USA). The stained sections were analysed using fluores-

cent microscopy. Mouse IgG1 and IgG2a (Dako) against

glucose oxidase of Aspergillus niger, not present or

inducible in human tissues, were used as negative staining

controls for laminin a4 and a5, respectively.

Double immunofluorescence staining

After fixation in acetone at –20�C for 10 min, cryostat

sections from RA patients were rinsed thoroughly with

0.1% Triton X-100 in PBS and washed in PBS. The

sections were incubated serially at +22�C in a humidified

chamber as follows: (i) normal goat serum (1:50; Dako);

(ii) 37 mg/mL polyclonal rabbit anti-human lactoferrin Ig

(Dako) together with monoclonal mouse anti-human

laminin a5 IgG2a [1:10 in 0.1% bovine serum albumin

(BSA) in PBS] (4C7; 12, 13); (iii) a mixture of Alexa

Fluor 568-conjugated goat anti-rabbit IgG and Alexa

Fluor 448-conjugated goat anti-mouse IgG (1:200;

Molecular Probes), all for 60 min at +22�C; betweenthe steps the sections were washed in PBS, and after the

last step they were also washed once in distilled water

before mounting in Vectashield (Vector Laboratories).

Similar to the immunofluorescence single staining, the

negative control staining was performed for double

immunofluorescence staining using irrelevant mouse

IgG2a and normal rabbit immunoglobulin (Dako).

Immunohistochemical staining

Paraffin sections were deparaffinized in xylene,

rehydrated through a graded ethanol series and washed

in PBS. For antigen retrieval the RA and OA slides were

pretreated with 0.4% pepsin in 1 N HCl for the detection

of neutrophil elastase, while for the detection of

cathepsin G the antigens were retrieved by heat treatment

for 30 min in 0.01 M sodium citrate buffer, pH 6.0, using

MicroMED T/T Mega Laboratory Microwave Systems

(Milestone, Sorisole, Italy). Endogenous peroxidase

activity was blocked using 0.3% H2O2 in methanol for

30 min. After washing in PBS, the sections were incu-

bated in the following reagents at +22�C unless stated

otherwise: (i) normal horse serum (for monoclonal

mouse anti-human antibodies) for 60 min to block

non-specific binding sites; (ii) monoclonal mouse anti-

human neutrophil elastase IgG1 (1:50, M 0752 Dako)

monoclonal mouse anti-human cathepsin G IgG1 (1:50,

PMN transmigration in RA lining 133

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

NCL-CATH-G, Novocastra Laboratories, Newcastle,

UK); (iii) biotinylated horse anti-mouse IgG (Vectastain

ABC Kits, 1:200, Vector Laboratories); (iv) avidin–

biotin peroxidase complex (ABC, 1:200 in dH2O,

Vector Laboratories); and (v) a combination of 0.023%

3,3-diaminobenzidine tetrahydrochloride and 0.006%

H2O2 for 7 min. Finally, the slides were dehydrated and

mounted inMountex (HistoLab,Gothenburg,Sweden).All

antibodies used above were diluted in 0.1% BSA in PBS.

Microscopic assessment

The single and double immunofluorescent sections were

observed under the fluorescence imaging microscope

Olympus AX70 (Olympus, Vienna, Austria) coupled to

a CCD camera under 200–400� magnification. Immuno-

peroxidase stained slides were analysed under 400�magnification using a light microscope (Leitz, Wetzlar,

Germany), coupled to a 12-bit cooled CCD camera

(Sensicam, PCO Imaging, Kelheim, Germany).

Cell cultures

Synovial membrane samples were collected from three

patients suffering from OA and from three patients

suffering from RA. Synovial fibroblast cell lines were

established using the explant culture method. In brief,

tissue samples were minced into small pieces and left

overnight in RPMI-1640 medium (BioWhittaker,

Walkersville, MD, USA) containing 10% foetal bovine

serum (BioWhittaker) and 10% penicillin/streptomycin.

The following day, the medium was changed and the

concentration of antibiotics was decreased to 1%. The

medium was changed twice a week and when about 60%

of the dish area was covered by a monolayer of cells, the

tissue pieces were removed and the cultures were allowed

to grow to confluence.

