Hypophosphorylation of Aqueous Humor sCD44 and Primary Open-Angle Glaucoma

Abstract

purpose. The ectodomain of CD44, the principal receptor for hyaluronic acid (HA), is shed as a 32-kDa fragment—soluble CD44 (sCD44)—which is cytotoxic to trabecular meshwork (TM) cells and retinal ganglion cells (RGCs) in culture. The purpose of this study was to characterize sCD44 further by determining the phosphorylation of aqueous humor sCD44 in normal and primary open-angle glaucoma (POAG).

methods. Aqueous humor samples of patients were subjected to CD44 enzyme-linked immunosorbent assay (ELISA) and two-dimensional (2-D) polyacrylamide gel electrophoresis, followed by Western blot analysis with anti-CD44, anti-serine/threonine, and anti-tyrosine phosphospecific antibodies, to determine sCD44 concentration, isoelectric point (pI), and phosphorylation, respectively. The bioactivity of hypophosphorylated sCD44 was tested in cell culture and HA affinity columns.

results. Two-dimensional Western blot analysis revealed that the representative pI of the 32-kDa sCD44 was 6.96 ± 0.07 in POAG versus 6.38 ± 0.08 in normal (P < 0.0004). Enzymatic dephosphorylation of sCD44 resulted in a basic shift in the pI. The normal aqueous humor sCD44 was positive for serine-threonine phosphorylation; however, POAG sCD44 was hypophosphorylated. Hypophosphorylated sCD44 was more toxic to TM and RGC cells than standard sCD44, and hypophosphorylated sCD44 had decreased affinity to HA, particularly with increased pressure.

conclusions. POAG aqueous is characterized by posttranslational change in the pI of sCD44 and hypophosphorylation, which clearly distinguished POAG from normal aqueous humor. The high toxicity and low HA-binding affinity of hypophosphorylated sCD44 may represent specific pathophysiologic features of the POAG disease process.

Primary open-angle glaucoma (POAG) is a common ocular neurodegenerative disease characterized by retinal ganglion cell (RGC) death, axon loss, and an excavated appearance of the optic nerve head, and POAG is often associated with elevated intraocular pressure (IOP). ¹ POAG is probably caused by a variety of cellular insults leading individually or collectively to cell death in (1) the trabecular meshwork (TM), which results in elevated IOP, ² and (2) RGCs, which results in visual field loss. Lowering IOP by topical medications or surgery decreases the risk of visual field loss in patients with ocular hypertension ³ or with glaucoma. ^{4 5}  

Alterations in the extracellular matrix, particularly the glycosaminoglycans (GAGs), play a causative role in POAG. ^{6 7} Our biochemical studies have indicated that POAG is associated with changes in two types of glycosaminoglycans in the TM: a marked decrease in hyaluronic acid (HA) and an increase in chondroitin sulfate. ^{7 8} One receptor for HA is CD44H, a type I transmembrane protein and member of the cartilage link protein family. ^{9 10 11} We have reported that CD44H is present in POAG ¹² and that the ectodomain of CD44, ¹³ —soluble CD44 (sCD44)—is significantly increased in POAG. Moreover, sCD44 is cytotoxic to TM cells and RGCs in cell culture. ¹⁴  

sCD44 is released from the cell surface by MT1-MMP, a membrane-bound metalloprotease. ¹⁵ sCD44 has different biological functions than does the intact CD44H protein. ⁹ The intramembrane portion is cleaved by a presenilin-dependent γ-secretase at two sites: One cleavage occurs close to the cytoplasmic border to release an intracytoplasmic fragment that translocates to the nucleus and promotes transcription, and a second cleavage site within the transmembrane domain generates a peptide similar to the amyloid β-peptide implicated in Alzheimer’s disease. ¹⁶  

In the present study, we identified a more basic, hypophosphorylated sCD44 that is also present in POAG aqueous humor. The hypophosphorylated form of sCD44 was more toxic to TM and RGC-5 cells. 

Materials and Methods

Aqueous Samples

Aqueous samples were obtained from normal patients and from definite glaucoma patients who were undergoing cataract surgery or filtration surgery, as previously described. ¹³ Exclusion criteria included prior incisional ocular surgery and diabetes mellitus. Patients provided informed consent after the nature and consequences of the study were explained, in accordance with the tenets of the Declaration of Helsinki, and the research protocols were approved by institutional review boards of each of the collaborators. 

