Secreted protein acidic and rich in cysteine (SPARC) knockout mice have greater outflow facility

Purpose Secreted protein acidic and rich in cysteine (SPARC) is a matricellular protein that regulates intraocular pressure (IOP) by altering extracellular matrix (ECM) homeostasis within the trabecular meshwork (TM). We hypothesized that the lower IOP previously observed in SPARC -/- mice is due to a greater outflow facility. Methods Mouse outflow facility (Clive) was determined by multiple flow rate infusion, and episcleral venous pressure (Pe) was estimated by manometry. The animals were then euthanized, eliminating aqueous formation rate (Fin) and Pe. The C value was determined again (Cdead) while Fin was reduced to zero. Additional mice were euthanized for immunohistochemistry to analyze ECM components of the TM. Results The Clive and Cdead of SPARC -/- mice were 0.014 ± 0.002 μL/min/mmHg and 0.015 ± 0.002 μL/min/mmHg, respectively (p = 0.376, N/S). Compared to the Clive = 0.010 ± 0.002 μL/min/mmHg and Cdead = 0.011 ± 0.002 μL/min/mmHg in the WT mice (p = 0.548, N/S), the Clive and Cdead values for the SPARC -/- mice were higher. Pe values were estimated to be 8.0 ± 0.2 mmHg and 8.3 ± 0.7 mmHg in SPARC -/- and WT mice, respectively (p = 0.304, N/S). Uveoscleral outflow (Fu) was 0.019 ± 0.007 μL/min and 0.022 ± 0.006 μL/min for SPARC -/- and WT mice, respectively (p = 0.561, N/S). Fin was 0.114 ± 0.002 μL/min and 0.120 ± 0.016 μL/min for SPARC -/- and WT mice (p = 0.591, N/S). Immunohistochemistry demonstrated decreases of collagen types IV and VI, fibronectin, laminin, PAI-1, and tenascin-C within the TM of SPARC -/- mice (p < 0.05). Conclusions The lower IOP of SPARC -/- mice is due to greater aqueous humor outflow facility through the conventional pathway. Corresponding changes in several matricellular proteins and ECM structural components were noted in the TM of SPARC -/- mice.

Matricellular proteins are nonstructural, secreted glycoproteins that facilitate cellular control over their surrounding ECM [15]. The matricellular protein Secreted Protein Acidic and Rich in Cysteine (SPARC) is one of the most highly transcribed genes in TM tissue and in TM cells undergoing physiological stress, indicating the important role of SPARC in normal physiology [16][17][18]. SPARC is present in many tissues throughout the body and has a prominent role in ECM homeostasis, especially fibrosis [19,20]. SPARC overexpression in perfused cadaveric human anterior segments increased IOP with a corresponding increase of fibronectin and collagen IV in the juxtacanalicular (JCT) TM. This increase of fibronectin and collagen IV was also observed in cultured human TM endothelial cells [21].
Transforming growth factor-β2 (TGF-β2) has been implicated in the pathogenesis of POAG [22][23][24][25][26][27][28][29][30][31]. In TM cells, SPARC is up-regulated by TGF-β2 via the smad 2/3, JNK, and p38 pathways [32,33]. The transgenic deletion of SPARC in mice prevents the ocular hypertensive effects of TGF-β2 [34]. We have previously shown that SPARC -/-mice have a 15-25% lower IOP compared to their wild-type (WT) counterparts that appears to be the result of enhanced aqueous drainage using an indirect measurement technique [35,36]. SPARC -/-mice also use more of the TM for outflow and exhibited decreased collagen fibril diameter in the JCT TM [36]. Although we have previously described the physiologic effects of SPARC deletion on IOP and central corneal thickness [35], the qualitative effect of SPARC deletion on the mouse JCT ECM is not yet known.
In this study we determined how SPARC expression affects outflow facility in mice. We hypothesized that SPARC regulates outflow facility by shifting the balance of ECM synthesis and turnover in the TM. We measured the outflow facility of SPARC -/-and WT mice using more direct measurement techniques and assessed the relative expression of selected structural ECM proteins in TM via immunohistochemistry (IHC) and in vitro using cultured murine TM endothelial cells.