Quantitative real-time polymerase chain reaction(qRT-PCR)

For stimulations, 105 cells/well were grown to confluence

in six-well plates. Cells from two parallel wells were used

for RNA extraction. The cells were stimulated for 48 h

with 0.1, 1, and 10 ng/mL recombinant human tumour

necrosis factor-a (rhTNFa) or recombinant human

interleukin-1b (rhIL-1b), both from R&D (Minneapolis,

MN, USA). Total RNA from cells was isolated using

TRIzol reagent (Invitrogen, Paisley, UK) according to the

manufacturer’s instructions. RNA quality was confirmed

with ethidium bromide-stained 1% agarose gel. The

mRNA was isolated with magnetic (dT)25-polystyrene

beads (Dynal, Oslo, Norway). mRNA concentration was

measured spectrophotometrically and cDNAwas synthe-

sized from 50 ng of sample mRNA using (dT)12–18

primers and Superscript enzyme, followed by RNase H

treatment (SuperScript Preamplification System, Invitro-

gen). Quantitative PCR was run using 4.8 ng first-strand

cDNA and 0.5 mM primers in LightCycler PCR mix in

a LightCycler PCR instrument (Roche, Mannheim,

Germany). qPCR runs were repeated twice with each

sample. For primers, the sequences were searched using

the NCBI Entrez Search system and sequence similarity

search was performed using the NCBI Blastn program.

The primers were designed, if possible, to produce an

amplicon that extended over two different exons (Table 1).

For the quantitative PCR standard curve, the gene of

interest was amplified in the PCR reaction, extracted

from agarose gel and cloned into pCRII-TOPO vector

(Invitrogen). After identification of the plasmid by restric-

tion enzyme analysis and sequencing, the concentration

was determined spectrophotometrically and serial dilu-

tions were prepared for quantitative PCR analysis. The

copy numbers of mRNAs were determined with two

separate runs for all samples and normalized against

1 � 106 housekeeping b-actin mRNA copies.

Statistical analysis

Statistical tools available in Microsoft Excel 2003

(Microsoft Corporation, Seattle, WA, USA) were used

to perform the statistical analysis. The results are pre-

sented as mean ± standard deviation. Student’s t-test wasused for pairwise comparisons.

Results

Immunofluorescence of laminin a chains

Immunoreactivity for laminin a4 and a5 chains was

found in the synovial lining in trauma samples and in

the synovial lining-like layer of interface membrane

samples retrieved during revision operations performed

for aseptic loosening of total hip replacement implants

(Figures 1A–1D). Laminin a4 was similarly labelled in

OA and RA samples (Figures 1E and 1G), but a5 chain

labelling was weaker in the synovial lining in OA

samples than in trauma and revision hip samples

(Figure 1F). Reaction for laminin a5 (Figure 1H) chains,although present in most areas of the hypertrophic lining,

Table 1. Primer sequences used in the qRT-PCR of laminin a4 and laminin a5; the analysis was standardized using the b-actinhousekeeping gene.

Left primer Right primer Amplicon length (bp)

Laminin a4 TGAAGCCAATGAAACAGCAG TGCTTAACGGCATCACTGAG 249Laminin a5 GGACACAGACGAGACAAGCA GTTGAACTTCATGGGCACCT 361b-actin TCACCCACACTGTGCCCATCTACGA CAGCGGAACCGCTCATTGCCAATGG 295

134 P Poduval et al

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

was also weak in RA lining; in addition, areas where

there was a striking lack of laminin a5 (Figure 1H,

between the arrowheads) in the presence of laminin

a4 (Figure 1G) were found in certain regions of the

RA synovial lining. In the laminin a chain staining, the

presence of laminin a4 and laminin a5 chain positive

basement membranes of the blood vessels provided a

useful internal positive staining control. Staining with the

irrelevant immunoglobulins of the same class and

subtype, using the same concentrations as were used

for immunolabelling, confirmed the specificity of the

staining results.

Immunoperoxidase staining of neutrophil elastase andcathepsin G

Neutrophil elastase and cathepsin G positive neutrophils

were relatively infrequent in the synovial membrane

samples from the trauma, revision and OA patients,

except for intravascular neutrophils. Intravascular neu-

trophils in OA samples stained strongly for both neutro-

phil elastase (Figure 2A) and cathepsin G (Figure 2B).