ELISA, Two-Dimensional Gel Electrophoresis and Western Blot Analysis of sCD44

The protein concentration of aqueous humor was determined with a protein assay (Bio-Rad, Hercules, CA), and a 1-μg aliquot was used to measure sCD44 concentration by ELISA (Bender Med Systems, Vienna, Austria), as previously described. ¹³ Two-dimensional (2-D) polyacrylamide gel electrophoresis (PAGE) was performed with 5 μg protein. The first and second dimensions were performed according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, and Bio-Rad, respectively), and the method was verified by a carbamylyte calibration kit (Amersham Bioscience, Piscataway, NJ). Western blot analysis was performed with anti-CD44 antibody (BU52; Ancell, Bayport, MN) to determine the isoelectric point (pI) of aqueous humor sCD44. Samples were treated with 8 M urea by using immobilized pH gradient strips containing an ampholyte solution (pH 3–10), loaded, and run at 200 V for 20 minutes, 450 V for 15 minutes, 750 V for 15 minutes, and 2000 V for 30 minutes. After the first dimension, the gel was equilibrated in sample buffer and then subjected to the second dimension on 4% to 15% SDS-PAGE in reducing conditions at 100 V for 60 minutes. The gels were transferred to nitrocellulose membranes, incubated with a mouse anti-CD44 antibody (1:500 dilution), incubated with a goat anti-mouse-horse radish peroxidase (HRP) conjugate (1:3000 dilution; Bio-Rad), and visualized on x-ray film by enhanced chemiluminescence (ECL), according to the manufacturer’s instructions (Amersham, Arlington Heights, IL). All 2-D gel electrophoreses were reproduced at least twice with identical results. The staining intensity was quantified to a densitometry scale of 0, no detectable product; 1+, trace; 2+, positive; and 3+, strongly positive for each pH interval and integrated to obtain the average pI. ¹² Data are expressed as the mean ± SEM and were analyzed by nonparametric Mann-Whitney test. 

Purification of 32-kDa sCD44 from Human Serum, Immunoprecipitation of sCD44 from Aqueous Humor, and sCD44 Phosphorylation

sCD44 was purified from 200 mL of human sera (Sigma-Aldrich, St. Louis, MO), as previously described. ¹⁴ Aliquots of 200 μL of pooled normal (n = 6; cup-disc ratio 0.4 ± 0.06; normal visual fields normal) and pooled POAG aqueous humor (n = 6; cup-disc ratio 0.86 ± 0.04; moderate to advanced visual field loss) were diluted 1:1 with PBS and incubated with mouse anti-CD44 monoclonal antibody (BU52, 1:50 dilution; Ancell) overnight at 4°C. The immune complexes were incubated with anti-mouse antibody conjugated with agarose (1:10 dilution; Sigma-Aldrich) at room temperature for 2 hours. The immune complex immobilized with agarose was centrifuged at 500g for 3 minutes. The resultant precipitate was rinsed three times with PBS and centrifuged at 500g for 3 minutes. The bound proteins were incubated with 0.2 M glycine (pH 2.2) for 5 minutes, to elute the protein, and centrifuged at 500g for 3 minutes. The resultant supernatant was neutralized with 1 M Tris-HCl (pH 7.8). 

To test the functional status of sCD44, 40 ng of sCD44 isolated from human serum, as determined by ELISA, was subject to hypo- and hyperphosphorylation in vitro. sCD44 was hypophosphorylated by incubation with 30 units alkaline phosphatase (MP Biochemicals, Irvine, CA) for 1 hour at 37°C and repeated. ¹⁷ sCD44 was hyperphosphorylated by treatment with 100 ng casein kinase II (CK II; Upstate Biotechnology, Charlottesville, VA) and 10 μCi of ATP (MP Biochemicals) for 1 hour at 37°C. ¹⁸  

To compare the sCD44 phosphorylation state in normal and POAG aqueous humor, sCD44 immunoprecipitated from pooled normal and pooled POAG aqueous humor. An aliquot of 20 pg of sCD44, as determined by ELISA, was diluted 1:1 in reducing SDS sample buffer, loaded on a 4% to 15% linear gradient gel and subjected to PAGE. Separated proteins were then transferred onto nitrocellulose membranes. After nonspecific sites were blocked with 10% nonfat dry milk (Bio-Rad), the nitrocellulose membranes were incubated CD44 antibody to verify the equal loading of sCD44 and with a battery of phosphospecific antibodies (Table 1) . The membranes were then incubated with the appropriate secondary HRP-conjugated antibody. Finally, sCD44 phosphorylation was visualized on x-ray film using an enhanced chemiluminescence system according to the manufacturer’s instructions (Amersham). 

To test whether the difference in sCD44 phosphorylation between normal and POAG aqueous humor was due to CK II, 20-pg loads of normal and POAG sCD44 were treated with 50 ng of CK II in the presence of 50 ng of ATP (MP Biochemicals). As a control, this treatment was also performed in the presence of either 50 μg heparin or 960 ng DRB (Sigma-Aldrich), two well-established inhibitors of CK II. ^{19 20} CK II-treated sCD44 was then analyzed by SDS-PAGE by Western blot analysis with the anti-phosphoserine/threonine antibody. 

