Abstract
purpose. To determine whether the cell adhesion molecule CD44, the principal
receptor of hyaluronan, is altered in the aqueous humor and the
anterior segment of patients with primary open-angle glaucoma (POAG).
methods. The trabecular meshwork (TM), iris, ciliary body, and sclera of POAG
and age-matched control eyes preserved in ethanol were microdissected
and subjected to 1% Triton X-100 solubilization at 4°C. Western blot
analysis was performed using monoclonal antibodies that recognize
either CD44H (hematopoietic; extracellular domain) or CD44S (soluble
ectodomain). The concentration of soluble CD44S in aqueous and
microdissected tissues was measured by enzyme-linked immunosorbent
assay (ELISA).
results. ELISA of soluble CD44S of aqueous from eyes of patients with POAG
indicated that the concentration of soluble CD44S is increased in
comparison with that of aqueous from normal eyes (P < 0.0003). Western blot analysis and densitometry of POAG iris and
ciliary body revealed a statistically significant increase in the
Triton X-100 extraction of CD44H. The predominant increases were in the
180-kDa (P < 0.001) and the 85-kDa
(P < 0.001) forms. ELISA of soluble CD44S
indicated that the concentration is statistically decreased in iris
(P < 0.05), ciliary body (P <
0.001), and TM (P < 0.005) of POAG eyes.
conclusions. Increased amounts of soluble CD44S in POAG aqueous and Triton
X-100–solubilized CD44H characterized POAG in the iris and ciliary
body. These soluble CD44 isoforms may influence the activity of the
transmembrane CD44H by acting as inhibitors of CD44H and, thereby,
adversely influence the cell survival of TM and retinal ganglion cells
in POAG.
Primary open-angle glaucoma (POAG) is a major blinding
disease affecting approximately 67 million persons
worldwide.
1 It is likely that several biochemical and
cellular factors influence the glaucoma process. A variety of cellular
insults or molecular defects
2 3 may intersect, leading
individually or collectively to cell death in the trabecular meshwork
(TM)
4 or retinal ganglion cells.
5 Moreover,
evidence suggests that activated immunity
6 and alterations
in the extracellular matrix
7 8 9 10 may be etiologic factors
in POAG. The extracellular matrix is a diverse group of macromolecules
that assemble to form a functional network. The majority of aqueous
outflow resistance in both normal and POAG eyes is within the TM,
especially within the extracellular matrix of the juxtacanalicular
connective tissue, which is a glycosaminoglycan-enriched
area.
10 11 12 Biochemical studies
10 and
computer-aided image analysis
11 indicate that POAG is
associated with a marked decrease in hyaluronan (HA) and an increase in
chondroitin sulfate in the TM. HA is a key factor in promoting cell
motility, adhesion, and proliferation.
13 14 These cell
events are orchestrated by three HA
15 cell
receptors—CD44, receptor for HA-mediated motility (RHAMM), and
intercellular adhesion molecule (ICAM)-1—all of which have been
identified in the human TM.
16
One receptor for HA, CD44, is a multifunctional receptor and
cell-adhesion molecule that increases with the aging of T
lymphocytes.
17 18 CD44H is an integral cell membrane
glycoprotein with postulated roles in a wide variety of biological
processes, including cell adhesion,
19 inflammation,
20 autoimmunity,
21 and
apoptosis.
22 CD44 exists as both an 80- to 90-kDa type 1
transmembrane glycoprotein, CD44 hematopoietic (CD44H), and a 30- to
50-kDa soluble form, soluble CD44S, generated by the release of
the extracellular domain by hydrolytic cleavage.
23 Hydrolysis of CD44H involves either the cleavage of
glycosylphosphatidylinositol (GPI)-anchor and/or the limited
proteolysis of the extracellular domain. These soluble CD44 isoforms
may regulate the effects of cognate ligands of CD44H by acting as
soluble inhibitors of CD44H. In addition, the proteolytic cleaved
soluble CD44S and the loss of the GPI-anchor of CD44H probably
constitute, at least in part, a turnover mechanism for downregulating
the CD44 receptor.
24
CD44H has diverse functional properties owing to sequence differences
arising from alternate splicing of mRNA, as well as extensive
posttranslational glycosylation.