By contrast, although not so easy to discern in

haematoxylin–eosin stained routine slides, neutrophil

elastase (Figure 2C) and cathepsin G (Figure 2D) immu-

noreactive neutrophils were fairly frequent in RA syno-

vial membrane samples compared to the other samples.

In such samples they were not only found in the intra-

vascular compartment but often also attached to the

vascular endothelium or transmigrating through it, or

found in the perivascular and interstitial stroma as well as

in the sublining and lining cell layers. Intravascular

neutrophils stained strongly and this staining was con-

fined to the cytoplasm of the cells (Figures 2C and 2D,

arrows). By contrast, many of the transmigrating neu-

trophils in the vascular wall and/or in the synovial lining

A B

C D

E F

G H 30 μm

Figure 1. Indirect immunofluorescence staining

of laminin a4 (red) and laminin a5 (green)

chains in trauma (A and B), revision total hip

replacement (C andD), OA (E and F) and RA (G

and H). Arrowheads mark the synovial lining.

Staining of laminin a4 and a5 chains in the

vascular basement membranes provides an

internal positive staining control. Original

magnification �400. Insert in G shows a nega-

tive staining control for laminin a4.

PMN transmigration in RA lining 135

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

stained more faintly and immunolabelled neutrophil

elastase and cathepsin G were at the same time also

seen outside the cell in the pericellular stroma

(Figures 2C and 2D, arrowheads).

Double staining of laminin a5 chain and neutrophillactoferrin

Double immunofluorescence disclosed lactoferrin-

positive neutrophils attached to and transmigrating

through the vascular basement membrane. This was

seen in basement membrane areas, where there seemed

to be a break in laminin a5 staining (Figures 3A and 3B).

Transmigrating cells or perivascular cells just outside the

blood vessels, already in the interstitial tissue, stained

relatively faintly and specific lactoferrin immuno-

labelling was also seen outside the cell in the extracel-

lular matrix (Figure 3B). In the RA synovial stroma small

aggregates of neutrophils were located between the sub-

lining blood vessels and overlying synovial lining

(Figures 3C and 3D). Finally, in the synovial lining these

cells converged to laminin a5 chain low expression

regions, where they were located just below or within

the synovial lining itself (Figures 3E and 3F), passing

through the laminin a5 low expression regions (Figures

3E–3G). Staining controls confirmed the specificity of

the staining (Figure 3H).

qRT-PCR

Three OA synovial fibroblast cell lines were cultured,

stimulated, and analysed separately. Despite the signif-

icant differences in the age of patients and their genetic

backgrounds, the variations between the individual cell

lines were minor. qRT-PCR disclosed that cultured OA

synovial fibroblasts contained 1203 ± 211 laminin a4mRNA copies� 10–6 b-actinmRNAs, but only 189 ± 126laminin a5 mRNA copies � 10–6 b-actin mRNAs.

Recombinant human TNFa and IL-1b did not signifi-

cantly decrease the laminin a4 chains in cultured OA

fibroblasts (data not shown), but laminina5 chain mRNA

copy numbers (per 106 b-actin copies) were significantlydecreased by TNFa (p < 0.05; Figure 4); the effect of

IL-1 b was not statistically significant (data not shown).

However, at the dose range used, 0.1–10 ng/mL, this

effect did not show any clear dose dependency.

By contrast, unstimulated RA fibroblasts contained

10-fold less laminin a4 mRNA copies than OA fibro-

blasts and RA fibroblast laminin a4 did not decrease butincreased upon stimulation with TNFa and IL-1b; thiseffect was not dose dependent within the dose range

0.1–10 ng/mL. Although this increase was not statisti-

cally significant, it clearly shows that, in RA fibroblasts,

cytokine stimulation did not decrease laminin a4 copy

numbers. Of note, unstimulated RA did not produce any

measurable laminin a5, and laminin a5 mRNA copy

levels were not stimulated by either TNFa or IL-1b(Table 2).