Cell Culture and sCD44 Treatment

Human TM ²¹ cells derived from three different donors and a transformed rat RGC-like cell line (RGC-5) ^{22 23} were grown in culture, as previously described, ¹⁴ and treated with 0.1 and 10 ng/mL of purified sCD44, hyperphosphorylated sCD44, and hypophosphorylated sCD44 for 24 and 48 hours. The cells were then counted with a Coulter counter and stained with trypan blue, to assess viability, as previously described. ¹⁴ The cell viability was expressed as the absolute number of viable cells in the experimental group (experimental percentage of cell viability × the total number of experimental cells) versus the absolute number of viable cells in the control group (control percentage of cell viability × the total number of control cells). Data are expressed as the mean ± SD and were analyzed by Student’s t-test. P < 0.05 was considered statistically significant. 

HA Preparation, Standard and Hypophosphorylated sCD44, and HA Affinity Columns

Human umbilical cord HA was standardized by size-exclusion chromatography on a Sepharose CL-4B column (1.6 × 22 cm; Amersham Pharmacia Biotech, Piscataway, NJ), equilibrated with 0.1 M ammonium acetate and standardized by the elution of blue dextran (MW = 2 × 10⁶; Sigma-Aldrich). EAH-Sepharose (2 mL; Amersham Pharmacia Biotech) was rinsed with 40 mL PBS in a sintered glass filter to remove azide from the storing buffer, loaded into a disposable minicolumn (Bio-Rad), and packed by using PBS. Seven milliliters of 50% glutaraldehyde-PBS solution was added, and the mixture was incubated for 10 minutes by end-over-end mixing, and the column was washed with 100 mL PBS. HA (0.5 mg) was dissolved in 7 mL PBS, added to the column, and incubated for 15 minutes while mixing. The solution was then passed through the column, and 100 mL PBS was used to wash the column. HA that did not bind to the EAH-Sepharose was removed from the column with 10 mL of 3 M ammonium thiocyanate. Ten milliliters of 0.2 M glycine (pH 8.5) was used as a blocking agent. The column was rinsed with 100 mL PBS. The columns were sealed at either 0 mm Hg (atmospheric pressure) or at 40 mm Hg, created by a PBS-filled pressure reservoir 54 cm above the column bed. sCD44 (0.5–5 ng) purified from human sera, was injected into the column and allowed to permeate the column bed. The column was incubated on a magnetic stirrer for 16 hours at 4°C at either 0 or 40 mm Hg. The column was rinsed with 15 mL PBS, and the unbound fraction was collected. Equal volumes of 0.5 M NaCl PBS and 0.2 M glycine (pH 2.5) were used to elute the bound sCD44 from the HA. The fractions eluted with glycine were neutralized with 0.1 M Tris-HCl (pH 7.5). Fractions were then concentrated with 30-kDa centrifugal concentrators (MWCO; Vivascience, Hannover, Germany) at 5400g for 4 minutes. The concentrated fractions were analyzed by ELISA for sCD44 concentration. Scatchard plots were constructed from the results to obtain the dissociation constant and maximum binding of standard and hypophosphorylated sCD44. 

Rotary Shadowing of HA

A 5-μL drop of HA (0.1 g/mL) in 0.5 M ammonium acetate buffer (pH 7.2) or in 0.2 M glycine (pH 2.5) ²⁴ was placed on a collodion-coated, 300-mesh grid; air dried; and rotary shadowed with gold palladium at a glancing angle of 4°, according to the method of Scott et al. ²⁵ A pressure chamber with access ports was constructed to maintain pressure at 40 mm Hg so that the 5-μL drop of HA on the collodion-coated, 300-mesh grid could be air dried under pressure. Each experiment was performed in triplicate with similar results. The HA preparation was visualized with an electron microscope (1200 EX; JEOL, Tokyo, Japan). 

Results

Isoelectric Variants of Aqueous Humor sCD44

sCD44 was immunoprecipitated from aqueous humor and analyzed by 2-D PAGE (Fig. 1) . In normal aqueous humor samples, the apparent pI of the 32-kDa sCD44 was 6.38 ± 0.08 with an isoelectric variance of 5.4 to 7.0 (Table 2) . In contrast, in POAG aqueous humor samples, the apparent pI of sCD44 was 6.96 ± 0.07 (P < 0.0004), with an isoelectric variance of 5.4 to 8.6; and, in normal-pressure glaucoma (NPG) aqueous humor samples, the apparent pI was 6.76 with an isoelectric variance of 5.2 to 8.6. In juvenile open-angle glaucoma (JOAG) and exfoliation glaucoma aqueous humor samples, the apparent pIs of sCD44 were 6.29 and 6.40, respectively. The clinical status and glaucoma medications of the patients are shown in Tables 3 and 4 . The apparent pI and presence of isoelectric variants of the 32-kDa sCD44 protein appeared to distinguish POAG aqueous humor from that of patients without glaucoma or with other clinical forms of glaucoma. 