Animal husbandry and handling
Animal procedures were conducted in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The protocol was approved by the Case Western Reserve University Institutional Animal Care and Use Committee (Protocol Number: 2013-0166). All procedures were carried out under ketamine/xylazine anesthesia, and all efforts were made to minimize suffering. To measure IOP and assess aqueous humor dynamics, animals were anesthetized with an intraperitoneal (IP) injection (0.1-0.2 ml/20 g body mass) of an anesthetic cocktail containing ketamine (16.5 mg/ml) and xylazine (1.65 mg/ml) which was provided by the animal facility. SPARC -/-mice and their wild type (C57BL6-SV129) strain had been previously generated [35]. The mice were bred at the Animal Resource Center of Case Western Reserve University. Equal numbers of male and female mice between 6 and 8 weeks in age were used. All animals were maintained on a 12-hour light/12-hour dark cycle (on 7:00 AM, off 7:00 PM) with food and water available ad libitum. Mice were sacrificed by anesthetic overdose followed by cervical neck dislocation once confirmed to be non-responsive in accordance to our approved IACUC protocol.

Intraocular Pressure (IOP) measurement
The Tonolab (Icare, Vantaa, Finland) was used to measure IOP through rebound tonometry. Our technique has been described in detail [35][36][37]. Briefly, mice were anesthetized via an I.P. injection of a cocktail of ketamine/xylazine. Following anesthesia, the animal was placed on a movable stand (BrandTech Support Jack; BrandTechScientific, Essex, CT) with its nose inside the facemask. The Tonolab (Colonial Medical Supply, Franconia, NH) was fixed horizontally, and a remote pedal was used to actuate measurements to eliminate potential artifacts caused by manual handling of the device. The mouse was positioned to allow the probe to contact the central cornea perpendicularly. Three sets of six measurements each were made, and the modes of each set were averaged; this value was recorded as the IOP. All measurements were conducted between 11 AM and 3 PM to minimize any potential artifact from circadian variability [16,35]. Additionally, the IOP was measured between 4 and 7 minutes after anesthetic injections on either the left or right eye (randomly chosen for each mouse) without adjusting the stand or tonometer. The IOP was not measured before 4 minutes because mice were not sufficiently anesthetized before this time to allow corneal contact with the probe [34]. Also, prior to 4 minutes, ketamine may have induced a temporary increase in IOP [38].

Episcleral venous pressure (P e ) measurement
We estimated P e using an adapted method that was previously described by two separate groups, Weinreb RN et al. and Millar JC et al. [39,40]. After the mice were anesthetized, a 35-gauge microneedle was inserted into the anterior chamber. The three-way valve was switched so that the manometer reservoir was open to the eye; OD or OS was randomly selected for measurement. The reservoir was then lowered at a rate of 1 mmHg/min until reflux of blood from the episcleral veins into Schlemm's canal was observed [39]. While monitoring the limbus under a professional ophthalmoscope (Keeler Instruments USA, Inc., Broomhall, PA), we recorded the pressure at which the reflux of blood was first observed and designated it the P e . The measurement procedure was repeated twice in each eye.

Outflow facility (C) measurement
We evaluated outflow facility using a previously published method [40]. The three-way valve was switched to close the reservoir and open the microdialysis infusion pump (SP101i Microdialysis Infusion Pump, World Precision Instruments (WPI), Sarasota, FL, USA) connected to 3 mL syringes. Air bubbles were washed out at a flow rate of 300 μl/min, and the pressure was monitored using LabVIEW software. After adjusting the pressure manometrically to pre-cannulation values (~10 mmHg) and allowing 10-15 minutes for pressure to stabilize, eyes were then initially perfused at a flow rate of 0.1 μL/min and the system was allowed to run for 12 minutes to achieve stable pressure. The flow rate was then increased to 0.2, 0.3, 0.4, and 0.5 μL/ min, and the stabilized pressures at each flow rate were recorded. C (μL/min/mmHg) was calculated as the reciprocal of the slope of the respective pressure-flow rate curves. An eye showing a regression value (R 2 ) less than 0.9 was excluded from analysis. After measurement of C, the animal was euthanized via anesthetic overdose. After approximately 30 minutes, C was measured again (C dead ). We assumed that after death (absence of heartbeat), aqueous production and P e would both be reduced to 0.