25 CD44H is expressed on a
variety of ocular cells, including retinal ganglion
cells
26 and axons.
27 CD44S is present in
serum
28 and aqueous humor.
29 In vitro, axon
growth of retinal ganglion cells is inhibited by the presence of
CD44H.
30 Aging leads to replacement of virgin T cells by
memory T cells and to the accumulation of cells with signal
transduction defects.
31 Aged CD44H-positive cells are less
responsive to antigenic stimuli.
32 Yonemura et
al.
33 demonstrated that CD44H is the membrane-binding
partner for ezrin-radixin-moesin (ERM) proteins. The CD44-ERM complex
acts as a regulatable protein scaffold that anchors actin filaments to
the plasma membrane. Consequently, increased turnover of CD44H in POAG
may disrupt several normal cell functions and adversely affect TM cell
function
34 and survival. Therefore, we determined the
concentration of soluble CD44S in the aqueous humor of normal subjects
and patients with POAG in comparison with CD44H and soluble CD44S in
donor eyes, both of which appear to characterize POAG. Our data support
the hypothesis that CD44H may represent a protein marker of the
disease.
34
Samples of aqueous humor were obtained from normal subjects
(n = 41; age range, 46–85 years; mean, 72.5 ±
8.6) who were undergoing elective cataract surgery, patients with
glaucoma (n = 29; age range, 46–85 years; mean,
71.5 ± 9.8) who were undergoing elective cataract surgery
(n = 9) or filtration surgery (n = 20),
and one patient with ocular hypertension who was undergoing elective
cataract surgery. None of the subjects had had incisional ocular
surgery. Seventy microliters of aqueous humor was aspirated by limbal
paracentesis. Patients provided informed consent after the nature and
consequences of the study were explained, in accordance with the
Declaration of Helsinki, and the research was approved by institutional
review boards.
Six normal eyes from three donors were obtained from the
Illinois Eye Bank and the Lions Eye Bank of Washington and Northern
Idaho. Six POAG eyes from three donors were obtained from the Rochester
Eye and Human Parts Bank, Inc., the Medical Eye Bank of Florida, and
the Michigan Eye Bank and Transplantation Center through the efforts of
the Foundation for Glaucoma Research (San Francisco, CA). After
enucleation, the eyes were either immersed in alcohol or refrigerated
and sent to the laboratory and placed in 100% ethanol within 12 hours
of death. The anterior segment was microdissected.
10 11 The mean age of the normal donor eyes was 72.9 years, and the mean age
of the POAG donor eyes was 67.0 years. None of the eyes had evidence of
uveitis or incisional ocular surgery. The available clinical
histories of the normal and POAG donor specimens are shown in
Table 1 .
The microdissected anterior segment tissues were incubated for 2 hours
at 4°C with end-over-end mixing in 1 mL 20 mM Tris (pH 7.5); 0.15 M
NaCl-1% Triton X-100
35 containing the inhibitors 1 mM
EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine HCl, 10 mM
N-methyl maleimide, 100 mM 6-amino hexanoic acid, 50 mM
iodoacetamide, and 1 μg/mL leupeptin; and 10 μg/mL soybean
trypsin inhibitor. The resultant homogenate was clarified by
centrifugation at 10,000
g for 10 minutes, and the
supernatant was stored at −80°C. The pellet was resuspended in 1 mL
of the above-mentioned buffer with the addition of 0.5% sodium
deoxycholate (Doc) and 0.1% sodium dodecyl sulfate (SDS) to further
extract CD44H.
35 The solubilization was performed for 2
hours at 4°C with end-over-end mixing and clarified by centrifugation
at 10,000
g for 10 minutes, and the supernatant was stored at−
80°C.
