Modulating the Transport Characteristics of Bruch's Membrane With Steroidal Glycosides and its Relevance to Age-Related Macular Degeneration (AMD)

Abstract

Purpose: Beneficial expectations of supplement therapies to increase the transport of nutrients, vitamins, and antioxidants across Bruch's membrane in AMD, by mass action alone, remain inconclusive. Therefore, the potential for targeting the transport pathways themselves to improve bidirectional exchange using amphipathic steroidal glycosides (ginsenosides) has been investigated.

Methods: Bruch's choroid preparations were mounted in modified Ussing chambers and basal levels of hydraulic conductivity (23 donors, age range, 12–89 years) and diffusional transport of FITC-albumin (21 donors, age range, 12–92 years) quantified. Then, following a 24-hour incubation with ginsenoside preparations, the transport parameters were re-evaluated and the resulting data analyzed with respect to aging and modulation by ginsenosides.

Results: Basal hydraulic conductivity of Bruch's showed an age-related exponential decline with a half-life of 19 years. Incubation with ginsenosides improved hydraulic conductivity with levels equivalent to donors 19 years younger. Across the age range examined, hydraulic conductivities were increased to 2.05-fold ± 0.38 (P < 0.001) of basal values. Diffusional transport of albumin across Bruch's also showed an age-related exponential decline with a half-life of 18 years. The decay curves were elevated on incubation with ginsenosides and diffusional rates were equivalent to donors 15 years younger. Diffusional rates were elevated 2.01-fold ± 0.49 over basal values (P < 0.001).

Conclusions: Transport characteristics of human Bruch's can be improved by ginsenosides, facilitating the bidirectional exchange of nutrients and waste products across the membrane. With improved transport pathways, the need for supplement therapies becomes redundant. Slowed aging of Bruch's is expected to delay the onset and/or progression of AMD.

Bruch's membrane is an extracellular matrix (ECM) that facilitates the exchange of nutrients and waste products between the choroidal blood supply and the RPE complex.¹ Aging was associated with a marked reduction in its transport capacity and in macular regions, the hydraulic conductivity of the membrane was shown to decline exponentially with the capacity for outward fluid transport halving for every 16 years of life.^2–4 

Diffusional studies on Bruch's membrane have also demonstrated gross age-related deterioration in the transport of protein-sized molecules. The permeability of serum proteins was observed to decrease from 3.5 × 10⁻⁶ cm/s in the first decade of life to 0.2 × 10⁻⁶ cm/s in the ninth decade, a decrease of over 10-fold.⁵ Similarly, over the nine decades of life examined, the permeability of a FITC-dextran probe (molecular weight 21.2 kDa, Stokes radius 3.4 nm) through Bruch's membrane declined from 1.58 × 10⁻⁵ cm/s to 1.39 × 10⁻⁶ cm/s, a decrease of over 10-fold.⁶ 

Because the outer retina is highly dependent on the choroidal circulation for delivery of nutrient products, aging of Bruch's is expected to considerably reduce the carrier-mediated transport of antioxidants, vitamins, essential metals, and lipids, undermining the protective machinery against oxidation and free radical damage in both the RPE and photoreceptors. Although this increased risk of damage is difficult to observe clinically, deficits in dark-adaptation are often noted in the healthy elderly due most likely to the inefficient transport across Bruch's of vitamin A.^7–9 In these subjects, the mechanism of diminished transport is perhaps substantiated by the fact that short-term, high-dose of vitamin A can often transiently reverse the scotopic changes.¹⁰ 

Advanced aging of Bruch's in AMD is associated with a much greater reduction in hydraulic and diffusional processes with deficits in scotopic thresholds due to inefficient transport of vitamin A being manifested in the early stages of the maculopathy.^4,6,11 Thus, the inefficient removal of toxic waste products and delivery of essential carrier mediated nutrients is hypothesized to inflict a metabolic insult that results (via inflammatory intervention) in the death of RPE and photoreceptor cells (dry AMD).¹² Approximately 10% to 20% of patients with this underlying pathology develop complications of neovascularization leading to rapid visual loss (wet AMD). 

Given the increasing elderly population, the projected increase in AMD patients, and the importance of oxidative damage in the retina, much attention has been focused on the use of antioxidant and vitamin supplementation regimes for preventive or therapeutic intervention.¹³ From a preventive viewpoint, several supplementation studies (incorporating vitamins A, C, and E, together with zinc and lutein) given to healthy elderly individuals did not provide any evidence for preventing or delaying the development of AMD.¹⁴ Supplementation studies to slow the progression of AMD have also not provided conclusive evidence in support of this strategy. The initial Age Related Eye Disease Study (AREDS) with a follow-up period of 6.3 years found some beneficial effect of antioxidant and zinc supplementation on progression to advanced AMD.¹⁵ However other, although much shorter trials, including AREDS 2 did not observe a detectable effect on progression to advanced AMD.^16,17 

The choroidal circulation is the major source of blood-borne metabolites to the outer retina, and therefore metabolite supplementation would be expected to increase the concentration gradient across Bruch's leading to greater transport by mass action alone. However, given the greater than 10-fold reduction in diffusion in the elderly and even higher in AMD, the amount of required supplementation would reach toxicologic levels and hence become unsustainable. More importantly, such supplementation strategies would not address the removal of toxic metabolites (accumulated in aging Bruch's) that are the likely source of damage in AMD. 

A more effective therapeutic strategy would be to improve the transport pathways themselves across Bruch's membrane. Potential targets would be the age-dependent deposition within Bruch's of proteins, lipids, lipoproteins, and debris from inefficient digestion of outer segment material by the RPE.^18,19 Proteinaceous content arises primarily from normal and abnormal ECM material of which damaged and denatured collagen accounts for nearly 50% of total collagen in the elderly.²⁰ These proteins (and lipid species) are often covalently modified and highly cross-linked due to oxidative and nonenzymic glycosylation reactions reducing the porosity of the membrane.^21,22 The lipid component includes neutral lipids, phospholipids, esterified and nonesterified cholesterol, and lipoproteins.^23–25 These components have been shown to be organized as 100-nm diameter particles forming a discreet sublayer within the inner collagenous layer of Bruch's and have been proposed to act like a ‘lipid barrier' that may impede transport.^26,27 The stabilization of these deposits may be facilitated by the age-related deposition of divalent metal ions, a deposition that is enhanced in Bruch's from donors with AMD.^28–30 

