Conceived and designed the experiments: MG AES RP JW RKL SKB. Performed the experiments: MG AES RP JW SKB. Analyzed the data: MG AES RP JW RKL SKB. Contributed reagents/materials/analysis tools: JW RKL SKB. Wrote the paper: SKB MG RKL. Conceived the overall idea, designed the experiments and wrote the manuscript: SKB. Designed and performed the experiments: MG AES RP JW. Assisted in refining experiments and writing the manuscript: RKL.
The authors have declared that no competing interests exist.
Fluid shear modulates many biological properties. How shear mechanosensing occurs in the extracellular matrix (ECM) and is transduced into cytoskeletal change remains unknown. Cochlin is an ECM protein of unknown function. Our investigation using a comprehensive spectrum of cutting-edge techniques has resulted in following major findings: (1) over-expression and down-regulation of cochlin increase and decrease intraocular pressure (IOP), respectively. The overexpression was achieved in DBA/2J-Gpnmb+/SjJ using lentiviral vectors, down-regulation was achieved in glaucomatous DBA/2J mice using targeted disruption (cochlin-null mice) and also using lentiviral vector mediated shRNA against cochlin coding region; (2) reintroduction of cochlin in cochlin-null mice increases IOP; (3) injection of exogenous cochlin also increased IOP; (4) increasing perfusion rates increased cochlin multimerization, which reduced the rate of cochlin proteolysis by trypsin and proteinase K; The cochlin multimerization in response to shear stress suggests its potential mechanosensing. Taken together with previous studies, we show cochlin is involved in regulation of intraocular pressure in DBA/2J potentially through mechanosensing of the shear stress.
Fluid shear is a mechanical stimulus experienced by cells and most organs involved with localized fluid flow. Cellular mechanosensing is linked to cytoskeletal remodeling to respond appropriately to altered fluid shear dynamics in single and multicellular organisms
Fluid flow abnormalities are associated with complex, late onset progressive diseases such as glaucoma (aqueous outflow), idiopathic intracranial hypertension (flow changes in cerebrospinal fluid) and the non-syndromic hearing disorder DFNA9
Cochlin was overexpressed in DBA/2J-Gpnmb+/SjJ mice, which do not develop elevated IOP or glaucomatous neurodegeneration with age, to determine its role in IOP elevation. Mice injected with COCH transgene with IRES mediated GFP expressing (COCH-GFP) lentiviral vector into the anterior eye chamber showed a rise in IOP concomitant with cochlin expression, reaching a peak between 8–30 days. The IOP remains elevated upto 35 days post-injection in animals injected at the age of 6 months (
(A) DBA/2J-Gpnmb+/SjJ mice (n = 42–48 for each vector at each time point, as indicated by the symbols) at six months of age were injected with a lentiviral vector bearing the COCH-GFP transgene or GFP alone (sham) or human serum albumin (HSA) in the anterior chamber (all under the control of a CMV promoter). IOP was recorded at the indicated time periods. (B) C57BL/6J mice at six months of age were injected with a lentiviral vector bearing the
To determine if downregulation of cochlin expression decreases IOP, cochlin shRNA or control GFP lentiviral vector was injected into the anterior chamber (
The role of cochlin in IOP elevation and its peak at 8 months was studied by abrogating cochlin. C57BL/6J mice with targeted disruption (knockout) of COCH
(A) Agarose gel (1.2%) showing the absence of PCR products for exons 8, 10 and 11 in the DBA/2J cochlin−/− mice. These exons are present in DBA/2J cochlin+/+ and cochlin+/− mice. DNA was isolated from mice tail and subjected to PCR amplification for the indicated exons. (B) Western blot analysis of TM extract of DBA/2J cochlin+/+, DBA/2J cochlin+/− and DBA/2J cochlin−/− mice showing the complete absence of cochlin in DBA/2J cochlin−/− mice. Coomassie blue stained gel shows equal loading of protein in all the lanes. (C) Intraocular pressure (IOP in mmHg) in different mice strains at indicated ages (n = 35–45 for each strain at each age group). The DBA/2J cochlin+/+ mice show a higher mean IOP as compared to DBA/2J-Gpnmb+/SjJ, DBA/2J cochlin+/− and DBA/2J cochlin−/− mice at 6 months and at 8 months of age. The difference is most marked at 8 months of age. By 12 months, the mean IOP is similar for the different strains. (D) Western blot analysis of the TM protein extract from the DBA/2J cochlin+/+ and DBA/2J-Gpnmb+/SjJ at the indicated ages, probed for cochlin. GAPDH has been shown as a loading control. (E) Generation 10 (Gen 10) DBA/2J cochlin−/− at six months of age were injected with a lentiviral vector bearing the COCH-GFP transgene (n = 20) or GFP alone (n = 20) in the anterior chamber. The mice were followed and IOP recorded at the indicated time periods. Error bars (C and E) depict ± standard deviation.