Calculating uveoscleral outflow (F u ) and eliminating aqueous formation rate (F in )
We referenced the same modified Goldmann equation as reported by Millar et al. [40]. As described, the IOP and P e in each eye were measured directly by rebound tonometry and manometric methods, respectively. When the microneedle was inserted into the anterior chamber and the infusion pump was switched on, the equation is as follows: where IOP p is IOP when the infusion pump is on, and F p is the flow rate incurred by the infusion pump. By adjusting F p to different values (0.1-0.5 μL/min), C was determined as the reciprocal of the slope of the IOP p /F p regression line. After mice were euthanized, F in and P e were assumed to be 0. The euthanized animals were perfused once again to determine C dead . F u and F in values for live mice were then estimated for live mice by calculation, as described by Millar, Clark and Pang (2011) [40].

Comparison of F u(live) and F u(dead)
To compare F u in live eyes with F u in eyes immediately after death, F u was measured directly as described previously [40]. Briefly, three live SPARC -/-mice, three live WT mice, three dead SPARC -/-mice, and three dead WT mice freshly euthanized by anesthetic overdose (i.e. six eyes in each group) were used. For the live mice, a 35-gauge needle was inserted into the anterior chamber for perfusion with fluorescein isothiocyanate (FITC)-dextran (10-4 M; 7000 ng/ μL; Sigma Chemical Co, St. Louis, MO) at a flow rate of 0.5 μL/min for 10 minutes. The live mice were then euthanized by anesthetic overdose. For the dead mice, the carcasses were perfused with sterile saline for 30 minutes immediately after death, after which the anterior chambers were perfused in the same manner as the live mice. The eyes were then enucleated and dissected to exclude the cornea and TM. The remaining portions of the eyes were dissected into three parts: the lens, the vitreous, and the retina/choroid/iris-ciliary body/scleral shell. Each part was homogenized and centrifuged in PBS. The supernatant was used to measure the fluorescence intensity (excitation 492 nm, emission 518 nm). F u for each eye was calculated as: where a is the volume of the supernatant analyzed (mL), b is the FITC-dextran concentration in the sample (ng/mL), and T is perfusion time (min).

Immunohistochemistry
Paraffin-embedded TM tissue slides were deparaffinized in xylene for 15 minutes and incubated with xylene for a second 15 minute period. Sections were subsequently hydrated with ethanol dilutions (100%, 95%, and 70%). The tissue was blocked in 5% bovine serum albumin for 1 hour at room temperature (RT), then permeabilized using 0.2% Triton X-100. The primary antibody was applied at 1:100 dilution overnight at 4˚C (Table 1). Slides were subsequently washed with 1×PBS-T, and secondary antibodies were applied at 1:200 for 1 hour at RT (Table 1). After two additional washes, a slide cover was mounted with SlowFade Gold antifade reagent with DAPI (S36938; Life technologies, Eugene, OR). The tissue was imaged using a Leica DMI 6000 B inverted microscope (at 40x) and a Retiga EXi Aqua Blue camera (Q-imaging, Vancouver, British Columbia, Canada). Three rectangular areas of equal size within the TM were selected at random for each section, and Metamorph Imaging Software (Molecular Devices, Downingtown, PA) was used to calculate the average fluorescence through the 488 nm and 594 nm channels.