The above age-related changes in Bruch's are partially addressed by coupled processes of continuous synthesis and degradation of the matrix.^31,32 Degradation is mediated by a family of 24 Zn²⁺-containing, Ca²⁺-dependent proteolytic enzymes referred to as the matrix metalloproteinases (MMPs). These MMPs are released into the matrix as inactive (latent) proenzymes and on activation (by proteolytic cleavage of a small peptide) are capable of digesting all components of an ECM.^33–36 Aging is associated with increased levels of latent MMPs but with diminishing levels of activated enzymes, compromising the degradation of the membrane.³⁷ In Bruch's of AMD donors, levels of activated MMPs 2 and 9 were reduced by 50% compared with age-matched controls.³⁸ Inefficiency of the degradation system appears to be due to polymerization of MMP species, and binding and sequestration within the matrix, thereby effectively removing them from the activation process.^39,40 

Destabilizing lipid and bound proteinaceous deposits using amphipathic molecules may offer a plausible interventionist possibility. These amphipathic molecules have hydrophilic and hydrophobic domains and can readily partition into lipid structures such as liposomes, micelles, membranes, or lipoidal deposits and assist dispersal.^41–44 

Ginsenosides constitute a group of approximately 30 steroidal amphipathic glycosides found in extracts of the ginseng plant (Panax ginseng CA Meyer).⁴⁵ Structurally, they are related to the cholesterol molecule and are divided into two groups, namely protopanaxatriols (e.g., Rg1, Rh2, Re, Rf) and protopanaxdiols (e.g., Rb1, Rb2, Rc, Rd), and individual species are characterized by the type and number of sugar attachments and the attachment site of the hydroxyl group (i.e., C-3, C-6, or C-20; Fig. 1). The hydroxyl group of ginsenosides allows interaction with both the polar head group in phospholipids and the beta-hydroxyl group of cholesterol, while the hydrophobic steroid backbone can interact with the hydrophobic side chains of fatty acids and cholesterol. Ginsenosides also display transition metal chelating properties, and thus the removal of deposited divalent metals from Bruch's is also expected to destabilize the lipid deposits.^46,47 

Figure 1

 

Structural similarities between cholesterol and the ginsenosides. Rg1 and Rh1 are examples of protopanaxtriols and Rb1, Rh2, and compound K are examples of the protopanaxdiols. Glc, glucosyl (C₆H₁₁O₆).

A pilot study with purified steroidal glycosides showed that these amphipathic molecules could indeed improve the fluid transport properties of donor Bruch's membrane. More exhaustive studies were then undertaken, using a natural mixture of ginsenosides obtained from the root extracts of the ginseng plant to evaluate the potential for modulating both the hydraulic and diffusional properties of human Bruch's membrane. Some preliminary work has also been undertaken to delineate the mechanisms underlying the mode of action of ginsenosides in Bruch's membrane. 

Methods

Having obtained research ethics approval, human donor eyes for these studies (age range, 12–92 years and postmortem times 24–48 hours) were obtained from the Bristol Eye Bank (Bristol, UK). The corneas were removed for use in transplantation surgery and the remaining globes were transported to the laboratory in saline specimen jars in an icebox. Transport studies were conducted with Bruch's choroid preparations mounted in Ussing chambers and the methodology has previously been published.^2,4,6 

Tissue Preparation

Following a preliminary fundal examination with a dissecting microscope to ensure that eyes were free from disease and gross handling artifacts, a circumferential incision was made 5 mm posterior to the scleral sulcus and the remaining anterior segment, lens, and vitreous discarded. The globe was then opened in the shape of a Maltese cross with one of the quadrants housing the macular region. The four quadrants were separated and transferred to a petri dish containing Tris-buffered saline (TBS; 50 mM Tris, 0.15M NaCl, 10 mM CaCl₂, 0.02% sodium azide, pH 7.4) and the retina gently peeled away. Exposed RPE cells were then removed by brushing away with a Camel hair brush and the samples trimmed to obtain one midperipheral region per quadrant. The Bruch's choroid preparation was then isolated from the underlying sclera by blunt dissection. After thorough rinsing in fresh TBS, the preparation was floated onto 8-μm nylon filter (Sigma-Aldrich, Dorset, UK) with Bruch's facing upward. This technique has previously been shown to preserve the structural integrity of Bruch's membrane.² The preparations were then transferred for mounting into Ussing chambers for either hydraulic conductivity or diffusion experiments. 

Ginsenoside Solutions

Ginsenoside solutions were of three types: purified ginsenosides, a mixture of purified ginsenosides, and ginsenosides present in the root extract of the ginseng plant. Ginsenoside Rb1 (Sigma-Aldrich) and Compound K (Ambo Institute, Seoul, Korea) were prepared in TBS at concentrations of 200 and 190 μg/mL, respectively. A ginsenoside mixture (Sigma-Aldrich) was supplemented with Rb1 to give a concentration of 102 μg/mL in TBS with other ginsenosides (Rb2, Rc, Rd, Rg1, Rg2, Re, and Rf) present at a level of 10 μg/mL each. Finally, a 10% extract was prepared from Korean Red Ginseng (KRG; Cheonjiyang Co. Ltd., Seoul, Korea) as follows. Ten grams of KRG was dissolved in 90 mL TBS and centrifuged at 3000g to remove particulate material. The supernatant was then filtered through a 0.22-μm mesh and pH and volume adjusted to 7.4 and 100 mL, respectively, to provide a 10% working solution (KRG-WS). The content of ginsenosides in the KRG-WS was determined by high performance liquid chromatography with evaporative light scattering detection (HPLC-ELSD)⁴⁸ at the Department of Food Science and Technology in Chungnam National University (Daejeon, South Korea). The major ginsenosides were found to be (in mg/mL): Rb1, 0.46; Rb2, 0.19, Rc, 0.17, Rd, 0.11, Rg3, 0.11, Rg2/Rh1, 0.11, with lower levels of Rg1, Re, Rf, and Rb3. 

