Retrieval and Preparation of Osteolytic Membranes
The osteolytic interfacial membranes were retrieved during nine revision
total joint arthroplasties according to a protocol approved by the
institutional review boards of the Virginia Commonwealth University Health
System and St. Mary's Hospital. At the time of the harvest of the interfacial
membrane, a small piece of iliotibial band, a fibrous strip of tissue that is
usually cut during the course of a hip replacement, was also excised from two
of the patients to be used as an internal control tissue that is known to
contain fibroblast cells. Patients were selected on the basis of a clinical
presentation and preoperative radiographic findings that were deemed to be
consistent with aseptic loosening. As seen on preoperative plain radiographs,
a lytic area was defined as an area of lucency of >2 mm. Intraoperatively,
indolent infection was ruled out by means of pathological evaluation of
fresh-frozen specimens and by bacterial culture. None of the cases of aseptic
loosening were associated with evidence of infection.
Surgical samples were chilled to 4°C and transported to our orthopaedic
research laboratory. Specimens for immunohistochemical analysis were frozen in
OCT (Optimal Cutting Temperature) embedding medium (Sakura Finetek USA,
Torrance, California) and oriented to allow sections through the
prosthesis-bone interface of the membrane. Tissue for Western blot analysis
was frozen to —70°C. Sixteen-micrometer cryostat sections (HM 500 M;
MICROM International, Waldorf, Germany) were placed onto charged slides
(Superfrost Plus; Fisher Scientific, Pittsburgh, Pennsylvania) at two sections
per slide.
Immunohistochemical Analysis
Slides were thawed at room temperature and then washed twice with
Tris-buffered saline solution. Small wells were created on each slide by using
an ImmEdge Pen (Vector Laboratories, Burlingame, California). Cells were fixed
with ice-cold methanol and then rinsed twice with Tris-buffered saline
solution. Tissues were permeabilized and blocked by adding a solution of 0.1%
Triton X-100 and 3% normal goat serum (Vector Laboratories) in Tris-buffered
saline solution and then washed twice with Tris-buffered saline solution.
Slides were then incubated with primary antibody for one hour. A goat
polyclonal RANKL antibody (Santa Cruz Biotechnology, Santa Cruz, California),
a mouse monoclonal antibody to prolyl 4-hydroxylase (5B5; DakoCytomation
California, Carpinteria, California), and rabbit polyclonal antibodies to
RANK, OPG, and the fibroblast intercellular adhesion molecule-1 (ICAM-1)
(Santa Cruz Biotechnology) were used at a 1:50 dilution. Macrophages were
visualized with use of a CD-163 antibody (Ber-MAC3; DakoCytomation California)
at a 1:20 dilution. The slides were washed twice with Tris-buffered saline
solution and the following fluorophore-conjugated secondary antibodies were
incubated for one hour at a dilution of 1:200 with either chicken anti-rabbit
Alexa Fluor 488 (green), chicken anti-rabbit Alexa Fluor 594 (red), donkey
anti-goat Alexa Fluor 488, or chicken anti-mouse Alexa Fluor 594 (Santa Cruz
Biotechnology). Slides were washed twice with Tris-buffered saline solution. A
fluorescent DNA-binding stain (Hoechst 34580; Molecular Probes, Eugene,
Oregon) was used at a concentration of 5 µg/mL and incubated for thirty
minutes to visualize nuclei. The slides were then washed twice with
Tris-buffered saline solution. Coverslips were applied with an anti-fading
medium (VECTASHIELD; Vector Laboratories). Slides were viewed and images were
stored with use of a laser scanning confocal microscope (model LSM 520 Meta;
Carl Zeiss, Thornwood, New York). Each tissue was excited with use of 405,
488, and 543-nm lasers. The tissue was excited separately by each laser to
eliminate any possible cross-excitation. All images were obtained with an
optical planar slice of 1 µm in thickness with use of a 40×
oil-immersion objective. A composite image was made by combining the images
from the three detection filters. In addition, negative controls were stained
with a secondary antibody only and used to verify the staining specificity of
binding of the secondary antibodies.
Western Blots
Interfacial membranes were homogenized with use of a PowerGen 125 Tissue
Homogenizer (Fisher Scientific) and then dissolved in lysis buffer (MPER;
Pierce Biotechnology, Rockford, Illinois). Samples were centrifuged at 20000
× G to separate cell lysates from tissue and cell debris. Aliquots of
soluble protein were measured by Bradford assay. The cell lysate proteins were
heated (at 95° to 100°C) in Laemmli sample buffer (Bio-Rad, Hercules,
California), separated with use of a 10% sodium dodecyl sulfate
(SDS)-polyacrylamide gel electrophoresis, and transferred to a nitrocellulose
membrane (BioRad). After transfer, the membranes were blocked with 5% powdered
nonfat milk in Tris-buffered saline solution-T, pH 8.0 (20-mM Tris-HCl, 137-mM
NaCl, and 0.05% Tween-20) for one hour at room temperature. The blocked
membranes were incubated with primary antibodies for Ber-MAC3, RANK, CD-14,
RANKL, 5B5, ICAM-1, or OPG diluted in 5% nonfat milk in Tris-buffered saline
solution containing 0.05% Tween-20 and incubated overnight at 4°C. The
membranes were washed four times with Tris-buffered saline solution containing
0.05% Tween-20. Horseradish peroxidase-conjugated secondary antibodies (Santa
Cruz Biotechnology) were diluted in 5% nonfat milk in Tris-buffered saline
solution containing 0.05% Tween-20 and incubated for one hour. The membranes
were washed four times with Tris-buffered saline solution containing 0.05%
Tween-20. Binding was visualized with use of enhanced chemiluminescence assay
(ECL; Amersham Biosciences, Piscataway, New Jersey).