Mean IOP rose from 17 to 19.7 mm Hg as DBA/2J cochlin+/+ mice age from 6 to 8 months and then decreased to 12 mm Hg by 12 months of age (
Re-introduction of COCH gene expression in DBA/2J cochlin−/− mice leads to increased IOP, in contrast to control GFP injected mice (
The fluid flow changes must be sensed by cells in order to regulate the structure of the TM that allows passage of aqueous humor and regulate its flow. Aqueous humor is a Newtonian fluid
Cochlin, at the protein level, possesses a short signal peptide and two von Willebrand factor A-like (vWFA1 and vWFA2) domains. The vWFA domain present in ECM proteins is associated with fluid shear responsiveness
(A) Western analysis of purified recombinant cochlin (*) using anti-cochlin antibody before after sheer stress, revealing shear stress-induced mulitmers. Cochlin was subjected to fluid shear of 3 µl/min for 150 cycles. (B) Comparison of proteolytic digestion (in percentage) of native (left side) and multimerized (right side) cochlin by trypsin (open bars) and proteinase K (filled bars) showing slower digestion of multimerized cochlin. Digestion of multimerized and native cochlin (10 µg) was performed using 0.01 µg/10 µg Trypsin or 0.05 µg/10 µg Proteinase K. After incubation at room temperature, the samples were boiled using 1 mM DTT at 100°C for 1 minute and analyzed on SDS-PAGE. Densitometric analyses data from three independent experiments (mean± standard deviation) has been presented. (C) A representative Western analysis of native and multimerized cochlin subjected to subcatalytic amounts of trypsin and proteinase K digestion for 10 minutes at room temperature. Purified recombinant cochlin (*), native monomeric cochlin and multimeric cochlin (initial amount 1 µg) has been depicted as Mo and Mu respectively. Digestion was performed by 0. 1 µg/10 µg Trypsin or 0.5 µg/10 µg Proteinase K. Digestion was stopped by heating at 100°C, separated on a 10% reducing SDS-PAGE and probed with chicken polyclonal antibody against cochlin. Proposed model of cochlin mechanosensing and associated global change in TM. (D) Illustration depicting that the cochlin constitutively secreted by the normal TM cells is degraded by the proteases (dashed lines denoting degraded proteins). (E) Cochlin in the presence of fluctuating shear stress (or elevated divalent cations), forms multimers which are resistant to proteolysis. Multimerized cochlin may potentially interact with the transmembrane proteins TREK-1 or slc44a2 either directly, functionally, or indirectly leading to events that may result in cytoskeletal reorganization. (F) Confocal microscopy image showing the sieve like structure of TM (arrow). A representative area of TM with rectangular arrangement of cells is highlighted by a yellow box. SC = Schlemm's canal; TM = trabecular meshwork. The magnified view of rectangular arrangement of cells highlighted by a yellow box also has been shown. (G) The line diagram depicts the filter like structure of the TM. A magnified view mimicking one such rectangle as in (F) with an additional cell in the middle of the rectangle has been shown. Notice the change in orientation of the middle cell opens up interstitial space for additional fluid flow indicated by dashed arrows. (H) When incremental space opening is insufficient, larger orientation changes by cells in the middle and those forming rectangle are necessary to increase fluid flow. Newly opened space is indicated by arrow head. The dark lines and dark dashed lines are to indicate orientation changes. (I) Cartoon illustrating concerted orientation changes in several individual cells (dashed lines) leading to a global