Studies were undertaken to assess the effect of ginsenosides on transport processes across Bruch's and on the release of bound MMPs, proteins, and lipids from the membrane. Gross ultrastructural integrity of the membrane following ginsenoside treatment was also assessed. The diverse nature of the studies required different experimental protocols and these are summarized in Table 1. Transport studies were initiated by firstly assessing sample integrity followed by a stabilization period of 3 hours. Basal rates of hydraulic conductivity and diffusional rate of albumin transport were then determined, the former taking approximately 6 to 9 hours while the latter required 12 hours of incubation and a further 3-hour wash to remove the FITC-albumin label prior to exposure to the ginsenosides. The morphometric protocol was identical to the procedure used to assess the effects on hydraulic conductivity. To assess the possible release of bound MMPs, proteins, and lipids, it was necessary to firstly remove endogenous unbound entities from the membrane by long perfusion periods with TBS before the effect of ginsenosides could be investigated. 

Table 1

 

Time Course of Procedures Employed to Assess the Effect of Ginsenosides on Transport Processes and release of MMPs, Proteins, and Lipids in Bruch's Membrane

 

Hydraulic Conductivity of Bruch's Membrane

Hydraulic conductivity was determined by quantifying the amount of fluid transported across the membrane in response to an applied hydrostatic pressure. The isolated Bruch's choroid preparation, mounted on the nylon support, was carefully placed over the central 6-mm diameter aperture of the bottom half of the open-type Ussing chamber (Fig. 2B). This half chamber incorporated channels providing feed for the preparation, an entry port for eluent, and an exit port leading to a pressure transducer. The top plate of the Ussing chamber (Fig. 2A) had a central aperture diameter of 8 mm, incorporating a 1-mm lip to facilitate withdraw of fluid without touching the preparation and two holes to accommodate the guiding pins protruding from the bottom plate. 

Figure 2

 

Schematic of the open-type Ussing chamber for determination of hydraulic conductivity of Bruch's membrane. The isolated Bruch's choroid preparation with Bruch's membrane facing upward was clamped directly between the two Perspex half-chambers (A, B) using the guiding pins for alignment. The lower compartment was filled with TBS taking care to remove all adherent air bubbles. A hydrostatic pressure was applied using a constant pressure-head device and from the amount of transported fluid over a given time interval, the hydraulic conductivity of the preparation was calculated.

Using the locating pins, the top plate was lowered and secured in place with three screws. The lower half chamber was then carefully filled with TBS using a syringe, with occasional tilting of the Perspex chamber to remove any trapped air bubbles from the system. Eluent reservoir and transducer lines were then connected and the hydrostatic pressure was adjusted to 1960 Pa. The preparation was initially perfused for a period of approximately 3 hours. Any defects due either to edge-damage, tears, or lacerations in the tissue resulted in perfuse leakage during this period and such preparations were discarded. 

Following stabilization, fluid in the top compartment was removed and perfusion reinstated. At timed intervals of 1 to 2 hours, the fluid entering the upper compartment was removed using micropipettes with the tip end being placed on the 1-mm lip of the chamber. The amount of fluid removed was determined by weighing the retrieved sample. Three to four fluid collections were obtained for each preparation and from the rate of fluid transported, the hydraulic conductivity (HC) of the preparation was calculated as  

 

where F is rate of flow (m³/s), A is the exposed surface area of the preparation (m²), and p is the applied hydrostatic pressure (Pa).  

After determination of basal hydraulic conductivity, the control preparations were further perfused with TBS for 3 hours and then the constant pressure-head device and pressure transducer connections were removed, chambers rinsed and filled with fresh TBS solution and transferred to a humidified incubator at 37°C for a period of 21 hours. 

Experimental chambers were perfused for 3 hours with ginsenoside containing solutions, chambers isolated from the connecting lines, rinsed and filled with fresh ginsenoside solutions, and transferred to the incubator for 21 hours. This procedure of 3-hour perfusion was instigated so as to allow entry of the ginsenosides into Bruch's membrane prior to the incubation. Following incubation, the preparations were thoroughly rinsed with TBS and after a preperfusion with TBS for 3 to 6 hours, the hydraulic conductivities of both control and experimental samples were determined as described above. 

A total of 55 donor eyes were used to assess the effect of ginsenosides on the hydraulic conductivity of Bruch's membrane. Of these, nine eyes (age range, 28–72, yielding 35 preparations) were used to assess the effect of purified ginsenosides. Thirty-eight preparations were obtained from 38 eyes (age range, 12–89 years; one macular quadrant per preparation) to assess the effect of ginsenosides present in ginseng extracts. These studies were restricted to macular quadrants so as to minimize topographical variability in hydraulic conductivity across the fundus. Four pairs of donor eyes (age range, 28–57 years; 28 preparations) were used to construct a dose-response profile using 0% to 10% ginseng extract. 

Diffusional Transport of Albumin Across Bruch's Membrane

Diffusional status was assessed by following the transport of fluorescein-isothiocyanate (FITC)-labeled albumin across the preparation in response to a constant concentration gradient. Diffusional studies were undertaken with FITC-albumin (MW 65 kDa) because its hydrodynamic radius of 3.5 nm was similar to that of most plasma carrier proteins, such as retinol binding protein, transferrin, and plasma zinc binding proteins. Diffusion studies were carried out in standard Ussing chambers as previously described.⁶ The central aperture allowing tissue exposure to the half-compartments was 6 mm in diameter. Tissue preparations floated onto nylon supports with Bruch's facing upward were carefully placed over the central aperture in one half-chamber and the other half-chamber lowered gently using guide pins to clamp the tissue, securing the chamber assembly with screws. The upright chamber was then rinsed several times with TBS. 

Tissue integrity was assessed by monitoring fluid transport (if any) across the preparation when exposed to an applied hydrostatic pressure of 2.0-cm H₂O for a period of 30 minutes. For tissues with very high hydraulic conductivities such as 140 × 10⁻¹⁰ m/s/Pa in the young,² the applied pressure of 2.0-cm H₂O (196 Pa) would result in a flow of approximately 140 μL over 30 minutes. This volume is sufficient to occupy the conduit between the two half-compartments without spilling into the opposing half chamber. Thus, presence of fluid in the opposing half chamber was taken as evidence of a damaged preparation and the sample was discarded. In reality, damaged preparations were easy to identify because flow was so rapid that levels in the two half chambers equilibrated within a few minutes. 