Preparation of Murine TM (MTM) cells.
Murine TM cells were harvested using a published methodology from SPARC -/-and WT eyes [41]. Briefly, 3-4 month old mice were anesthetized by intraperitoneal (IP) injection of the ketamine/xylazine mixture. They were subsequently placed on a stereotaxic mouse adaptor, and the head was externally secured with two jaw holder cuffs as well as a tooth bar and nose clamp (Stoelting Co, Wood Dale, IL). A 30-gauge needle attached to a 10 μL syringe was mounted onto a microsyringe pump (World Precision Instruments, Inc., Sarasota FL). The pump was mounted to a micromanipulator (World Precision Instruments, Inc.). Under 35x stereotaxic magnification (Kent Scientific Co., Torrington, CT), proptosis of the eye was produced by mild pressure over the medial and lateral canthi. A paracentesis was performed to drain aqueous humor. The syringe was then advanced so that the needle was nearly parallel to the surface of the iris. Only one eye was injected, following the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research policy. Two microliters of magnetic microbeads (Spherotech, Lake Forest, IL) were injected into the anterior chamber. After completing the injection, the needle was left in the anterior chamber for 5 minutes before removal. To isolate MTM cells, we used a previously published method [41]. Briefly, mice were sacrificed 7 days after injection and the injected eyes were enucleated and dissected to remove the retina, choroid, vitreous, and lens. Tissue was pooled from 12 mice and digested in 4 mg/mL collagenase A (Worthington Biochemical Corporation, Lakewood Township, NJ) with 4 mg/mL BSA dissolved in PBS. It was incubated at 37˚C for 2-4 hours and then attached to a magnet. Cells that did not phagocytize the beads were aspirated from the tube and replaced with fresh media. Resuspended cells were passed through a 100 μm cell strainer (Thermo Scientific, Worcester, MA) and then centrifuged at 600g for 10 minutes. The cell pellet was resuspended in 1 mL media and seeded into a 48-well plate with approximately 500 μL of cell suspension per well.

Immunoblotting
Our immunoblotting technique has previously been described in detail [21]. The conditioned media (CM) from the cultured MTM cells was harvested, and MTM cells were lysed with 1x radioimmunoprecipitation assay (RIPA) buffer. The conditioned media was concentrated 30-fold using a 10 kDa centrifugal filter unit (MilliporeSigma, St. Louis, MO). Equal amounts of total protein in conditioned media or cell lysate (CL) were mixed with 6x reducing (CL) or non-reducing (CM) SDS sample buffer (Boston BioProducts, Ashland, MA). SDS-PAGE was used to resolve proteins in the samples on 10% polyacrylamide gels. Proteins were transferred onto nitrocellulose membranes with a pore size of 0.2 μm (Invitrogen, Carlsbad, CA). Membranes were incubated for one hour in 0.5x blocking buffer (Rockland, Limerick, PA) at RT. Primary antibodies were added to the membranes, which were incubated overnight at 4˚C (Table 1). Membranes were washed three times with 1x TBS/T for 10 minutes at RT, then incubated with secondary antibodies in 0.5x blocking buffer (Rockland) for one hour at RT (Table 1). Membranes were washed three more times with 1x TBS/T and imaged using the Odyssey Infrared Imaging System (Li-Cor, Lincoln, NE). Odyssey densitometric software was used to analyze relative band intensities.

Statistical analysis
All data were analyzed in Microsoft Excel (Microsoft, Redmond, WA,). IOP, C, F in , and F u in SPARC -/-compared to WT mice were calculated and assessed for statistical significance using the 2-tailed unpaired Student's t-test. Statistical analyses for IHC data were completed using unpaired Student's t-tests. p < 0.05 was considered statistically significant. Additionally, 95% confidence intervals were calculated for each parameter.

Aqueous humor hydrodynamics in SPARC -/-and WT mice
The IOP, P e , and C (before and after euthanasia) of the eight SPARC -/-and WT mice eyes were measured, and F in and F u were calculated ( Table 2, Fig 2). IOP was 14.7 ± 1.0 mmHg in SPARC -/-mice compared to 17.3 ± 0.5 mmHg in WT mice (p = 1.98 x 10 −5 , n = 8 in each group). The outflow facility before death (C live ) in SPARC -/-mice (0.014 ± 0.002 μL/min/ mmHg), was greater than C live in WT mice (0.010 ± 0.002 μL/min/mmHg) (p = 0.002). We compared C live and outflow facility after euthanasia (C dead ) in SPARC -/-and WT mice and found there was no significant difference between C live and C dead in either strain (p = 0.376, p = 0.548, respectively). P e values were the same in SPARC -/-and WT mice, 8.0 ± 0.3 mmHg and 8.3 ± 0.8 mmHg, respectively (p = 0.304, n = 8 in each group). F u and F in were the same for both SPARC -/-(n = 3) and WT (n = 4) mice ( Table 2, Fig 2).