change in the structure of the sieve like TM. (J) Confocal microscopy image of cochlin transfected TM cells (left; i) expressing cochlin (pink) and tiny aggregates of cochlin (arrows) secreted into the media. These small cochlin aggregates coalesce (arrow) to form larger deposits as shown in human glaucomatous TM probed for cochlin (green; right; ii). Scale bar = 25 µm (left) and 50 µm (right). (K) Confocal microscopy image of trilayer of cochlin transfected TM cells on a PVDF membrane exposed to continuous fluctuation in fluid shear stress in an Ussing type chamber showing the development of cochlin deposits (red; arrows).
TM cells function in an environment of continuous varying mechanical and fluid shear forces
Fluid shear, cyclic strain and osmotic shock are forms of biomechanical stress
We hypothesize that the fluid shear responsive property of cochlin plays a role in tissue remodeling, in consonance with transmembrane shear transducing proteins (SACs). Expression of cochlin has been shown to result in co-expression of an SAC, TREK-1 in the primary TM cells
SACs (TREK-1) are activated by shear and osmolarity
We propose a model where such mechanosensing and mechanotransduction enables some TM cells (
We also found that prolonged overexpression of cochlin results in formation of cochlin deposits in cell culture (
The work was conducted adhering to the guidelines of the Institute Review board of the University of Miami. All human samples were handled in keeping with the principles expressed in the Declaration of Helsinki. All experiments with the human samples were conducted at the SKB lab and the protocol was approved by the Institute Review board of the University of Miami. A written informed consent was obtained from all patients undergoing trabeculectomy for POAG and donating the tissue so obtained for research. Cadaveric human eyes were obtained from Bascom Palmer Eye Bank with the approval of Institute Review board of the University of Miami. Human TM cell culture protocol was approved by the Institute Review board of the University of Miami.
Adult mice from inbred strains DBA/2J and DBA/2J-Gpnmb+/SjJ were obtained from the Jackson Laboratory (Bar Harbor, ME). Mice with targeted cochlin knockout allele
Human eyes from normal and POAG donors, all between 40 and 85 years of age (Table S1), were used in this study, and were obtained through the Bascom Palmer Eye Bank (BPEI) and National Disease Research Interchange (NDRI). The protocol for use of human tissue was approved by the Institutional Review Board of the University of Miami. Eyes were enucleated within 12 h of death and stored at −80°C until TM tissue was isolated by dissection. Normal control eyes were from donors with no visual field defects, no evidence of glaucoma, and without central nervous system abnormalities. Fixed human TM tissues used for immunohistochemistry were obtained from the Eye Donor Program of the Foundation Fighting Blindness (Owings Mills, MD). Glaucomatous eyes and tissues were from clinically documented POAG donors. Glaucomatous TM tissues (∼1–2 mm3) were obtained by trabeculectomy from POAG patients in the BPEI and Mundorf Eye Center (Charlotte, NC) with institutional review board approval. Human tissue obtained by trabeculectomy consisted predominantly of TM; however, possible contamination with small amounts of surrounding tissue (e.g. sclera) cannot be excluded. TM cells for cell culture were isolated from the rim tissue associated with corneas used for transplantation at the BPEI and were obtained from healthy human eyes within 3 h of death and stored until use in Optisol-GS medium (Chiron Vision, Claremont, CA).