Discussion

The basement membrane of the blood vessels represents

a true, full-thickness basement membrane. Unlike the

vascular basement membrane, the synovial lining layer

does not contain a basement membrane, which would

separate the synovial lining cell layer from the under-

lying connective tissue stroma. However, synovial lining

fibroblast-like type B and macrophage-like type A cells

have a thin layer of intercellular substance between them

acting as a putty or intercellular cement that contains

50μA B 50μ

C 50μ D 50μ

Figure 2. Avidin–biotin peroxidase complex

staining of neutrophil elastase and cathepsin

G in OA (A and B, respectively) and RA

(C and D, respectively). Staining in OA is

mostly confined to the cell cytoplasm in intra-

vascular neutrophils, whereas in RA some neu-

trophils below the synovial lining in the

perivascular matrix are in part degranulated

so that extracellular proteinase aggregates can

be seen in the synovial lining area (arrow-

heads). Original magnification �400.

136 P Poduval et al

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

components of the basement membrane (3,14–19).

Therefore, it is only natural that the intense staining

pattern of the basement membrane of the blood vessel

is not reproduced in the synovial lining, which displays a

much weaker staining of the thin layers of the

intercellular substance.

RA synovial fluid is characterized by a large number of

polymorphonuclear neutrophilic leucocytes. In RA syno-

vial fluid the number of leucocytes can be up to

5–50�109/L andmost of these leucocytes are neutrophils

(16). Furthermore, although neutrophils mature in the

bone marrow for 10–14 days, they only circulate for

6–12 h, and after transmigration into tissues and body

fluids live for 1–3more days before undergoing apoptosis

(17). Thismeans that the neutrophil efflux to, and turnover

in, synovial fluid is enormous in arthritis. Our results

suggest that the neutrophils pass through the synovial

lining using laminin a5 low expression regions. Such

gateway regions were found in the synovial lining, and

the double staining revealed that exactly these regions

seemed to attract transmigrating neutrophils, identifiedby

lactoferrin staining (18–19).Lactoferrin, used in thiswork

as a neutrophil marker, is formed during the myelocyte

stage of neutrophil development. It forms an important

component of the granules specific for neutrophils’, there-fore, called specific granules (or secondary granules

because they are formed after the primary or azurophil

granules that have already appeared during the

A B

C D

E F

10

10

20

Pixe

l int

ensi

ty

30

40

50

60

201 401 601 801Distance in pixels

1001 1201

G

50 μm

H

Figure 3. Double immunofluorescence staining

of laminin a5 chain (green) and neutrophil

biomarker lactoferrin (red). (A) A laminin a5low expression area of the vascular basement

membrane with an attached neutrophil (original

magnification �400). (B, C) Neutrophils at

different stages of (trans)migration in, at, and

outside the post-capillary venules (B: �400;

C: �200). The staining of intravascular neutro-

phils is very strong and confined to the cell

cytoplasm (thick arrow). By contrast, the extra-

vascular neutrophils are in part degranulated

and stain more faintly (arrowheads, C and D).

Neutrophils seem to aggregate, migrate, and

transmigrate en masse, which is clearly seen

in D (�200). They seem to converge to laminin

a5 low expression areas in the synovial lining

(E, arrow, �200), which is co-localized with

transmigrating, lactoferrin-immunoreactive,

and partly degranulated stream of neutrophils

(F, arrows, �200). Laminin a5 low expression

regions in E were subjected to densito-

metry; G shows the densitogram, with two

low expression regions. Inclination of the cut-

ting angle in relation to the checkpoint region

and the width between the two arrows in

E area may contribute to some background

laminin a5 immunoreactivity (G). (H) Negative

staining control for laminin a5/lactoferrindouble staining.

PMN transmigration in RA lining 137

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

promyelocyte stage and are, in contrast to the neutrophil

specific granules, also found in othermyeloid cells such as

monocyte/macrophages). We did not observe laminin a5low expression regions except in the RA lining. As the

synovialfluidneutrophil countsare fairly lowin, forexam-

ple, traumatic and prosthetic joints, it is not clear whether

such low expression regions are not present at all or

whether there are only very few inmildly inflamed joints.