After determination of integrity, both half-chambers were filled with TBS, glass-encased metal stirring bars inserted and the whole assembly placed on a magnetic stirrer for a period of 3 hours to stabilize. Basal diffusional rates were then determined by replacing the buffers with 1.5 mL TBS in the choroidal half chamber and 1.5 mL 0.1 mM FITC-albumin (Sigma-Aldrich) prepared in TBS in the Bruch's facing half chamber. Both half chambers were filled simultaneously using a double-barreled syringe so as to avoid hydrostatic insults to the membrane preparation and to minimize even a transient exposure to a pressure differential across the membrane. The openings of the solution compartments were sealed with masking tape containing a pin-hole to minimize pressure differentials and to reduce evaporation. 

Following a 12-hour incubation period, the solutions in the half-chambers were withdrawn and the amount of FITC-albumin diffusing across the preparation determined by measuring absorbance at 490 nm on a Perkin-Elmer UV-Vis Spectrophotometer (Llantrisant, UK) and calculating with reference to a calibration curve prepared with the working FITC-albumin solution. The 12-hour incubation period was chosen based on progress curves (obtained in donors aged 29 and 78 years) that showed linearity of transport up to the 30 hours of incubation examined. 

Having established basal rates of diffusion, all chambers were thoroughly rinsed with TBS over a period of 3 hours. Control chambers were then filled with TBS and experimental chambers with 10% ginseng extract and the whole assembly transferred to a humidified incubator at 37°C for a period of 24 hours. Following incubation, chambers were rinsed with TBS for 3 hours and the rate of FITC-albumin diffusion reassessed as described above. 

Stability of the FITC-albumin solution is an important parameter in these experiments because its degradation leading to the release of free fluorescein would lead to erroneous interpretation of diffusional rates. Stability was assessed by fractionating the FITC-albumin present in the donor compartment after a 12-hour incubation on a Sephadex G-100 column (1.5-cm inner diameter, 30-cm length; Sigma-Aldrich) with TBS as eluent. Free fluorescein was not detected in any of the FITC-albumin solutions, and hence any absorbance detected in the half chamber opposite to the FITC-albumin solution was due to the intact FITC-albumin molecule that had diffused across the membrane preparation. 

A total of 33 Bruch's choroid preparations (age range, 12–92 years) were employed in the diffusion studies, 12 as control and 21 to assess the effect of ginsenosides on the transport process. 

Morphologic Examination of Ginseng-Treated Bruch's Membrane

Adjacent tissue samples were obtained from the same quadrant (from donors aged 69 and 79 years) and mounted in Ussing chambers. Following stabilization and determination of basal hydraulic conductivity, experimental chambers were perfused for 3-hours with 10% ginseng extract followed by a 21-hour overnight incubation. Control chambers were processed similarly but with TBS throughout. All chambers were then rinsed briefly in TBS, and following determination of hydraulic conductivity, samples were removed and fixed for 1 hour in 2.5% glutaraldehyde. After a rinse in 0.1 M sodium cacodylate containing 7.5% sucrose, samples were postfixed for 1 hour in 2% osmium tetroxide in 0.2 M sodium cacodylate. They were then dehydrated in a series of ascending levels of alcohols and embedded in Araldite. Thick (1 μm) sections were cut on a Huxley microtome (Leica, Milton Keynes, UK) and stained with toludine blue. For electron microscopy, thin sections were cut using a diamond knife on a Reichert OM4 microtome (Leica) and were viewed in a Jeol electron microscope (model 1200, Ex II; Welwyn Garden City, UK). 

Ginsenoside-Mediated Release of Proteins From Bruch's Membrane

Bruch's choroid preparations were obtained from three donors (aged 63, 83, and 97 years) and mounted to provide seven 6-mm Ussing chambers for experimentation. They were perfused with TBS at a hydrostatic pressure of 2940 Pa and fluid traversing the preparation was collected for four periods of 5 hours each. Four of the chambers were perfused with TBS for two further periods of 5 hours while the remaining three chambers were perfused with 10% ginseng solution. Proteins in the collected fluid samples were precipitated by addition of an equal volume of 12% trichloroacetic acid (TCA) and overnight incubation at 4°C. Samples were then spun at 5000 g and supernatant discarded. The pellet was washed once with 6% TCA and proteins solubilized in 0.5 mL of 0.5 M NaOH by an overnight incubation at room temperature. Protein content was estimated by the Folin-Ciocalteau method⁴⁹ and results expressed as micrograms of protein released per 100 μL collected fluid. 

A second set of eight Ussing chambers were prepared using Bruch's tissue from four donors aged 49 to 77 years and perfused as above but after the fourth collection, half the chambers were perfused with TBS and the other half with ginsenoside Rb1 at a concentration of 167 μg/ml. Proteins released during the perfusion periods were quantified as described earlier. 

MMPs in Eluted Fractions From Bruch's Membrane

Bruch's choroid samples obtained from three donors aged 49, 50, and 71 years were used to prepare 12 Ussing chambers; six designated as control and six experimental. All chambers were perfused with TBS at a hydrostatic pressure of 2744 Pa for 12 hours to remove mobile and loosely adherent proteins within the matrix. Basal hydraulic conductivity of the preparations was determined over a period of 3-hour perfusion and fluid traversing the membrane collected and stored at −70°C for later analysis (first collection). Control and experimental chambers were then perfused with TBS and 10% ginseng extract respectively for 3 hours and again the fluid crossing the preparations stored (second collection). All chambers were then incubated at 37°C for 12 hours. Following thermal equilibration to room temperature, a final assessment of hydraulic conductivity was made over a perfusion period of 3 hours (third collection). 

Collected fluid samples were then subjected to gelatin zymography to assess the potential release of MMP species (if any). Fluid samples (30 μL) were mixed with an equal volume of nonreducing 2% SDS sample buffer and 30 μL of the mixture applied to the gel lanes. Zymographic methods have been extensively described in our previous publications.^38–40 Briefly, 10% SDS-PAGE gels (1.0-mm thick) were prepared containing a 4% stacking layer and 0.1% gelatin in the separating layer. Samples for analysis were loaded into lanes together with prestained protein molecular weight markers, spanning a molecular weight range of 6 to 500 kDa (Invitrogen, Life Technologies, Grand Island, NY, USA) and 10% fetal calf serum (FCS; Sigma-Aldrich) as an internal standard to identify bands for MMPs 2 and 9. Electrophoresis was performed using the X-Cell SureLock Mini-Cell system (Invitrogen). 