The mice were anaesthetized using intraperitoneal injection (0.1 µl) of ketamine (100 mg/kg) and xylazine (9 mg/kg). One mouse was anaesthetized at a time and IOP was measured as soon as the mouse failed to respond to touch. All care was taken to ensure that the mice were in a similar level of anesthesia when the IOP measurements were made. To measure the IOP, hand held tonometer, TonoLab (Colonial Medical Supply, Franconia, NH) was used. To ensure accuracy of the measurements, TonoLab measurements were confirmed using anterior chamber cannulation as described previously
Mice were euthanized at the indicated ages and the eyes were immediately enucleated. TM was dissected out of the enucleated eyes under a dissecting microscope by a trained ophthalmologist. Such dissected tissue can have contamination from surrounding tissue. Previously with immunohistochemistry we have shown the presence of cochlin largely to remain confined to TM
Mice were euthanized at 12 months of age and the eyes were immediately enucleated and fixed in 2% Glutraldehyde and left overnight at 4°C. To asses for glaucomatous damage, the optic nerves were sectioned and subsequently stained with paraphenylenediamine (PPD) Images were taken using Zeiss Axiostar Plus CYL#621 microscope (Carl Zeiss Microimaging Inc, Thornwood, NY).
To quantify the percentage of mice suffering from mild, moderate or severe optic nerve damage, PPD stained sections of each mouse were graded by three observers in a blinded fashion independent of each other. Three sections from each mouse (n = 25 in each group) were graded and given a score of 1–10 with one being the least severe and ten being the most severe damage. The scores from the two observers were averaged. Mice getting a final score of 0–4 were labeled as having mild damage, 5–7 moderate damage and 8–10 severe damage.
Mice were anaesthetized as for IOP measurement. A topical anesthetic (tetracaine hydrochloride 0.5%, Bausch & Lomb) was applied to the desired eye. 1 µl of the lentiviral vector bearing the gene of interest (
C57BL/6J mice with targeted disruption (knockout) of
Mice were anaesthetized as described above for IOP measurement. Genotyping was performed on mice approximately 3 weeks of age. Mice tails were tattooed (AIMS, ATS-3 general rodent tattoo system) for identification. Then 2.5–3.0 mm of tail was clipped. The tail was cauterized using Kwik-Stop styptic powder for hemostasis. Three hundred µl of Lysis reagent (Direct PCR-Tail, cat#102-T, Viagen Biotech, Los Angeles, CA) and 9 µl of 10 mg/ml proteinase-K solution (cat# P6556, Sigma-Aldrich, St. Louis, MO) was added to the excised tail and incubated overnight at 55°C. The following morning the tail samples were heated for 45 minutes at 85°C. DNA isolated from the tail was then subjected to polymerase chain reaction (PCR) amplification for exons 4, 5, 8, 10 and 11. PCR products were run on a 1.2% agarose gel (Ultra-pure agarose, cat# 16500-100, Invitrogen) to assess for the presence of the above exons. Following primer pairs were used for identifying DBA/2J cochlin−/− mice:
Exon 4: Sense:
Exon 5: Sense:
Exon 8: Sense:
Exon 10: Sense:
Exon 11: Sense:
The sizes of different exons are as follows Exon 4: 157 bp, Exon 5: 134 bp, Exon 8: 148 bp, Exon 10: 227 bp and Exon 11: 517 bp.
The cochlin KO mice were bred onto DBA/2J background for 10 generations (Gen10) as described below:
Step 1. Crossed DBA/2J Coch+/+ X Coch KO† to obtain Gen1 (Heterozygote Coch+/−)
Step 2. Crossed Gen1 X Gen1 to get Gen1*
Step 3. When Gen1* is genotyped and its KO status was established, it was termed as Gen1 KO.
Step 4. Crossed Gen1KO X DBA/2J Coch+/+ which was termed Gen2
Step 5. Crossed Gen 2 X Gen 2 to get Gen2*
Step 6. When Gen2* is genotyped and its KO status was established, it was termed as Gen2 KO.