Protein concentration was determined using bovine serum albumin
as a standard (Bio-Rad Laboratories, Hercules, CA) according to the
manufacturer’s protocol. Five μg of aqueous protein or 20 μg of
microdissected tissue protein from each of the Triton X-100 extracts
and the Doc-SDS extracts were electrophoresed on 10% SDS-PAGE,
transferred to nitrocellulose membranes (Immobilon-P; Millipore,
Bedford, MA), and immunoblotted with 1:500 dilution anti-CD44H antibody
(R&D Systems, Inc., Minneapolis, MN) or with 1:100 dilution anti-CD44S
antibody (Biosource International, Camarillo, CA). The immune complex
was visualized by an enhanced chemiluminescence (ECL) detection system,
according to the protocol supplied by the manufacturer (Amersham
Pharmacia Biotech, Arlington Heights, IL), using the appropriate
horseradish peroxidase (HRP)–conjugated secondary antibody (1:2500
dilution), and was quantified according to a densitometry scale of 0,
no detectable product; 1+, trace; 2+, positive; 3+, strongly positive;
and 4+, intensely positive.
36 37
Soluble CD44S was measured in the aqueous humor from patients,
the Triton X-100 extract, and the Doc-SDS extract, using the standard
commercially available CD44 soluble ELISA kit (Bender Med Systems,
Vienna, Austria) as described by Asplund and Heldin.
35 Briefly, 1 μg protein equivalent of each sample of aqueous, Triton
X-100 extract, and Doc-SDS extract protein was subjected to 1:60
dilution in the sample dilution buffer; 20 μL of the diluted sample
and 80 μL of the sample diluent buffer were added to each well. The
samples were treated with 50 μL of diluted HRP-CD44S antibody
conjugate and incubated for 3 hours at room temperature. Each well was
washed three times with buffer, and 100 μL freshly prepared
tetramethyl-benzidine substrate buffer was added and incubated for 15
minutes at room temperature. Color development was stopped by adding
100 μL of 4 N H
2SO
4; the
optical density was read in a ELISA reader at 450 nm. The concentration
of soluble CD44S was determined by the standard curve, using the
manufacturer’s standard CD44S soluble protein.
CD44H and CD44S immunoblots of normal and POAG tissues were
performed in duplicate, and CD44S ELISA assays were performed in
triplicate. For validation of the standard curve, the slope, intercept,
and correlation coefficient were within the manufacturer’s suggested
ranges. Data are presented as mean ± SEM. All statistical
analyses were conducted with Student’s t-test. P < 0.05 was considered the level of significance.
In 41 normal aqueous samples the soluble CD44S concentration was
9.58 ± 0.79 ng/mL (SEM), whereas in 26 POAG aqueous samples the
soluble CD44S concentration was 15.96 ± 1.37 ng/mL (an
approximately twofold increase,
P < 0.0003;
Table 2 ). Western blot analysis of aqueous humor showed three immunoreactive
bands: a predominant 31-kDa band, the soluble form of CD44, which was
markedly increased in the POAG aqueous, a minor 55-kDa band, another
soluble form of CD44, and an 85-kDa band
(Fig. 1) . The aqueous humor profile of immunoreactive bands was distinct from
that of the positive control, serum, which has 60- to 80-kDa and 100-
to 150-kDa bands (data not shown).
Donor Eyes and Triton X-100 Extraction: CD44H in the Iris and in
the Ciliary Body
In the iris and ciliary body microdissected tissues, Triton X-100
extracted proteins were identified on immunoblots with an anti-CD44H
monoclonal antibody that recognizes the extracellular domain of CD44. A
CD44H-positive protein of an apparent molecular weight of 85 kDa was
identified in almost all cases
(Fig. 2A 2B) . In addition, heterogenous high-molecular-weight proteins were
observed, typical of glycosylated CD44H.
32 Immunoblots of
the iris and ciliary body revealed a distinctive pattern in POAG eyes
compared with normal eyes. All six cases of POAG but only one of the
normal eyes, 3L, demonstrated a marked increase in 180-kDa and 85-kDa
CD44H proteins in the iris and in the ciliary body
(Figs. 2A 2B) .
Analysis of the intensity of CD44H immunostaining between the POAG and
normal eyes was highly significant for the 180-kDa protein
(
P < 0.001) and for the 85-kDa protein
(
P < 0.001). The CD44H immunoblot pattern of iris and
ciliary body was correlated with the available clinical histories
(Table 1) . In eyes from donor 2, the most advanced glaucoma, was the
most positive for CD44H; eyes from donor 3, with moderate glaucoma,
were moderately positive; those from donor 1, early glaucoma, were
slightly positive.