sCD44 Phosphorylation

To determine the possible mechanisms for a more basic pI in POAG aqueous humor sCD44, pooled normal and pooled POAG aqueous humor samples were immunoprecipitated with anti-CD44 antibody and immunoblotted with phosphospecific anti-serine/threonine antibodies (Fig. 2A) . The sCD44 concentration of the pooled normal aqueous humor was 5.55 ± 3.61, whereas the sCD44 concentration of the pooled POAG aqueous humor was 14.00 ± 2.63 (P < 0.001). The clinical status and glaucoma medications of the patients are shown in Table 4 . Western blot analysis showed that 32-kDa sCD44 from normal aqueous humor was strongly immunopositive for serine and threonine phosphorylation. In marked contrast, 32-kDa sCD44 from POAG aqueous humor was less immunopositive for serine and threonine phosphorylation, which indicates that POAG sCD44 is hypophosphorylated in comparison with normal sCD44. Both normal and POAG aqueous humor were negative on Western blot with anti-tyrosine phosphospecific antibody. sCD44 phosphorylation was also studied with phosphospecific antibodies for PKA, PKB, PKC, and MAPK motifs, and the Western blot analysis indicated that the phosphorylation status of these motifs was similar in normal and POAG. 

To explore further the observation of hypophosphorylation of sCD44 in POAG aqueous humor, pooled aqueous humor of normal and POAG samples was phosphorylated with CK II in vitro. To verify the specificity of CK II phosphorylation, aliquots were also treated with CK II in the presence of specific CK II inhibitors. The CK II-treated aqueous humor samples were immunoblotted with phosphospecific anti-serine/threonine antibodies (Fig. 2B) . After CK II phosphorylation of sCD44, the immunoblots of POAG were identical with normal, indicating that sCD44 is hypophosphorylated in POAG aqueous humor. 

To examine the biological significance of sCD44 hypophosphorylation in POAG, we isolated the sCD44 from human sera and modified its phosphorylation state in vitro (Fig. 3) . sCD44 was hypophosphorylated with alkaline phosphatase and also hyperphosphorylated with CK II. In standard sCD44, the apparent pI was 6.15 with a pH range of 5.65 to 6.15. In hypophosphorylated sCD44, the apparent pI was 6.25, with a pH range of 5.3 to 6.7, whereas in the hyperphosphorylated sCD44, the apparent pI was 6.05, with a pH range of 5.6 to 6.2. The basic shift caused by alkaline phosphatase treatment led to a more basic pI, as observed in sCD44 in POAG and NPG aqueous humor. 

sCD44 Phosphorylation and Cell Viability

To determine whether phosphorylation of sCD44 influences cell viability, human TM cells were treated with standard, hypophosphorylated, or hyperphosphorylated sCD44 (Fig. 4) . After 24 hours, the cell viability of human TM cells treated with 10 ng/mL of standard sCD44 was 66% (P < 0.0002), and the cell viability of human TM treated with 10 ng/mL of hypophosphorylated sCD44 was 58% (P < 0.0002). In contrast, hyperphosphorylated sCD44 was less toxic to TM cells than to cells treated with standard sCD44. The cell viability of cells treated with 10 ng/mL hyperphosphorylated sCD44 was 88% (P < 0.001), in comparison with cells treated with standard sCD44. 

Similarly, phosphorylation of sCD44 altered the effects of sCD44 on RGC-5 survival (Fig. 5) . Hypophosphorylated sCD44 was significantly more toxic than the standard sCD44. At a low dose, 0.1 ng/mL, hypophosphorylated sCD44 decreased cell viability to 71% after only 24 hours (P < 0.05). At this dose, it took 48 hours for standard and hyperphosphorylated sCD44 to cause a significant decrease. The cell viability of cells treated with 10 ng/mL hypophosphorylated sCD44 was 31% after 48 hours. In comparison, the viability of cells treated with 10 ng/mL standard sCD44 was 76% after 24 hours (P < 0.02) and 59% after 48 hours (P < 0.000002). The cell viability of RGCs treated with hyperphosphorylated sCD44 was similar to that in cells treated with standard sCD44. 

sCD44 Phosphorylation, HA Binding, and Pressure

To evaluate the effect of pressure on the bioavailability of sCD44, binding assays of standard and hypophosphorylated sCD44 to immobilized HA were performed to analyze the physiological role of phosphorylated and hypophosphorylated sCD44 in normal and POAG aqueous humor (Fig. 6) . At 0 mm Hg, the dissociation constant (K _d) of hypophosphorylated sCD44 (K _d = 0.0868 nM) to HA was five times greater than that of standard sCD44 (K _d = 0.0167 nM). The maximum binding of standard and hypophosphorylated sCD44 were similar. At 40 mm Hg, the K _d of hypophosphorylated sCD44 (K _d = 0.0686 nM) to HA was four times greater than that of standard sCD44 (K _d = 0.0157 nM). 