Step 7. Crossed Gen2KO X DBA/2J Coch+/+ which was termed Gen3
Step 8. Repeated steps 4 to 7 till Gen 10 KO has been obtained.
(† Coch KO was obtained as a research gift from Dr. C. Stewart at National Institutes of Health
Cochlin was purified according to the established protocols
Shear force (F) = η Av/l, for aqueous humor, where η is the viscosity of the fluid, A is the cross sectional area of the chamber through which the fluid is flowing and l is the length of the chamber. The factor l for aqueous humor is the distance between the ciliary epithelium and the episcleral veins. For aqueous humor outflow whether in normal or in glaucomatous eyes, η, A and l remain constant. The flow rate (v) is the greatest determinant of fluid shear or Force.
DBA/2J-Gpnmb+/SjJ mice were euthanized and the eyes were enucleated immediately. The globes were fixed using 4% paraformaldehyde (cat# 19943, USB Corp, Cleveland, OH) overnight, washed with 1XPBS 3 times for 10 m per wash, placed in 15% sucrose for 4 h and then switched to 30% sucrose for another 4 h. The globes were then placed in a cryomold (cat# Tissue-Tek 4565, Sakura Finetek, USA) and embedded in O.C.T. (Optimum Cutting Temperature) compound (cat# Tissue-Tek 4583, Sakura Finetek, USA) using liquid nitrogen. The embedded globes were then sectioned and subjected to immunohistochemcal analysis probing for cochlin (hCochlin#3, Aves Labs Inc.) and GFP (cat#, 600-301-215 Rockland Immunochemicals Inc., Gilbertsville, PA).
To overexpress cochlin in the TM of congenic DBA/2J mice, cochlin transgene lentivirus was constructed in the HEK-293T [cat# 293T/17 (CRL-11268), ATCC, Manassas, VA] cells. Cochlin expression clone (cat# EX-Q0226-Lv31, GeneCopoeia Inc., Rockville, MD) was packaged into a lentiviral vector by using the Lenti-Pac FIV expression packaging kit and the protocol provided by the manufacturer. This protocol typically yielded 107 ifu/ml of the recombinant lentivirus. The
To down-regulate cochlin expression in the TM of DBA/2J mice cochlin shRNA was constructed in HEK-293T cells using the Trans-Lentiviral™ GIPZ Packaging System (cat# TLP4614, Open Biosystem, Huntsville, AL) and the protocol provided by the manufacturer. This typically yielded a viral stock of 108 tranduction unit (TU)/ml. For control experiments a lentiviral vector expressing GFP was constructed in a similar fashion. The shRNA used was set of 5 clones. The sequence of the clones used is shown below.
Primary human TM cells were cultured from cadaveric corneo-scleral sections obtained from the Bascom Palmer Eye Bank (BPEI) and Mundorf Eye Centre (Charlotte, NC). The cells were isolated through a blunt dissection of the area containing and adjacent to the canal of Schlemm, followed by 2 h digestion in 1X PBS (cat# 21-030-CV, Mediatech Inc.) suspension of 20% 0.01 µg/µl collagenase-A (cat# LS004194, Worthington, Lakewood NJ). The blunt dissection and the proteolytic treatment were performed inside a 12 well culture plate (cat# 665-180 Greiner Bio-One, Neuburg, Germany). Culture media containing [DMEM 1X (cat# 10-017-CM, Mediatech Inc.), 10% heat inactivated fetal bovine serum (FBS) (cat# 35-016-CV, Mediatech Inc.), 0.5% 1.7 mM L-Glutamine (cat# G6392, Sigma-Aldrich), 1% Antibiotic-Antimycotic solution (cat# 30-004-CL, Mediatech Inc.)] was added after 2 h to terminate digestion. A sterile microscopy slide (cat# 56700-194, VMR) was placed on top of the tissue fragment to ensure bottom-contact and immobility inside the media well. The sections were cultured at 37°C, 5% CO2 cell culture incubator. Culture was washed with 1X PBS 7 days later to remove tissue debris; media change occurred every 3–4 days. Thus obtained cells were trypsin treated (cat# 25-050-Cl, Mediatech Inc.) and accordingly distributed the day before the transfection. Transfection complexes were created using a ratio of 0.4 µg/µl of respective vector DNA to the transfection agent (Lipofectamine 2000, cat# 11668-019, Invitrogen), prepared according to the manufacturer's instructions. Following the addition of the complexes to the selected culture wells, the transfection reactions were allowed to take place over a 2.5 h span, after which they were terminated through the addition of cell culture media.