TNFa decreased laminin a5 chain mRNA levels in

synovial fibroblasts from OA patients. IL-1b had a

similar effect although this was not statistically signifi-

cant. Together theymay significantly decrease the overall

concentration of the interstitial laminin a5 in the hyper-

trophic RA synovial lining. In line with these results,

laminin a4 chain mRNAwas found to be very low in RA

synovial fibroblasts, but in contrast to OA cells, laminin

a4 mRNA levels in RA cells did not decrease but

increased upon stimulation with TNFa or IL-1b. In

contrast to OA synovial fibroblasts, the laminin a5 chainwas not found at all in unstimulated RA fibroblasts and

could not be induced by these cytokines. This supports

our interpretation that one reason for the increased

neutrophil leakage through the RA synovial lining is

the low in situ synthesis of the laminin a5 chain. There is

a discrepancy here between the RA synovial lining and

cultured RA fibroblasts because the lining clearly con-

tains immunoreactive laminin a5 chain suggesting local

synthesis, whereas cultured RA fibroblasts totally lacked

laminin a5 chain mRNA. This suggests that the specific

micro-milieu in the lining plays a local role stimulating

the synthesis of laminin a5 chains, and this effect is not

seen when RA fibroblasts are cultured alone, in isolation

from other cells and their regular intercellular matrix.

This is a remarkable observation because very few clear-

cut differences have been hitherto described between

cultured OA and RA synovial fibroblasts (20–21).

Of note, these same cytokines induce chemokines,

including IL-8, which is released by stimulated synovial

cells and accumulates in RA synovial tissue and fluid

(22–26). IL-8 has also been detected at the pannus–

cartilage junction (27). The chemotactic activity of

synovial fluid for neutrophils was inhibited by 40%

when IL-8 was neutralized by an appropriate antibody

(26). RANTES is another chemokine found in increased

concentrations in inflamed synovial fluid. A previous

study showed a positive correlation between RANTES

levels and mononuclear cell migration to synovial fluid

(28). More chemotactic stimuli are generated by comple-

ment degradation products (29, 30), produced by immune

complex-mediated complement activation. Indeed, RA

has been considered an extravascular, intra-articular

immune complex disease, driven by IgG and rheumatoid

factor complexes and complement activation. This par-

adigm, formulated by Professor Nathan J. Zvaifler, still

holds true today (29, 31). The current immunohistopatho-

logical findings are in agreement with an effective and

massive neutrophil migration from the circulatory com-

partment to the synovial fluid and suggest laminina5 lowexpression regions as the transit gateways.

Some studies have investigated the cross-talk between

neuronal growth factor (NGF), its receptor NGF-R, and

fibroblast-like synovial cells. Pro-inflammatory cyto-

kines, such as TNFa and IL-1b, upregulate NGF and

its receptor tyrosine kinase (TrkA) in synovial fibro-

blasts. Fibroblast-like synovial cells proliferate following

stimulation with NGF in a dose-dependent manner (32).

Therefore, NGF signalling might also play a role for the

current finding.

Laminin a5 low expression regions in the synovial

lining of patients with RA were not due to an artefact

caused by tissue processing, because nuclear staining of

the samples disclosed intact a synovial lining cell layer at

these laminin a5 low expression regions. Furthermore,

a cutting and processing artefact is also refuted by the

neutrophils heading towards these immigrant check-

points, because these neutrophils represent a vital sign

proving the true existence of such transit areas in vivo.

Neutrophil serine proteinases, such as neutrophil

elastase and cathepsin G, are stored as active proteinases

in the neutrophil granules and upon stimulation they are

released into the extracellular space. Neutrophil elastase

may also be important as an in vivo activator of

0C α4 α5

20

mR

NA

cop

ies

40

60*

80

100

Figure 4. Quantitative real-time polymerase chain reaction (qRT-PCR)

analysis of laminin a4 and a5 chain mRNA copy numbers per 106

b-actin copies in cultured human synovial fibroblasts from OA. Com-

pared to the unstimulated synovial fibroblasts (control, C; normalized

to 100), in the presence of 1 ng/mLTNFa the laminina4 chainwas onlyslightly affected (this decreasewas not statistically significant), but the

laminina5messengerRNA levels fell significantly to about half of that

of the controls (p < 0.05, t-test).

Table 2. Effect of 0.1, 1, and 10 ng/mL TNFa or IL-1b on laminina4 chain mRNA copy numbers per 106 b actin copies in RAfibroblasts. The value for non-stimulated cultures was 100(range 76–126). Laminin a5 mRNA was not found in any ofthe RA samples.