After electrophoresis (150 V, 1 hour), the gels were removed from their cassettes, rinsed in distilled water and incubated for two 30 minute periods in 2.5% Triton X-100 to remove SDS and renature the proteins. They were then transferred to reaction buffer (50 mM Tris-HCI, 10 mM CaCl₂, 75 mM NaCI, and 0.02% NaN₃, pH 7.4) and incubated at 37°C for 20 hours to allow proteolytic digestion of the gelatin substrate. Gels were then rinsed in distilled water and photographed so that the position of the prestained molecular weight markers could be recorded, as after staining, their positions were masked by the presence of stained gelatin in the gels. Gels were stained with SimplyBlue SafeStain (Invitrogen) containing Coomassie G-250 for a period of 3 hours. Destaining was carried out with distilled water for 1.5 hours. 

Gelatinase activity was observed as clear bands on a blue background. These gels were scanned at a resolution of 2400 dpi (Epson 3490 scanner, Epson, Hemel Hempstead, UK), and after converting to grayscale format, colors were inverted so that MMPs were now visualized as dark bands against a whitish background. 

Lipid Composition of Bruch's Membrane Perfused With Ginseng Extracts

A screening procedure for the presence of major lipid classes in Bruch's preparations was undertaken using thin layer chromatography (TLC). Bruch's choroid samples were obtained from four donors (age: 64–75 years) and mounted to give a total of 14 hydraulic conductivity chambers. They were perfused for a period of 9 hours with TBS to remove free lipids and proteins from the preparation. Chambers were then randomly allocated to two groups and initially perfused for 2.5 hours with TBS. A second 2.5 hour perfusion was undertaken with TBS for the control group and 10% ginseng for the experimental group, and chambers transferred to a humidified incubator at 37°C for 12 hours. Following incubation, control, and experimental chambers were again perfused with TBS and 10% ginseng, respectively, for 2.5 hours. Chambers were then thoroughly rinsed with TBS and perfused with TBS for 6 hours to remove resident ginseng components from the preparation. Finally, the exposed 6-mm diameter tissue samples were cut out and sonicated in 1.0 mL water. 

Lipids present in the tissue samples were extracted by vigorous agitation with four volumes of chloroform: methanol (2:1 vol/vol). Following centrifugation of samples, the bottom organic phase was removed, evaporated to dryness over a stream of nitrogen and finally the residue was reconstituted in 50-μL chloroform: methanol (2:1 vol/vol). 

Silica gel TLC plates and all reagents for chromatography were obtained from Sigma-Aldrich. The rapid screening method for lipids developed by Pai⁵⁰ and Plekhanov⁵¹ was adopted for the analysis. This method used two solvent systems for separation of the major lipid classes; solvent system #1, chloroform: methanol: acetic acid: water (50:30:8:3; vol/vol); solvent system #2, heptane: diethyl ether; acetic acid (70:30:2; vol/vol). A lipid standard mixture was prepared in chloroform: methanol (2:1 vol/vol) containing cholesteryl oleyl carbonate (cholesterol ester, ChE, 1 mg/mL), cholesterol (Ch, 2 mg/mL), glyceryl tripalmitate (TG, 2 mg/mL), and phosphatidylcholine (PC, 2 mg/mL). Dilutions were also made to construct a calibration curve. 

Having spotted samples (40 μL) and standards (10 μL) onto 17 × 20-cm silica gel TLC plates (on an aluminum backing), solvent system #1 was run half-way up the plate, plate removed and air-dried. Full development was then undertaken with solvent system #2. Lipids were stained by immersing the TLC plate in filtered amido-black 10B solution (0.2% [wt/vol] in 1M sodium chloride) for 10 minutes followed by destaining in 1.0 M sodium chloride for 10 to 20 seconds. 

Dried plates were scanned on an Epsom 3490 photo scanner and images stored in jpeg format. Densitometric quantification was undertaken using the GelQuant.NET software provided by biochemlabsolutions.com (in the public domain) with reference to calibration curves generated from the lipid standards loaded onto the TLC plate. 

Statistical Analysis

Data is presented as mean ± 1 SD. Standard nonlinear regression analyses were performed to assess the age-dependency of fluid and diffusional transport using a commercial statistical package (Fig-Sys; Biosoft, Cambridge, UK) that used the Marquardt-Levenberg algorithm. Paired Student's t-tests were used to compare basal versus treatment levels and unpaired tests to compare between control versus ginsenoside treated samples. GraphPad Prism software was used to obtain the statistical parameters (GraphPad Software, Inc., La Jolla, CA, USA) with results being considered significant for P less than 0.05. 

Results

Effect of Ginsenosides on the Hydraulic Conductivity of Bruch's Membrane

Exposure of Bruch's to purified ginsenoside species (Rb1, Compound K, and a mixture of several ginsenosides) was associated with a significant improvement in the transport properties of the membrane (Table 2; P < 0.01). All subsequent experiments on hydraulic conductivity were carried out with 10% extracts of Korean Red Ginseng (KRG-WS). 

Table 2

 

Effect of Purified Ginsenosides on the Hydraulic Conductivity of Bruch's Membrane

 

Basal hydraulic conductivities of Bruch's membrane taken from the midperipheral region of the macular quadrant were assessed by nonlinear regression analysis and showed an age-dependent exponential decay described by the function below (Fig. 3):  

 

where A is the maximum hydraulic conductivity (22.3 × 10⁻¹⁰ m/s/Pa) and k is the decay constant (0.036). The half-life of the function is given by ln2/k and was calculated as approximately 19 years (i.e., transport capacity was halved for every 19 years of life). A semilogarithm transformation of the data is shown in the inset to Figure 3 with the regression line representing a single exponential decay function. Also shown on the plot is a failure threshold, defined as the minimal hydraulic conductivity of Bruch's required to transport fluid pumped out by the RPE. This was previously reported as 0.4 to 0.65 × 10⁻¹⁰ m/s/Pa, calculated from the rate of RPE fluid transport obtained from in vivo and in vitro human and animal experimentation and the hydrostatic and oncotic driving pressures across Bruch's membrane.^4,52 Thus, aging of human Bruch's membrane was associated with an exponential decrease in hydraulic conductivity with levels approaching failure thresholds in the very elderly.  

Figure 3

 

Basal hydraulic conductivity of Bruch's membrane in the donor samples used in the study. The data was fitted to an exponential decay function where HC = 22.3 exp^{(−0.036 × age)} and the half-life of the function was calculated as 19 years (correlation coefficient = 0.85, P < 0.001). A semilogarithmic transformation of the data showing the exponential nature of the age-related decay in transportation is given in the inset. Number of donors, 38; age range, 12 to 89 years.