A comparison of normal and POAG immunoblots of the TM and sclera
was unrevealing
(Figs. 2C 2D) ; the intensity of the CD44H profiles was
statistically insignificant. There was a large amount of CD44H in the
TM in the most advanced POAG (in donor eye 2R but less in 2L, which
recently had undergone laser trabeculoplasty), a moderate amount of
CD44H in the eyes with a less-advanced form (3R and 3L), and a minimal
amount in an early diagnosed POAG (eyes 1R and 1L).
To test whether the Triton X-100 extracted all the CD44H, we
subjected the Triton X-100-insoluble material to extraction with 0.5%
Doc-0.1% SDS
(Fig. 3) . In both the normal and POAG tissues, Doc-SDS extracted additional
amounts of CD44; however, results of Doc-SDS extraction and Western
blot analysis were similar for normal and POAG and statistically
insignificant in all tissues.
To explore whether a CD44S monoclonal antibody, which
preferentially recognizes soluble CD44S, was useful in distinguishing
normal and POAG tissues, we used Triton X-100 and Doc-SDS extracts of
all tissues on immunoblots. The CD44S-positive proteins were
predominantly 55, 67, and 85 kDa, which was distinctly different from
the CD44H immunoblots (compare
Figs. 2 and 4 ). There were no differences between normal and POAG tissues when CD44S
was used on immunoblots
(Fig. 4) .
The concentrations of tissue CD44S, as determined by ELISA, are
shown in
Figure 5 for iris, TM, ciliary body, and sclera. The POAG ciliary body was
significantly decreased (range, 150–200 pg/mg) in comparison with the
normal ciliary body (range, 300–400 pg/mg;
P <
0.001). The POAG iris also contained lower concentrations of CD44S than
did the normal iris (
P < 0.05). The POAG TM was
statistically different from the normal TM. The range of in normal TM
was 80 to 120 pg/mg whereas the range in POAG was 10 to 80 pg/mg
(
P < 0.005). There was no significant difference in
laser- versus non–laser-treated TM. There was no statistically
significant difference between normal and POAG in the concentration of
CD44S in sclera.
The purpose of our study was to determine the concentration of
soluble CD44S in aqueous humor from normal subjects and patients with
POAG and CD44H in microdissected anterior segment tissues of normal and
POAG-affected donor eyes. Although the multifunctional CD44 protein is
widely distributed, its presence, absence, or alteration may represent
a protein marker of POAG.
34 The ectodomain of CD44H is
shed as soluble CD44S, rather than internalized, and this shedding of
CD44S may allow for the downregulation or renewal of the CD44H
receptor.
21 24 In addition, soluble CD44S may have
regulatory functions.
24
The concentration of soluble CD44S was 9.58 ng/mL in normal aqueous
humor compared with 15.96 ng/mL in POAG. The concentration of soluble
CD44S increased with age. Linear regression analysis showed that
soluble CD44S increased approximately 0.075 ng/mL per year in normal
aqueous and approximately 0.12 ng/mL per year in POAG aqueous. In eyes
at age 50, the normal projected soluble CD44S was approximately 7.0
ng/mL; by age 80, it was 9.25 ng/mL. The data for patients with POAG
indicate that at age 50 the projected soluble CD44S was 12.5 ng/mL and
by age 80 was 16.2 ng/mL in this study. It is quite possible that
within the 41 normal aqueous specimens a few individuals may have
undiagnosed POAG or will manifest POAG in the future. It appears that
the soluble CD44S concentration in POAG is higher than that in juvenile
open-angle glaucoma. Notably, the soluble CD44S concentration in
patients in whom filtration surgery was unsuccessful was higher than in
patients with successful filtration surgery.
Soluble CD44S is shed as the result of ligand
binding,
24 32 alternate splicing,
38 and
enzymatic cleavage from the cell surface.
24 The source of
soluble CD44S is presumably the ciliary body–iris epithelium, because
these tissues have the highest concentrations of CD44H and are known to
release proteins.
39 ELISA analysis of the soluble CD44S of
POAG iris, ciliary body, and TM showed a statistically significant
decrease; however, ELISA of CD44S in aqueous showed a statistically
significant increase in soluble CD44S, which may indicate that soluble
CD44S is shed into the aqueous.