TNFa IL-1b

0.1 ng/mL 658 (150–924) 294 (274–181)1 ng/mL 616 (133–854) 309 (229–389)10 ng/mL 619 (64–801) 171 (151–119)

138 P Poduval et al

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

pro-matrix metalloproteinases (MMPs), including the

inducible 92-kDa type IV collagenase or MMP-9 (but

not of the constitutive pro-form of the 72-kDa type IV

collagenase or MMP-2) (33). Cross-linking of neutrophil

Fcg receptors by immune complexes leads to neutrophil

activation (34). Consequently, immune complex-

activated neutrophils discharge large quantities of toxic

products, overwhelming the anti-proteinase shields of

synovial fluid (35, 36). This neutrophil activation and

degranulation has been considered to be important for

destruction of the hyaline articular cartilage (37–39).

Proinflammatory cytokines capable of prolonging neu-

trophil survival are detectable in RA synovial fluid

(40–41). We found low laminin a5 chain expression

regions in RA synovial lining. We therefore speculate

that the migrating neutrophils contribute to their genera-

tion by a localized degranulation and that a concerted

serine proteinase and MMP attack in part creates these

regions, as they do in transendothelial migration (10).

In the vascular basement membrane, low expression

regions of laminins have been reported to be constitu-

tively expressed but enhanced by neutrophil proteinases

so this may represent a difference between blood vessels

and synovial lining. However, based on the current data,

it is not possible to concludewhether such low expression

migration checkpoints in synovial lining are permanent

or transient. However, degradation products of laminins

and other basement membrane components could act as

locally produced and site-specific chemoattractants to

direct neutrophil migration (convergence) towards such

transit gateways (42, 43).

This work shows laminin a4 for the first time in the

synovial lining. As suggested by Hallmann et al (6), it is

possible that, in basementmembranes containing laminin

a4 but lacking laminin a5, the ability of the truncated

‘tau-cross’ laminin a4 alone to be incorporated into the

interstitial matrix is compromised. This would result in a

‘loose’ network facilitating neutrophil penetration (6).

The present data suggest a simple but effective mecha-

nism by which neutrophils could converge towards and

migrate across the synovial lining transit gateways into

the synovial fluid.

Acknowledgements

This work was supported in part by grants from Finska Lakaresallskapet,

the Finnish Dental Society Apollonia, the Research Foundation of

Farmos, the Finnish Society for Rheumatology, EVO grants, Invalid

Foundation, the Finnish Diabetes Research Foundation, and the Sigrid

Juselius Foundation.

References

1. LeBleu VS, MacDonald B, Kalluri R. Structure and function of

basement membranes. Exp Biol Med 2007;232:1121–9.

2. Konttinen YT, Li T-F, Xu J-W, Takagi M, Pirila L, Silvennoinen T,

et al. Expression of laminins and their integrin receptors in different

conditions of synovial membrane and synovial membrane-like

interface tissue. Ann Rheum Dis 1999;58:683–90.

3. Poduval P, Sillat T, Beklen A, Kouri VP, Virtanen I, Konttinen YT.

Type IV collagen a-chain composition in normal and arthritic

synovial lining. Arthritis Rheum 2007;56:3959–67.

4. Aumailley M, Bruckner-Tuderman L, Carter WG, Deutzmann R,

Edgar D, Ekblom P, et al. A simplified laminin nomenclature.

Matrix Biol 2005;24:326–32.

5. Miner JH, Yurchenco PD. Laminin functions in tissue morpho-

genesis. Annu Rev Cell Dev Biol 2004;20:255–84.

6. Hallmann R, Horn N, Selg M, Wendler O, Pausch F, Sorokin LM.

Expression and function of laminins in the embryonic and mature

vasculature. Physiol Rev 2005;85:979–1000.

7. Fujiwara H, Gu J, Sekiguchi K. Rac regulates integrin-mediated

endothelial cell adhesion and migration of laminin-8. Exp Cell Res

2004;292:67–77.

8. Sixt M, Engelhardt B, Pausch F, Hallmann R, Wendler O,

Sorokin LM. Endothelial cell laminin isoforms, laminins 8 and

10, play decisive roles in T cell recruitment across the blood-brain

barrier in experimental autoimmune encephalomyelitis. J Cell Biol

2001;153:933–46.