Trabecular meshwork cells were cultured as described above on the PVDF membrane. Before plating the cells, a layer of collagen matrix (∼0.1 mg per layer; Rat Tail Collagen, cat# 354249, BD Biosciences, San Jose, CA) was formed on the membrane to facilitate cell adherence. The cells were allowed to form a confluent monolayer over a period of 16–24 h, following the addition of another layer. This process was repeated to ultimately achieve a confluent tri-layer of cells. Primary TM cells of the each layer for all multilayer cell containing experiments were transfected as described previously
To demonstrate the formation of cochlin deposits in vitro, cochlin transfected TM cells were cultured on a PVDF membrane (Pall Life Sciences) with a pore size of 0.45 µM in a series of three confluent layers as described above. The membrane was then placed between the two hemi-vessels of the Ussing-type chamber connected to a standing column of serum-free cell culture media (DMEM 1X, Mediatech Inc) which was allowed to pass through the membrane by gravity, producing shear stress upon the cell layers as a result of the passage. Thus arranged setup was maintained for 12 h, following which the membrane was embedded in optimum cutting temperature medium and solidified using liquid nitrogen to permit sectioning. Obtained sections were subjected to immunohistochemical analysis probing for cochlin protein using cochlin primary antibody (usually hCochlin#3, Aves Labs Inc.), followed by respective secondary antibody incubation (A647: A21449, Invitrogen) and imaged using the confocal setup (Leica TSP5; DM6000 B).
Mice were anaesthetized as for IOP measurement and a drop of topical anesthetic (tetracaine hydrochloride 0.5%, Bausch & Lomb, Tampa, FL) was applied to the desired eye. Custom generated 2 µl anti-cochlin antibody (hcochlin#3, Aves Labs Inc.), coupled to an infra-red dye with an absorption peak at 800 nm (cat# CUST154, Rockland Inc.), was injected into the anterior chamber of mice eyes. The anesthetized animals were restrained on a mounting tube, which was fixed on a 6-axis movable platform for imaging. An ultra-high resolution (∼3 µm) SDOCT was used to scan the mouse. The mice were imaged before injection and 30 minutes post-injection using the SDOCT. This analysis was performed on 3, 6, 8 and 12 month old DBA/2J mice. The images obtained before and after injection were subtracted digitally and the resultant images were analyzed for quantifying the OCT signal intensity using NIH ImageJ (v.1.43u) software which is representative of the amount of cochlin present. We imaged 10 mice each at 3, 6, 8 and 12 months of age. The maximum signal intensity (found in mice at 8 months) was regarded as 100% and all the readings are expressed relative to this intensity.
DBA/2J-Gpnmb+/SjJ mice injected with the lentiviral vector bearing the
(TIF)
(EPS)
(EPS)
We are indebted to Dr. W. Daniel Stamer for generous gifts of primary TM cells for biological replicate experiments. We are thankful to W. Feuer for statistical analysis, Drs. Simon John and Iok-Hou Peng for helpful discussions. We thank the following lab members and collaborators for their meaningful contributions: H. Ebrahemi, E. Hernandez and G. Gaidosh.