Effect of SPARC Deletion on ECM Proteins in SPARC -/-and WT mice
IHC was performed to investigate changes of ECM proteins in the TM. The fluorescence intensities of collagen IV, collagen VI, fibronectin, laminin, PAI-1, and tenascin-C exhibited significant decreases in SPARC -/-mice compared to WT mice (p < 0.05 for all). There were no differences in fluorescence intensities between SPARC -/-and WT mice for collagen I, hevin, osteopontin, tenascin-X, TSP-1, and TSP-2 (p > 0.05). There was no fluorescence staining for SPARC in SPARC -/-mice. (Table 3, Figs 3 and 4).

PLOS ONE
The expression of ECM and matricellular proteins from murine TM (MTM) cell culture demonstrated fewer significant changes between WT and SPARC -/-mice than the IHC fluorescence. Namely, laminin did not show a statistically significant change in band intensity in Table 3. ECM and matricellular protein percent change in SPARC -/-mice compared to WT mice. Percent change in TM tissue was determined through IHC. Percent change in MTM culture was determined through immunoblotting.  western blots of protein from MTM cell cultures (Table 3). Unlike the other proteins investigated, PAI-1 expression changed in opposite directions in the absence of SPARC between in vivo and in vitro settings; PAI-1 expression was decreased in mice but increased in cell culture. Collagens IV and VI, fibronectin, and tenascin-C were significantly decreased in SPARC -/mice via the MTM method, which is consistent with the findings from the IHC fluorescence method.