Rotary Shadowing

To determine whether HA polymer is altered by pressure, electron microscopy was used to visualize a rotary-shadowed preparation of HA at 0 (atmospheric pressure) and 40 mm Hg. At 0 mm Hg, HA polymers formed an intertwining network containing loops, cables, and hairpins (Fig. 7A) , as previously observed by Scott et al. ²⁵ In contrast, at 40 mm Hg, HA formed cables without an intertwining network or loops (Fig. 7B) . When the HA preparation at 0 mm Hg contained 0.2 M glycine (Fig. 7C) , the intertwining network was disrupted, indicating the presence of hydrophobic patches within the HA polymer and network. Thus, increased pressure influenced the HA polymer and its available sCD44 binding sites. This observation and results of the HA affinity column suggest that the effective bioavailable concentration of sCD44 is influenced by HA binding and IOP. 

Discussion

Although the multifunctional CD44H transmembrane protein is widely distributed and once was considered to function primarily as a cell-adhesion molecule, CD44H is now recognized as having broader functions in cell signaling and survival. ^{26 27 28} Proteolytic cleavage of the extracellular domain of CD44H releases sCD44, which has different biological functions than the intact CD44 protein. ^{9 29 30} The purpose of this study was to determine the pI and phosphorylation of sCD44 in aqueous samples of patients with POAG, in comparison with normal patients without glaucoma. The increased concentration ¹³ and the pI of aqueous sCD44 on 2-D Western blot analysis distinguished patients with POAG or NPG from normal subjects. All the patients with glaucoma were treated with a variety of glaucoma medications (Tables 3 4) . It is possible that glaucoma medications influence sCD44 concentration by decreasing aqueous turnover rate and/or modifying posttranslational modification of sCD44. An expanded cohort of patients with POAG, glaucoma medications, and sCD44 concentrations may assist in the analysis of sCD44 concentration and the POAG process. Nonetheless, there was no obvious correlation between glaucoma medications and sCD44 concentration in the current set of samples and the pI of POAG aqueous sCD44 was more basic than that of normal aqueous sCD44. 

To determine the possible mechanisms of a more basic pI in POAG aqueous sCD44, consideration was given to phosphorylation and glycosylation. ^{29 31 32} Previously, only the cytoplasmic domain of CD44 has been recognized as phosphorylated. ³³ Phosphorylation of sCD44 was analyzed by Western blot analysis with phosphospecific antibodies. The ectodomain of sCD44 was phosphorylated in normal aqueous; however, the ectodomain of sCD44 was hypophosphorylated in POAG aqueous. Aqueous humor sCD44 from patients with POAG was phosphorylated by CK II in vitro. CK II activity is responsive to stress ³⁴ resulting in translocation of the enzyme from the cytoplasmic compartment to the nuclear compartment and also redistribution of the kinase within the nuclear compartment. ³⁵ It is not known whether operative CK II is active in the anterior segment. Thus, POAG and NPG aqueous are characterized by a posttranslational modification of sCD44, which is more basic and less phosphorylated than the standard sCD44 present in normal aqueous. 

The phosphorylation of the ectodomain of standard sCD44 and the hypophosphorylation of sCD44 in POAG are novel findings. The presence of hypophosphorylated sCD44, which distinguished POAG aqueous humor, may be related to the CD44-TGF-β signaling pathway, which is also known to be increased in POAG aqueous humor. ³⁶ First, it has been recognized that CD44H interacts with TGF-β, which influences CD44 phosphorylation through autophosphorylation and CK II activity. ³⁷ If CD44 interaction with TGF-β is disrupted, CK II is turned off, ³⁸ which results in hypophosphorylation of sCD44. Secondly, in cells overexpressing sCD44, CH44H signaling is blocked, preventing the activation of TGF-β2, and cells die by apoptosis. ³⁹ In a previous study, sCD44 cytotoxicity was verified by four controls: heat inactivation, premixing with HA, anti-CD44 antibody, and coadministration of a pan caspase inhibitor. ¹⁴  

CD44 is characteristically phosphorylated in the cytoplasm that facilitates interaction with the cytoskeleton. ^{40 41} Our previous studies in which we used computer-aided color image analysis of normal and POAG sections analyzed by immunostaining with CD44H antibody clearly separated individual cases of POAG from normal. ¹² The sections were treated with Triton X-100, and CD44H was extracted. CD44H in POAG eyes may relate to a change in the associated GPI-anchor in the cell membrane, cell activity, and cytoplasmic phosphorylation of CD44H. ⁴²  

The ectodomain of CD44 contains seven possible CK II consensus phosphorylation sites. Although the phosphorylation sites are not known in sCD44, the phosphorylation status of cellular proteins is controlled by the opposing actions of protein kinases and phosphatases. ⁴³ Alterations in the phosphorylation of sCD44 may involve metabolic stress, free radicals, or age-related modification of the kinase–phosphatase regulation mechanism that results in the release of the more toxic hypophosphorylated sCD44. Therefore, POAG may be another neurodegenerative disease in which the phosphorylation status of a putative toxic protein plays an important role in the disease process. 