Control incubations (24 hours with TBS solution) were undertaken with 15 donor preparations (age range, 22–89 years) and a paired Student's t-test showed that incubation with TBS alone was without effect on the hydraulic conductivity of Bruch's (P > 0.05; Fig. 4A). 

Figure 4

 

Effect of control (A) and 10% ginseng extract (B) incubations on the hydraulic conductivity of Bruch's membrane. (A) Control incubations (n = 15) did not alter the hydraulic conductivity of Bruch's; open and filled symbols represent basal and 24-hour incubations respectively. (B) Incubation with 10% ginseng extract (n = 23) improved the hydraulic conductivity of Bruch's (P < 0.001) effectively displacing the aging curve toward the right by 19 years; open and filled symbols represent basal and 24-hour incubations respectively. Note that the aging decay constant (i.e., the gradient) was not altered by the ginseng treatment.

Incubation with 10% ginseng extract was associated with a marked improvement in the hydraulic conductivity of Bruch's membrane (Fig. 4B; paired t-test, P < 0.001). In the age-plot of Figure 4B, ginseng treatment resulted in the translation of the regression line toward the right by 19 years whilst the intrinsic decay constants (k) remained unchanged (0.036 for basal versus 0.037 for experimental). The maximum hydraulic conductivity (A) in the age plot was increased from 21.7 to 45.12 × 10⁻¹⁰ m/s/Pa. 

Changes in hydraulic conductivity in both control and experimental samples following the incubation procedure have also been expressed as fold change compared with basal preincubation values (Fig. 5). In the absence of an age-related relationship between fold-change in hydraulic conductivity and age, the data has been pooled and the Mean ± SD calculated as 1.11 ± 0.22(15) for control and 2.05 ± 0.38(23) for the group treated with 10% ginseng extract. Thus incubation with ginseng extract was associated with an improvement in hydraulic conductivity of nearly 2-fold (P < 0.001). 

Figure 5

 

Effect of ginseng extract on the hydraulic conductivity of Bruch's expressed as fold change in basal levels. Open and filled symbols represent control and ginsenoside incubations respectively. Incubation with ginsenosides increased hydraulic conductivity to a fold increase of 2.05 ± 0.38(23) over basal levels (P < 0.001). Bars represent mean ± SD.

Dose-Response Curve

A dose-response curve was constructed using 28 preparations obtained from four pairs of donor eyes. The concentration of ginseng extract was varied between 0% and 10% and the resulting fold change in hydraulic conductivity is shown in Figure 6. A hyperbolic Michaelis-Menten type function was fitted giving half-maximal change in hydraulic conductivity at a ginseng extract level of 0.6%. 

Figure 6

 

Dose-response curve for the effect of ginseng extract on the hydraulic conductivity of Bruch's membrane. Saturation of response was observed at approximately 4% ginseng extract and the concentration for half-maximal response was 0.6% extract. Data values are plotted as mean ± SD.

Effect of Ginsenosides on the Diffusional Characteristics of Bruch's Membrane

Progress curves were undertaken with Bruch's choroid preparations from a young (29-year-old) and an elderly (78-year-old) donor. The rate of albumin transport (with a concentration difference across the sample of 0.1 mM in both tissue preparations showed linear characteristics up to the 30 hours of incubation (Fig. 7). Thus, the concentration gradient of albumin driving the diffusional process was assumed to have remained fairly constant during the incubation period. A simple calculation shows that in the younger sample, with the high rate of diffusional flux (0.33 nmol/h), over an incubation period of 30 hours, the concentration of albumin in the FITC-albumin containing half compartment was only slightly reduced from 0.1 to 0.093 mM, and hence the linearity of Figure 7. An incubation period of 12 hours was therefore used in subsequent experiments because this would minimize any change in albumin level in the donor half compartment and at the same time allow sufficient transport across the preparation for its detection by the spectrophotometric method. 

Figure 7

 

Progress curves for the diffusion of FITC-albumin across Bruch's membrane down a concentration gradient of 0.1 mM in a younger (29-year) and an older (78-year) donor.

For diffusional studies, a total of 33 donor preparations were used and the basal diffusional rates are shown as a function of donor age in Figure 8. Aging was associated with an exponential decrease in the diffusional flux of albumin across Bruch's membrane represented by the relationship shown below:  

 

A semilogarithmic plot of the data is shown in the inset to Figure 8. The half-life of the decay process was calculated to be 18 years leading to a greater than 10-fold reduction in transport capacity over a human life-span.  

Figure 8

 

Age-dependent variation in the diffusional characteristics of Bruch's membrane to protein-sized molecules. Aging was associated with an exponential decline in the diffusion of albumin and the half-life of the decay process was approximately 18 years. A semilogarithmic plot of the data is presented in the inset (n = 33, age range, 12–92 years).

Incubation of donor samples with TBS alone (12 donors, age range, 17–90 years) did not affect the diffusional status of the Bruch's choroid preparations (Fig. 9A; paired t-test, P > 0.05). Preparations incubated with 10% ginseng extract (21 donors, age range, 12–92 years) showed a highly significant improvement in the diffusional characteristics of Bruch's membrane (Fig. 9B; paired t-test, P < 0.001). In the age-plot, ginseng incubation resulted in translation of the regression line such that the diffusional status of a donor after ginsenoside treatment was equivalent to that of a donor 15 years younger. 

Figure 9

 

Effect of control (A) and 10% ginseng extract (B) incubations on the diffusional characteristics of Bruch's membrane. (A) Control incubations (n = 12, ages 17–90 years) did not alter the diffusion of albumin across Bruch's; open and filled symbols represent basal and 24-hour incubations respectively. (B) Incubation with 10% ginseng extract (n = 21, ages 12–92 years) improved albumin diffusion across Bruch's (P < 0.001) effectively displacing the aging curve toward the right by 15 years; open and filled symbols represent basal and ginsenoside incubations respectively. Note that the aging decay constant (i.e., the gradient) was not altered by the ginsenoside treatment.