Discussion
We have previously shown that SPARC -/-mice have a 15-25% lower IOP and greater areas of TM utilized for filtration compared to wild-type mice [35,36]. Using the technique of constant-flow infusion reported by Millar et al. [40], we demonstrated that the differences in IOP between SPARC -/-and WT mice are attributable to differences in C values, i.e. outflow facility. In this study, SPARC -/-mice showed a 14.7 ± 6.4% decrease in IOP, corresponding to a 37.8% ± 16.1% increase in C value compared to WT mice; the P e , F in , and F u were unchanged between SPARC -/-and WT mice. Regression analysis applied to individual mice revealed a close-to-linear linear correlation between infusion rate and pressure over the range of flow rates studied. As mentioned by Millar et al. [40], this technique can also measure C in the eyes of euthanized animals. Combining this value with IOP and P e measurements allows for determination of all aqueous humor hydrodynamics parameters, e.g. F u and F in , as described in the Goldmann equation. Millar et al. [40] validated their assumptions that euthanasia does not affect C and F u . It was assumed that euthanasia reduced F in and P e to zero. Additionally, we used FITC-dextran perfusion to directly measure F u and found that F u was not affected by euthanasia in SPARC -/-and WT mice. In our WT mice, aqueous humor production and outflow facility values were similar, but not identical, to values of other mice reported in previous studies [39,40]; differences in mouse strains, measurement techniques, and anesthesia likely account for the differences of values. We and others have found that IOPs and conventional outflow vary among different mouse strains and durations of anesthesia [42][43][44][45][46][47][48].
We observed a decrease of collagen types IV and VI, fibronectin, and laminin within the JCT TM of SPARC -/-mice relative to WT mice. In MTM cell cultures, these changes were also seen with the exception of laminin. Additionally, PAI-1 was decreased in the JCT TM of SPARC -/-mice but increased in MTM cultures from SPARC -/-mice. The different results between IHC and MTM cell cultures could be due to a true lack of change in expression, as observed with hevin, TSP-1, and TSP-2, or due to differences in substrate, i.e. cell culture versus in vivo setting. Alterations of ECM in JCT have been shown by numerous authors to correlate with IOP changes in various model systems [49][50][51]. In perfused human cadaveric anterior segments, we found that SPARC overexpression increases IOP and correlates with an increase in collagen IV, fibronectin, and laminin [21]. Thus, it is likely that the observed decreases in collagens and fibronectin are responsible for the reduced outflow resistance. Taken together, the murine data in this study along with our data in perfused cadaveric human anterior segments and TM cell cultures strongly indicate that changes in collagen IV, collagen VI, and fibronectin closely follow changes in SPARC [21].
Our observations of ECM differences in the JCT are consistent with other observations in non-ocular tissues of SPARC -/-mice. SPARC -/-mice have differences in collagen fibril morphology and less collagen in non-ocular tissues [52]. SPARC -/-mice have decreased laminin and collagen IV deposition in renal tissue, decreasing damage from experimental diabetic nephropathy [53]. Similar decreases in interstitial collagen are apparent in hearts and in fat depots of SPARC -/-mice [54,55]. We previously reported that collagen fibril diameter was significantly decreased in the JCT of SPARC -/-mice, reflecting the importance of SPARC in collagen processing and a potential mechanism by which IOP is reduced in SPARC -/-mice [2]. Significant changes in collagen fibril diameter in other tissues of SPARC -/-mice have been reported in previous studies [56,57]. For instance, Martinek et al. showed that collagen fibrils formed in the absence of SPARC are smaller and more uniform in diameter than those of wild-type animals [58]. Additionally, overexpression of SPARC by adenoviral delivery in WT animals subjected to myocardial infarction enhanced collagen assembly in these mice and improved cardiac function [59]. Taken together, these studies provide evidence that SPARC regulates collagen assembly and ECM homeostasis.
It is plausible that other members of the matricellular family may compensate for the loss of SPARC. We found that tenascin-C was decreased in the JCT TM of SPARC -/-mice and in MTM cultures from SPARC -/-mice. The functional significance of this change is unclear. However, we have previously shown that single gene deletions of TSP-1 and -2 result in a lower IOP in mice [46,60]. Single gene deletions of osteopontin, hevin, and tenascins-C and -X do not alter IOP in mice [37,46,61]. The lack of change of hevin, osteopontin, tenascin-X, TSP-1, and TSP-2 in SPARC -/-mice suggests that they may work through SPARC-independent pathways to modulate their effects.
Immunohistological examination revealed decreases in fibronectin and collagen types IV and VI within the TM of SPARC -/-mice. Furthermore, immunoblot of MTM cells revealed similar decreases in fibronectin and collagen types IV and VI. The JCT region contains Schlemm's canal (SC) inner wall cells, subendothelial ECM containing an incomplete basement membrane, and TM endothelial cells; these components represent the anatomic location of maximal outflow resistance [62,63]. Regulation of ECM homeostasis has been shown to influence IOP and outflow resistance [9][10][11]21,64,65]. Collagen IV, fibronectin, and laminin are the main components of the basement membrane of the JCT [66]. An increase of basement membrane components collagen IV and fibronectin within the JCT is a significant structural finding in corticosteroid-induced glaucoma, in which IOP is elevated most likely due to increased TM resistance [67][68][69]. Decreased permeability is also evident when high glucose and dexamethasone induce collagen IV and laminin in monolayers of TM endothelial cells [70]. This structure-function relationship between ECM and outflow is evident in an inverse manner in our study, where a decrease of ECM components in the JCT seen in SPARC -/mice correlates with increased TM outflow. Therefore, it is likely that ECM components are significant contributors to the mechanism of decreased IOP in SPARC -/-mice.
We have demonstrated that the lower IOP of SPARC -/-mice is due to greater aqueous humor outflow facility. The observed changes in ECM components in the JCT region provide additional evidence of the importance of ECM homeostasis for IOP regulation. The role of SPARC in influencing these changes in ECM organization and the associated changes in IOP makes SPARC a promising target of further study for its apparent relevance to the pathogenesis of POAG. Future work is planned to characterize the signaling pathways of SPARC in the TM and the mechanisms by which SPARC regulates IOP and JCT ECM homeostasis.