Study Demographics
The study comprised nine patients, seven women and two men, with ages
ranging between fifty-eight and eighty-two years. The osteolytic interfacial
membranes were taken during six hip replacements and three knee replacements.
Four were from the left side and five, from the right. Five replacements were
cemented and four were cementless. All articulations were
metal-on-polyethylene. RANKL-positive cell detection in the periprosthetic
membrane was found in all nine patients tested.
Immunofluorescent Confocal Microscopy
The harvested interface tissue samples were handled so that multiple
cryosections could be prepared for multiple independent colocalization
experiments on the same patient's interfacial membrane tissue.
Figure 1 is representative of
the colocalization experiments comparing staining patterns of key fibroblast
markers (5B5 and ICAM-1; Fig. 1,
A) and myeloid markers (Ber-MAC3 and RANK;
Fig. 1, B) in areas of
the membrane containing cells, as demonstrated by nuclear staining. In
addition to being present on fibroblasts, ICAM-1 is a cell adhesion molecule
that has been shown to be present on the cell surface of preosteoclast cells
(osteoblasts)17 as
well as others. The 5B5 antibody is specific for the prolyl 4-hydroxylase
involved in collagen synthesis that is highly expressed in the endoplasmic
reticulum of
fibroblasts18. The
composite staining comparison of ICAM-1 with 5B5 demonstrated strong
colocalization across areas of nuclear staining within the interfacial
membrane (Fig. 1,
A).
Ber-MAC3 Staining
Staining for Ber-MAC3, an antibody to CD-163 protein expressed on
macrophages, was highly colocalized with staining for RANK
(Fig. 1, B). The high
colocalization of RANK and Ber-MAC3 is evident in the composite image, in
which the red image from Ber-MAC3 staining and the green image from RANK
staining combine to make a yellow color in overlapping areas. In contrast,
RANKL and Ber-MAC3 staining were not colocalized in the interfacial membrane;
however, RANKL-positive cells were seen to be in close proximity to
Ber-MAC3-positive cells (as discussed below).
5B5/ICAM-1 and RANKL Staining
Diffuse RANKL staining was detected throughout the membrane, but in every
tissue viewed there were also areas that displayed substantially higher
expression of RANKL than was seen in surrounding areas. These "active
zones" of RANKL expression were distributed throughout the tissue and
consisted of a large number of closely associated cells that were
characterized by their high levels of RANKL expression. The interior of these
active zones showed positive staining for both 5B5 and ICAM-1, as depicted in
two separate patients' interfacial membranes
(Fig. 2, A and C, and 2,
B and D, respectively), suggesting that
fibroblasts are the cell source responsible for RANKL expression. There was
also evidence of 5B5 and ICAM-1 staining apart from staining of the RANKL.
This secondary staining of 5B5 and ICAM-1 was seen in areas of the interfacial
membrane containing cells and in areas without cells as demonstrated by
nuclear staining.
OPG and RANKL Staining
OPG, the soluble decoy receptor for RANKL, was not seen in large amounts in
the soluble form in the membrane. We expected to see little OPG staining
throughout the membrane because the repeated washing of the slides in the
staining procedure would remove the secreted portion of OPG. However, staining
revealed the presence of apparently intracellular OPG in the interfacial
membrane tissues from two separate patients
(Fig. 3, A and
C). The cells containing OPG were highly colocalized with
RANKL. OPG and 5B5 were also present in the same areas of tissue, with OPG
staining being more diffuse (data not shown). The OPG staining associated with
the 5B5 appeared to be less intracellular than that in the images of OPG and
RANKL staining in active zones, indicating the lack of extracellular OPG
staining with regard to RANKL.
Ber-MAC3 and RANKL Staining
Figure 3, B and
D, shows the staining of Ber-MAC3 and RANKL in
interfacial membranes from two separate patients. These results demonstrate
clear separation of the signals for Ber-MAC3 and RANKL. Positively stained
macrophages are depicted as single nucleated cells both in close proximity to
the signal for RANKL and as distinct cells present throughout the interfacial
membrane tissue section. In addition, in contrast to the stain for 5B5 and
ICAM-1, we found no evidence of Ber-MAC3 signal apart from that in cell
nuclei.
Western Blot Analysis
Osteolytic interfacial membrane was also processed for nonquantitative
Western blot analysis. Figure 4
shows results of separate immunoblots of each epitope used in the study. The
size of protein targets in the interfacial membrane was 40 kDa for RANKL, 55
kDa for 5B5, 54 kDa for OPG, a doublet of 77 and 85 kDa for RANK, 90 kDa for
ICAM-1, and 95 kDa for Ber-MAC3. Interestingly, the smaller soluble form of
RANKL (28 kDa) was not detected.
Controls
For confocal images, a small piece of iliotibial band was used as the
control. The iliotibial band tissue showed staining for high numbers of
nucleated cells, and expression of 5B5, ICAM-1, Ber-MAC3, and RANK. There was
minimal RANKL detection in the nucleated portions of the iliotibial bone
tissue and no staining of any active zones was observed (data not shown). In
contrast, definitive staining of RANKL in regions of high nuclear density
(Fig. 5), displayed as strong
RANKL staining from a ring-like cell structure characteristic of the described
active zones, was observed in the interfacial membrane tissues from all nine
patients.
Note: The authors thank Ms. Carol Coats for coordinating the
patient recruitment and consent in the study and the tissue retrieval and
delivery between the involved hospitals and the Virginia Commonwealth
University Orthopaedic Research Laboratory. We also thank Dr. John Cardea for
his assistance in obtaining interfacial membrane tissues at the Virginia
Commonwealth University Health System.