The fold-change in diffusion following incubation with either TBS or ginseng extract was calculated and has been plotted as a function of age in Figure 10. In the absence of an age-related relationship between fold change in diffusion and age, the data has been expressed as mean ± SD(n). In control incubations, the change in diffusion was 1.11-fold ± 0.24(12) and in ginseng treated samples, 2.01-fold ± 0.49(21) compared with basal levels. Thus incubation with ginseng extract significantly improved the diffusional characteristics of human Bruch's membrane (unpaired t-test, P < 0.001). 

Figure 10

 

Effect of ginseng extract on the diffusional status of Bruch's expressed as fold change in basal levels. Open and filled symbols represent control and ginsenoside incubations, respectively. Incubation with ginsenoside-containing solutions increased diffusional rates for albumin transport by 2.01-fold ± 0.49(21) over basal levels (P < 0.001). Bars represent mean ± SD.

Morphologic Examination of Ginseng-Treated Bruch's Membrane

Electron micrographs of Bruch's membrane samples perfused with TBS and 10% ginseng extract are shown in Figure 11. The gross pentalaminated structural organization of Bruch's membrane was preserved after treatment with the ginseng extract. Treated samples still showed the presence of a large amount of debris and deposits within Bruch's membrane. 

Figure 11

 

Effect of ginseng extract on the structural integrity of Bruch's membrane. Control and experimental samples of Bruch's membrane were obtained from adjacent regions in the same quadrant of the fundus and perfused with TBS or 10% ginseng extract, respectively. The pentalaminated structure of Bruch's remained intact after the ginseng treatment. Considerable amounts of debris and deposits were still present in membranes treated with ginseng. (A–D) 79-year-old donor; (E–F) 69-year-old donor. BM-RPE, BM-CC, basement membranes of the RPE and choriocapillaris respectively; ICZ, OCZ, inner and outer collagenous zones; EL, elastin layer. Scale bars: 1 μm.

Ginsenoside-Mediated Release of Proteins From Bruch's Membrane

Perfusion of Bruch's choroid preparations with TBS was associated with the release of proteins but this release declined rapidly as the perfusion progressed (Fig. 12). Switching the perfusion fluid from TBS to 10% ginseng extract after the fourth period of perfusion (Fig. 12A) resulted in additional release of proteins (P < 0.001, in periods 5 and 6 compared with TBS controls). Similarly, perfusion with ginsenoside Rb1 (167 μg/mL; Fig. 12B) also resulted in increased release of proteins in the fifth period (P < 0.001), with release being reduced in the sixth period (P < 0.01). The level of released proteins was however too low to be separated and detected by standard SDS-PAGE using Coomassie Blue dye for detection (data not shown). 

Figure 12

 

Perfusion-mediated release of proteins from Bruch's choroid preparations. Bruch's choroid samples mounted in Ussing chambers were perfused for six periods of 5 hours each and the fluid traversing the preparation was collected and assessed for protein content. Perfusion with TBS was associated with declining levels of released protein. Switching perfusion after the fourth period to 10% ginseng extract (A) resulted in increased levels of released proteins. Similarly, perfusion with ginsenoside Rb1 (B) also increased the amount of proteins released from the preparation. (A) Three donors aged 63 to 97 years; seven chambers prepared of which three were perfused with 10% ginseng extract. (B) Four donors aged 49 to 77 years; eight chambers prepared of which four were perfused with ginsenoside Rb1. ***P < 0.001; **P < 0.01.

MMPs in Eluted Fractions From Bruch's Membrane

All tissue samples were initially perfused with TBS for 12 hours to remove mobile components from the preparation. Subsequent perfusion for 3 hours with TBS (first collection) showed either the complete absence or trace presence of MMP species (representative zymograms are given in Fig. 13). Continued perfusion with TBS resulting in the second and third fluid collections showed the complete absence of released MMP species (Fig. 13A). Switching the perfusion medium to 10% ginseng extract resulted in release of various MMP species (Fig. 13B). These included both latent (pro-) and activated forms of MMPs 2 and 9. An overnight incubation and subsequent perfusion (third collection) showed further release of these MMP components. Of particular note is the absence of active-MMP2 in the first collection but its presence in the second and third collections. 

Figure 13

 

Ginseng mediated release of MMPs from Bruch's membrane. Following a 12-hour perfusion with TBS, samples were either further perfused with TBS or 10% ginseng extract and the ensuing fluid transported across the membrane analyzed for MMP content by zymography. Samples perfused with TBS alone did not lead to the release of MMP species (A). In the experimental chambers (B), the first 3-hour collection with TBS showed the trace presence of MMP species. However, switching to 10% ginseng extract (second and third collections) resulted in the release of MMP species including activated forms of MMPs 2 and 9. P- and A- denote pro- or latent and activated forms of MMPs respectively. HMW1, HMW2: represent high molecular weight MMP species 1 and 2.

Lipid Composition of Bruch's Membrane Perfused With Ginseng Extracts

In these experiments, following ginseng perfusion, the hydraulic conductivity of the preparations was increased 1.83-fold ± 0.42 SD, P less than 0.005. Control incubations with TBS did not show any significant change (hydraulic conductivity = 1.18-fold ± 0.16 SD, P > 0.05). 

The lipid content of Bruch's choroid preparations isolated following perfusion with TBS and 10% ginseng is shown in Figure 14. Lipid profiles were dominated by the large presence of cholesterol esters. Visual inspection of the TLC plates showed that ginseng perfusion was associated with a slight reduction in levels of cholesterol esters and triglycerides, a much greater reduction in cholesterol, and the virtual absence of phosphatidylcholine, and unidentified lipids UL-2 and UL-3. Quantitative assessment of the changes in lipid composition following perfusion with ginseng is shown in Table 3. Ginseng perfusion effectively removed cholesterol, phosphatidylcholine, and unknown lipids UL-2 and UL-3 from Bruch's membrane. Triglycerides and UL-1 were reduced by 33% and 23%, respectively (P < 0.05). Cholesterol esters, the major lipid deposits in Bruch's membrane were only reduced by 26% (P < 0.05). 

Figure 14

 

Lipid content of Bruch's membrane after perfusion with 10% ginseng. After perfusion with TBS and 10% ginseng extract, all chambers underwent a further perfusion for 6 hours with TBS to remove contaminating ginseng. The 6-mm-diameter exposed Bruch's preparation was then cut out, sonicated in 1.0 mL water and processed for lipid extraction. Perfusion with ginseng resulted in slightly reduced levels of cholesterol esters (ChE), triglycerides (TG), and unidentified lipid UL-1 but with the virtual absence of cholesterol (Ch), phosphatidylcholine (PC), and unidentified lipids UL-2, UL-3, and UL-4. C, control perfusion with TBS; G, experimental perfusion with 10% ginseng extract. Std, lipid standard; Std-4 to Std-1, dilutions of lipid standard.