9. Walker DC, Behzad AR, Chu F. Neutrophil migration through

preexisting holes in the basal laminae of alveolar capillaries and

epithelium during streptococcal pneumonia. Microvasc Res

1995;50:397–416.

10. Wang S, Voisin MB, Larbi KY, Dangerfield J, Scheiermann C,

Tran M, et al. Venular basement membranes contain

specific matrix protein low expression regions that act as exit

points for emigrating neutrophils. J Exp Med 2006;203:

1519–32.

11. Petajaniemi N, Korhonen M, Kortesmaa J, Tryggvason K,

Sekiguchi K, Fujiwara H, et al. Localization of laminin

alpha4-chain in developing and adult human tissues. J Histochem

Cytochem 2002;50:1113–30.

12. Engvall E, Davis GE, Dickerson K, Ruoslahti E, Varon S,

Manthorpe M. Mapping of domains in human laminin using

monoclonal antibodies: localization of the neurite-promoting

site. J Cell Biol 1986;103:2457–65.

13. Tiger CF, Champliaud MF, Pedrosa-Domellof F, Thornell LE,

Ekblom P, Gullberg D. Presence of laminin alpha5 chain

and lack of laminin alpha1 chain during human muscle

development and in muscular dystrophies. J Biol Chem

1997;272:28590–5.

14. Revell PA, Al-Saffar N, Fish S, Osei D. Extracellular

matrix of the synovial intimal layer. Ann Rheum Dis 1995;54:

404–7.

15. Konttinen YT, Li T-F, Xu J-W, Takagi M, Pirila L, Silvennoinen T,

et al. Expression of laminins and their integrin receptors in

different conditions of synovial membrane and synovial

membrane-like interface tissue. Ann Rheum Dis 1999;58:

683–90.

16. Harris DE. Rheumatoid arthritis. Pathophysiology and implications

for therapy. N Engl J Med 1990;322:1277–89.

17. Bainton DF. Neutrophil granules. Br J Haematol 1975;29:

17–22.

18. Konttinen YT, Reitamo S. Effect of fixation on the antigenicity of

human lactoferrin in paraffin embedded tissues and cytocentrifuged

cell smears. Histochemistry 1979;62:55–64.

19. Konttinen YT, Reitamo S, Ranki A, Hayry P, Kankaanpaa U,

Wegelius O. Characterization of the immunocompetent cells of

rheumatoid synovium from tissue sections and eluates. Arthritis

Rheum 1981;24:71–9.

20. Konttinen YT, Saari H, Santavirta S, Antti-Poika I, Sorsa T,

Nykanen P, Kemppinen P. Synovial fibroblasts. Scand J Rheumatol

Suppl 1988;76:95–103.

21. Konttinen YT, Li TF, Hukkanen M, Ma J, Xu JW, Virtanen I.

Signals targeting the synovial fibroblast in arthritis. Arthritis Res

2000;2:348–55.

22. Koch AE, Kunkel SL, Burrows JC, Evanoff HL, Haines GK,

Pope RM, et al. Synovial tissue macrophage as a source of the

chemotactic cytokine IL-8. J Immunol 1991;147:2187–95.

PMN transmigration in RA lining 139

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

23. Golds EE, Mason P, Nyirkos P. Inflammatory cytokines induce

synthesis and secretion of gro protein and a neutrophil chemotactic

factor but not beta 2-microglobulin in human synovial cells and

fibroblasts. Biochem J 1989;259:585–8.

24. Seitz M, Dewald B, Gerber N, Baggiolini M. Enhanced production

of neutrophil-activating peptide-1/interleukin-8 in rheumatoid

arthritis. J Clin Invest 1991;87:463–9.

25. Brennan FM, Zachariae CO, Chantry D, Larsen CG, Turner M,

Maini RN, et al. Detection of interleukin 8 biological activity in

synovial fluids from patients with rheumatoid arthritis and produc-

tion of interleukin 8 mRNA by isolated synovial cells. Eur J

Immunol 1990;20:2141–4.