CD44 is the principal ligand for extracellular matrix HA. A potentially important function of CD44 is its role in the turnover of HA. The expression of and proteolytic release of the sCD44 by cells allows them to bind and internalize HA, as well as any HA-binding proteins. ⁴⁴ Several proteins (e.g., sCD44, ^{45 46} RHAMM ⁴⁷ and metastatin ⁴⁸ ) that can bind to HA have antitumor properties. ⁴⁸ Notably, the antitumor activity of metastatin is inhibited by HA, and cells that have higher concentrations of HA are resistant to metastatin. ⁴⁸ Recently, it has been reported that a synthetic peptide containing the HA-binding motif inhibits tumor growth by targeting mitochondria and triggering the intrinsic pathway of apoptosis. ⁴⁴ In addition, CD44H is essential for certain high-affinity receptors (e.g., erbB2 phosphorylation and erbB2–erbB3 heterodimerization) for cell survival. ²⁸ If sCD44 interferes with CD44H activity, erbB2 is less active, which results in an extrinsic pathway of apoptotic cell death. 

The bioavailability of sCD44 depends on its binding to HA, and the results of the binding of sCD44 to HA are influenced by pressure. HA has hydrophobic patches ⁴⁹ that are affected by elevated pressure and are disrupted by glycine, as demonstrated by electron microscopy and rotary shadowing. The apparent change in the HA polymer with increased pressure and the decreased binding of sCD44 to HA with increased pressure may be one reason why increased IOP is a clinical risk factor in POAG. Topical ocular hypotensive medications delay and prevent the onset of POAG in individuals with elevated IOP ³ and in patients with NPG. ⁵⁰ We hypothesize that IOP and HA play pivotal roles in blocking the potential cytotoxicity of sCD44. In normal aqueous humor, HA binds and inactivates sCD44. In POAG TM ^{7 8} and aqueous humor (Navajas JR, et al. IOVS 2004;45;ARVO E-Abstract 3665), HA concentration is decreased. Aqueous humor sCD44 concentration is twice that of normal, ¹³ and sCD44 is now more bioavailable, particularly if IOP is increased, because sCD44 dissociates more freely from HA. Consequently, in POAG, sCD44 concentration reaches a threshold and is cytotoxic to certain target cells (e.g., TM, RGCs, or supporting cells in the prelaminar portion of the optic nerve). 

In this study we identified a novel, hypophosphorylated form of sCD44 in the aqueous humor of POAG. Since a hypophosphorylated sCD44 was cytotoxic to TM and RGC cells in vitro, a hypophosphorylated sCD44 in POAG aqueous humor is a possible protein candidate for the cause of cell death in POAG. The presence of hypophosphorylated sCD44 clearly distinguished POAG from normal aqueous humor, suggesting a specific pathophysiology and a biochemical hallmark of the disease. 

 Supported by National Eye Institute Grant EY-12043, Rosemary O’Meara and Kathleen F. Connelly Memorial Funds, Illinois Society for the Prevention of Blindness, Midwest Eye-Banks and Transplantation Center, and an unrestricted grant from the Research to Prevent Blindness.

 Submitted for publication December 14, 2004; revised March 2, 2005; accepted March 31, 2005.

 Disclosure: P.A. Knepper, None; A.M. Miller, None; J. Choi, None; R.D. Wertz, None; M.J. Nolan, None; W. Goossens, None; S. Whitmer, None; B.Y.J.T. Yue, None; R. Ritch, None; J.M. Liebmann, None; R.R. Allingham, None; J.R. Samples, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Primary Antibodies and Source

Antibodies	Specificity*	Category	Dilution	Source
PKA	R-X-X-pS/T	Rabbit polyclonal	×1000	Cell Signaling
PKB	R/K-X-R/K-X-X-pS/T	Rabbit polyclonal	×1000	Cell Signaling
PKC	pT-X-R/K	Rabbit polyclonal	×100	Cell Signaling
MAPK/CDK	pS/T-P	Mouse monoclonal	×1000	Cell Signaling
S/T	pS/T	Mouse monoclonal	×1000	Upstate USA
S	pS	Mouse monoclonal	×1000	Sigma-Aldrich
T	pT	Mouse monoclonal	×800	Sigma-Aldrich
Y	pY	Mouse monoclonal	×200	Upstate USA

*  The code for specificity is the probable motif for phosphorylation, which is denoted as p; capital letters represent the corresponding amino acid, and X is any amino acid.

Figure 1.

 

Two-dimensional PAGE and Western blot analysis of aqueous humor aliquots equivalent to 5 μg protein were separated by electrophoresis and immunoblotted with anti-CD44 antibody, to identify isoelectric variants of 32-kDa sCD44. Two representative Western blot analyses are shown for normal eyes: a 72-year old patient (H-2004) and a 60-year-old patient (H-2005) with normal cup-disc ratios and visual fields. Three are shown for POAG: an 86-year-old patient (H-1646) with a 0.8 cup-disc ratio and moderate visual field loss, a 77-year-old patient (H-2027) with a 0.7 cup-disc ratio and moderate visual field loss, and a 61-year-old patient (H-1940) with a 0.9 cup-disc ratio and severe visual loss, who had undergone successful glaucoma filtration surgery. Representative Western blot analyses are shown for NPG, a 59-year-old patient (H-1965) with a 0.8 cup-disc ratio and moderate visual field loss; JOAG, a 30-year-old patient (H-1953) with a 0.9 cup-disc ratio and severe visual field loss; and exfoliation glaucoma (EG), a 74-year-old patient (H-1975) with a 0.9 cup-disc ratio and severe visual field loss.