The ectodomain shedding of proteins from the cell surface is common.
Many proteins lead a dual existence as both integral membrane proteins
and soluble proteins; the classic examples are immunoglobulin and
acetylcholinesterase.
23 Alternate processing of each
produces membrane-bound or soluble forms, each form having distinct
physiologic roles.
40 CD44H also has a dual existence.
Soluble CD44S is shed into aqueous and may act as a signaling molecule
that may influence target cells, both in the anterior and the posterior
segments. It is interesting that increased concentrations of soluble
CD44S have been found in the sera of patients with autoimmune
diseases.
41 If CD44S acts as a decoy receptor, it would
interfere with normal ligand-induced signal transduction by membrane
CD44H.
42 Notably, the shedding of CD44S participates in
turnover and remodeling of the extracellular matrix and cell
surface,
43 and in binding studies of CD44S, the shed
soluble CD44S retains its biologic activity.
24 44
Western blot analysis in POAG demonstrated a marked increase of the
Triton X-100–extracted CD44H in the iris and ciliary body. Western
blot analysis of the iris and ciliary body in all six cases of POAG
demonstrated a statistically significant increase in CD44H—the 85- to
180-kDa proteins. Western blot analyses using the CD44S antibody were
insignificant but confirmed the specificity of the antibody, which
preferentially recognized the lower molecular weight CD44S. Triton
X-100, a nonionic detergent, is useful in solubilizing a number of
integral membrane proteins,
45 such as
CD44H.
46 Triton X-100 binds to the hydrophobic domains of
proteins without disrupting protein–protein interactions. A
20-amino-acid hydrophobic region in the transmembrane portion of CD44H
is an absolute requirement for detergent solubility.
46 47 Because the transmembrane portion of CD44H associates with phospholipid
microdomains, the activity of the cell determines the Triton X-100
solubility of CD44H.
48 In POAG, the predominant increase
in the Triton X-100-solubilized CD44H was in the 180- and 85-kDa forms
of CD44, the membrane CD44.
Computer-aided color image analysis of normal and POAG sections
analyzed by immunostaining with CD44H antibody clearly separated
individual cases of POAG from the normal.
34 The sections
were treated with Triton X-100, and CD44H was extracted. CD44H is an
unusual cell receptor, having an unusually low diffusion
coefficient.
49 50 Thus, the difference in the solubility
of CD44H in POAG eyes on tissue sections or microdissected tissue may
relate to a change in the associated GPI anchor in the cell membrane
and/or cell activity.
CD44H promotes the migration of fibroblasts through interactions with
its ligands in the matrix. In addition to exhibiting influence over
migration, CD44H influences cell survival.
51 Previous
studies have shown that fibroblasts undergo apoptosis after anti-CD44
antibody treatment.
52 In addition to anti-CD44
antibodies’ inducing fibroblast apoptosis, soluble CD44S has been
shown to impair tumor metastasis by inhibiting the function of CD44H as
a cell-adhesion molecule and causing mammary carcinoma cells to undergo
apoptosis.
53 The ability of soluble CD44S to induce
apoptosis by binding to membrane-bound CD44H elicits the hypothesis
that elevated levels of soluble CD44S will be accompanied by programmed
cell death. Shed soluble CD44S binds to membrane-bound CD44H and
abrogates the CD44 receptor.
Thus, CD44H is a multifunctional receptor involved in
cell–cell
54 55 and cell–matrix
interactions,
56 cell trafficking, presentation of
chemokines and growth factors, and transmission of growth
signals.
57 An increased concentration of soluble CD44S in
POAG may block the binding of CD44H to HA.
58 Also CD44H is
required for certain high-affinity receptors, e.g., erbB2
phosphorylation and erbB2-erbB3 heterodimerization, for cell
survival.
51 If soluble CD44S interferes with CD44H
activity, then it is downregulated and erbB2 is less active, and this
causes programmed cell death. Thus, our working hypothesis is that POAG
is characterized biochemically by decreased concentration of HA and
increased turnover and downregulation of the HA receptor CD44, which,
in turn, may influence cell survival of TM and retinal ganglion cells.