26. Koch AE, Volin MV, Woods JM, Kunkel SL, Connors MA,

Harlow LA, et al. Regulation of angiogenesis by the C-X-C

chemokines interleukin-8 and epithelial neutrophil activating pep-

tide 78 in the rheumatoid joint. Arthritis Rheum 2001;44:31–40.

27. Deleuran BW. Cytokines in rheumatoid arthritis. Localization in

arthritic joint tissue and regulation in vitro. Scand J Rheumatol

Suppl 1996;104:1–34.

28. Ellingsen T, Buus A, Møller BK, Stengaard-Pedersen K. In vitro

migration of mononuclear cells towards synovial fluid and plasma

from rheumatoid arthritis patients correlates to RANTES synovial

fluid levels and to clinical pain parameters. Scand J Rheumatol

2000;29:216–21.

29. Zvaifler NJ. Rheumatoid synovitis. An extravascular immune

complex disease. Arthritis Rheum 1974;17:297–305.

30. Low JM, Moore TL. A role for the complement system in rheu-

matoid arthritis. Curr Pharm Des 2005;11:655–70.

31. Edwards SW, Hallett MB. Seeing the wood for the trees: the

forgotten role of neutrophils in rheumatoid arthritis. Immunol

Today 1997;18:320–4.

32. Raychaudhuri SP, Raychaudhuri SK. The regulatory role of nerve

growth factor and its receptor system in fibroblast-like synovial

cells. Scand J Rheumatol 2009;28:1–9.

33. Nagase H. Activation mechanisms of matrix metalloproteinases.

Biol Chem 1997;378:151–60.

34. Willis HE, Browder B, Feister AJ, Mohanakumar T, Ruddy S.

Monoclonal antibody to human IgG Fc receptors. Cross-linking of

receptors induces lysosomal enzyme release and superoxide gen-

eration by neutrophils. J Immunol 1988;140:234–9.

35. Ossanna PJ, Test ST, Matheson NR, Regiani S, Weiss SJ. Oxidative

regulation of neutrophil elastase-alpha-1-proteinase inhibitor

interactions. J Clin Invest 1986;77:1939–51.

36. Abbink JJ, Kamp AM, Nuijens JH, Swaak TJ, Hack CE.

Proteolytic inactivation of alpha 1-antitrypsin and alpha

1-antichymotrypsin by neutrophils in arthritic joints. Arthritis

Rheum 1993;36:168–80.

37. Ugai K, Ishikawa H, Hirohata K, Shirane H. Interaction of poly-

morphonuclear leukocytes with immune complexes trapped in

rheumatoid articular cartilage. Arthritis Rheum 1983;26:1434–41.

38. Weiss SJ. Tissue destruction by neutrophils. N Engl J Med

1989;320:365–76.

39. Chatham WW, Swaim R, Frohsin H Jr, Heck LW, Miller EJ,

Blackburn WD Jr. Degradation of human articular cartilage by

neutrophils in synovial fluid. Arthritis Rheum 1993;36:51–8.

40. Xu WD, Firestein GS, Taetle R, Kaushansky K, Zvaifler NJ.

Cytokines in chronic inflammatory arthritis. II. Granulocyte-

macrophage colony-stimulating factor in rheumatoid synovial effu-

sions. J Clin Invest 1989;83:876–82.

41. Alvaro-Gracia JM, Zvaifler NJ, Brown CB, Kaushansky K,

Firestein GS. Cytokines in chronic inflammatory arthritis. VI.

Analysis of the synovial cells involved in granulocyte-macrophage

colony-stimulating factor production and gene expression in rheu-

matoid arthritis and its regulation by IL-1 and tumor necrosis

factor-alpha. J Immunol 1991;146:3365–71.

42. Harvath L, Brownson NE, Fields GB, Skubitz AP. Laminin

peptides stimulate human neutrophil motility. J Immunol 1994;

152:5447–56.

43. Steadman R, Irwin MH, St John PL, Blackburn WD, Heck LW,

Abrahamson DR. Laminin cleavage by activated human neutro-

phils yields proteolytic fragments with selective migratory

properties. J Leukoc Biol 1993;53:354–65.

140 P Poduval et al

www.scandjrheumatol.dk

Scan

d J

Rhe

umat

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y D

icle

Uni

v. o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.