Table 2.

 

2-D PAGE Isoelectric Focusing of Aqueous sCD44^*

	n	Average pH, †	pH Range, ‡	P , §
Normal	4	6.38 ± 0.08	5.4 to 7.0	—
POAG	9	6.96 ± 0.07	5.4 to 8.6	0.0004
Normal-tension glaucoma	2	6.76 ± 0.09	5.2 to 8.6	—
JOAG	2	6.29 ± 0.01	5.8 to 6.6	—
Exfoliation glaucoma	2	6.40 ± 0.01	5.2 to 7.0	—

*  2-D PAGE was used to separate proteins according to their net charge, and the sCD44 isoelectric variants were visualized by Western blot with anti-CD44 antibody and ECL detection.

†  The average pH was obtained by densitometry of the Western blots.

‡  The pH range of the immunopositive staining for sCD44 on Western blots.

§  Significance using non-parametric Mann-Whitney test.

Table 3.

 

Clinical Status, Glaucoma Medications, and 2-D Gel Electrophoresis of Individual Aqueous Humor Samples

Clinical Status/Patient Number	Age	C/D*	VF, †	Glaucoma Medications											CD44, ‡
								Pilocarpine	β-Blocker	α-Adrenergic	Carbonic Anhydrase Inhibitor			Prostaglandin Analogs
																Systemic	Topical
Normal
H-2005	60	0.3	1	−	−	−	−	−	−	8.07
H-2004	72	0.4	1	−	−	−	−	−	−	9.62
H-2035	85	0.4	1	−	−	−	−	−	−	10.04
H-2033	60	0.3	1	−	−	−	−	−	−	7.96
Mean	69.3 ± 11.9									8.92 ± 1.06
Primary Open-Angle Glaucoma
H-2027	77	0.7	3	−	+	−	−	+	+	16.23
H-1982	55	0.9	3	−	+	+	−	+	+	15.87
H-2504	59	0.9	4	+	+	+	+	+	+	32.12
H-1646	86	0.9	3	+	+	−	−	−	−	23.38
H-2090	80	0.7	2	−	−	+	−	−	+	14.15
H-1940, §	61	0.95	4	−	+	+	−	+	+	5.80
H-2036	75	0.9	4	−	−	−	+	−	+	8.96
H-2024	80	0.8	3	−	+	−	−	+	+	18.51
H-2022	75	0.6	3	−	+	−	−	−	−	16.46
Mean	72.0 ± 10.9									16.83 ± 7.68, ∥
Normal–pressure glaucoma
H-1965	59	0.8	3	−	−	+	−	−	+	14.38
H-2132, §	64	0.7	2	+	+	−	−	+	−	6.64
Mean	61.5 ± 3.5									10.51 ± 5.47
Exfoliation glaucoma
H-1967	81	0.9	4	+	+	+	−	+	+	11.01
H-1975	74	0.9	4	−	+	+	−	+	+	10.22
Mean	77.5 ± 4.9									10.61 ± 0.56
Juvenile open-angle glaucoma
H-1953	27	0.95	4	−	−	+	−	+	+	10.12
H-2131	38	0.9	4	−	+	−	−	−	+	6.22
Mean	32.5 ± 7.8									8.17 ± 2.76

*  Cup-disc ratio.

†  Visual field classification was 1, normal; 2, arcuate defect; 3, abnormal in one hemifield and not within 5° of fixation; 4, abnormal in both hemifields or within 5° of fixation

‡  CD44 concentration of aqueous humor was determined by ELISA and is expressed as nanograms per milliliter.

§  Patient had successful filtration surgery at least 6 months before cataract surgery, when the aqueous humor sample was obtained.

∥   P < 0.02 versus normal.

Table 4.

 

Clinical Status and Glaucoma Medications of Pooled Aqueous Humor Samples

Clinical Status/Patient Number	Age	C/D	VF	Glaucoma Medications											CD44
								Pilocarpine	β-Blocker	α-Adrenergic	Carbonic Anhydrase Inhibitor			Prostaglandin Analogs
																Systemic	Topical
Normal
H-2040	65	0.5	1	−	−	−	−	−	−	8.52
H-2042	61	0.5	1	−	−	−	−	−	−	9.55
H-2047	85	0.6	1	−	−	−	−	−	−	9.58
H-2609	60	0.2	1	−	−	−	−	−	−	7.62
H-2610	76	0.4	1	−	−	−	−	−	−	3.35
H-2612	75	0.3	1	−	−	−	−	−	−	1.08
H-2614	56	0.3	1	−	−	−	−	−	−	2.93
H-2619	53	0.4	1	−	−	−	−	−	−	1.75
Mean	66.3 ± 11.2									5.55 ± 3.61
Primary open-angle glaucoma
H-1982	55	0.9	3	−	+	+	−	+	+	15.87
H-1987	58	0.9	4	+	+	+	+	+	+	15.43
H-1988	79	0.95	4	−	+	−	−	−	+	14.28
H-2000	69	0.8	4	−	+	+	−	−	+	15.85
H-2029	74	0.7	1	−	−	−	−	+	−	13.63
H-2036	75	0.9	4	−	−	−	+	−	+	8.96
Mean	68.3 ± 9.8									14.00 ± 2.63*

*   P < 0.001 versus normal.

 Data are as described in Table 3 .

Figure 2.

 

Phosphospecific Western blot analysis of immunoprecipitated sCD44 from pooled normal and POAG aqueous humor. A 20-pg equivalent of sCD44 was subjected to PAGE and immunoblotted with (A) a CD44 antibody, to ensure equal sCD44 loads and sCD44 transfer (CD44), and with phosphospecific antibodies against the following motifs: serine/threonine (S/T); serine (S); threonine (T); tyrosine (Y); PKA; PKB; PKC; and MAPK. (B) Twenty-picogram equivalents of sCD44 were treated with 50 ng CK II alone or in the presence of CK II inhibitors, subjected to PAGE, and immunoblotted with anti-phosphoserine/threonine antibody. C, control, sCD44 with no enzyme treatment; CK II, sCD44 treated with CK II; CK II+Hep, sCD44 treated with CK II and te CK inhibitor heparin; CK II+DRB, sCD44 treated with CK II and the CK inhibitor DRB.

Figure 3.

 

2-D PAGE and Western blot analysis of isolated 32-kDa sCD44 with a 100-pg sCD44 equivalent, as determined by ELISA. Standard sCD44 (sCD44); hypophosphorylated sCD44 (−p sCD44) is isolated standard sCD44 treated with alkaline phosphatase to dephosphorylate sCD44, and hyperphosphorylated sCD44 (+p sCD44) is isolated standard sCD44 treated with CK II to phosphorylate sCD44.

Figure 4.

 

Human TM cell viability and phosphorylation of sCD44. Human TM cells were treated with 10 ng/mL standard (sCD44), 10 ng/mL hyperphosphorylated (+p sCD44), and 10 ng/mL hypophosphorylated (−p sCD44) for 24 hours. Data represent the mean ± SD of three experiments; **P < 0.01;***P < 0.001 compared with control; ††P < 0.01 compared with standard sCD44; ‡‡‡P < 0.001 compared with hyperphosphorylated sCD44.

Figure 5.

 

RGC-5 cell viability and phosphorylation of sCD44. RGC-5 cells were treated with 0.1 ng/mL or 10 ng/mL standard (sCD44), hyperphosphorylated (+p sCD44), or hypophosphorylated (−p sCD44) sCD44 for 24 (A) or 48 (B) hours. Data represent the mean ± SD of results in three experiments; *P < 0.05; **P < 0.01; ***P < 0.001 compared with the control; †P < 0.05; ††P < 0.01; †††P < 0.001 compared with standard sCD44; ‡P < 0.05; ‡‡P < 0.01; ‡‡‡P < 0.001 compared with hyperphosphorylated sCD44.

Figure 6.

 

sCD44 binding to HA as a function of sCD44 phosphorylation and pressure. Graded amounts of 0.5 to 5.0 ng of standard sCD44 and hypophosphorylated sCD44 (−p sCD44) were applied to mini HA-affinity columns at 0 (A) or 40 (B) mm Hg. Bound and unbound concentrations were measured by ELISA, and each data point represents the mean results of duplicate ELISAs. The saturation curves (inset) and Scatchard plots of standard and hypophosphorylated sCD44 were analyzed for K _d and maximum binding (B _max). The x-axis is the amount of sCD44 bound, expressed in micromoles per mole of hyaluronic acid (mole HA). The y-axis is the amount of sCD44 bound, expressed in micromoles per mole of hyaluronic acid (mole HA) divided by picomoles of unbound sCD44 (pM).

Figure 7.

 

Rotary shadowing of HA as a function of pressure. Electron micrographs show results of (A) 0.1 mg/mL 100-kDa HA in 0.5 M ammonium acetate buffer (pH 7.2) at 0 mm Hg; (B) 0.1 mg/mL 100 kDa HA in 0.5 M ammonium acetate buffer (pH 7.2) at 40 mm Hg; and (C) 0.1 mg/mL 100-kDa HA in 0.2 M glycine (pH 7.2) at 0 mm Hg, viewed with an electron microscope. Note that HA formed several polymeric chains that formed an intertwining network at 0 mm Hg and that the HA network was disrupted by pressure of 40 mm Hg and 0.2 M glycine at 0 mm Hg. Magnification, ×70,000; scale bar, 100 nm.