Table 3

 

Effect of Ginsenoside Perfusion on the Lipid Content of Bruch's Membrane

 

Discussion

Aging of Bruch's was associated with an exponential decrease in hydraulic conductivity with a half-life of 19 years, results compatible with previous studies.^2–4 In elderly subjects therefore, the risk of crossing the failure threshold was elevated. Under these circumstances, the pressure that needs to be generated by the RPE to drive fluid through Bruch's risks the likelihood of RPE detachments, and as such, 12% to 20% of AMD patients show these complications.²⁷ 

Basal diffusional rates for FITC-albumin across Bruch's membrane showed an exponential age-related decline with a half-life of approximately 18 years, so that the expected reduction in transport over a human lifespan was greater than 10-fold, again compatible with previous experimental studies.^5,6 In the elderly, compromised diffusional capacity for carrier sized molecules across Bruch's is expected to severely curtail the delivery of essential metals, vitamins, and antioxidants to the RPE/retina increasing their susceptibility to damage. Deficits in the transport of vitamin A in aged and early maculopathy leading to visual disturbance of dark adaptation are well documented.^8,10,11 

The mechanisms underlying the age-related demise in transport across Bruch's are poorly understood but factors that compromise the porosity of the membrane are expected to undermine its transport characteristics.^53,54 In Bruch's, these include the age-related accumulation of normal and denatured/cross-linked extracellular matrix (ECM) material, deposition of oxidatively damaged protein and lipid products, introduction of oxidative and glycosylated cross-links, and the deposition of lipids and lipoproteins.^20,23,26,55 

Quick-freeze/Deep-Etch microscopy has demonstrated the presence of specific lipoprotein containing particles (typically 60–100 nm in diameter but as large as 300 nm) in predominantly the inner collagenous layer of Bruch's membrane.^26,55 These particles appear to have a well-defined morphology with triglycerides and esterified cholesterol within the core and cholesterol, phospholipids, and proteinaceous material on the surface. Their large size argues against their entry ‘en masse' from either the RPE or choroidal compartments and the more likely explanation is the free-energy driven aggregation of constituents with the surface phospholipid enriched layer being stabilized by divalent metal ions. Distinct lipid particles begin to accumulate after 30 to 40 years of life but their precursors are present earlier and the correlation between the age-related accumulation of lipids and decrease in hydraulic conduction strongly suggests a direct relationship.^2,52 

Both lipid and protein derived aggregates are held together by weak hydrophobic or electrostatic interactions and the intact complex is trapped within Bruch's either due to the large size or by interaction with the fibrillar matrix (mediated by hydrophobic, ionic, and/or metal-dependent stabilization). Harsh amphipathic detergents such as sodium dodecyl-sulphate can release bound forms from the matrix and disintegrate the large complexes.³⁹ 

Ginsenosides are less harsh amphipathic molecules that can destabilize loosely held lipid structures such as liposomes and micelles by both direct interaction with the lipid moieties and chelation of metal ions that normally serve to stabilize the hydrophilic surface of these particles^41,47,56,57 Even in stable cellular domains associated with membrane lipid rafts, ginsenosides are able to dissociate the components within this microdomain leading to a destabilized structure^43,44,58 Their potential for interaction with deposits in Bruch's was therefore investigated. 

The limited morphologic examination of Bruch's did not appear to show any gross alterations in the ultrastructural organization of the membrane following treatment with ginsenosides (Fig. 11). The pentalaminated structure remained intact with much evidence of the persistent presence of debris and deposits. However, compositional analyses demonstrated the release of proteins and lipids. The major lipid components of Bruch's membrane, namely cholesterol esters and triglycerides, were reduced by 23% and 33% respectively on perfusion with ginsenosides (Table 3). Other components such as cholesterol, phosphatidylcholine, and several as yet unidentified lipids were completely removed. Thus the major lipid deposits in Bruch's are therefore amenable to therapeutic manipulation. 

Perfusion of Bruch's with TBS leads to the slow removal of the mobile protein components from the membrane. After clearance of this major fraction, the potential for removing the bound and trapped proteinaceous debris was assessed. Ginsenoside perfusion resulted in displacing the bound proteinaceous fraction (Fig. 12) but the amount of protein released per 6-mm diameter disc of tissue was too low to be examined by standard SDS-PAGE. However, the technique of gelatin zymography that increases detection sensitivity because of enzymic amplification, showed the release of both latent and activated forms of MMPs 2 and 9 following ginsenoside intervention. Matrix metalloproteinases are capable of degrading the extracellular matrix, and therefore have a role in both the normal maintenance of the matrix and in assisting the neovascularization process in response to VEGF-mediated stimulation of vessel growth. In Bruch's, both latent and activated forms are present and constitute the degradation arm of a coupled synthesis-degradation process that by continuously turning over the matrix serves to maintain the structural and functional characteristics of the membrane.^37,59 Aging is associated with an increase in latent MMPs but a decrease in the amount of activated MMP species and this change is thought to underlie the large accumulation of denatured collagen in elderly Bruch's membrane.^20,37 In AMD, these changes in MMP levels are much more advanced with increased levels of latent MMPs but a 50% reduction in the amount of activated species.³⁸ Since the introduction of exogenously activated MMPs 2 and 9 into donor Bruch's has been shown to substantially improve the hydraulic conductivity of the membrane,⁵⁹ the ginsenoside-mediated removal of lipids and release of activated MMPs may also be conducive for improving the transport characteristics of Bruch's membrane. The release of MMPs was from the bound fraction within the matrix and not the result of VEGF-induced overexpression and release from the RPE (because the Bruch's preparations were devoid of RPE cover). This ginsenoside-mediated increase in MMPs is therefore expected to supplement the normal turnover of the matrix. 

Rb1 is the major ginsenoside present in extracts of Korean Red Ginseng and it is partially converted to Compound K by intestinal flora.^60–62 Both ginsenosides elicited marked improvements in the hydraulic conductivity of Bruch's membrane (P < 0.01). Because of the great expense of commercially purified ginsenosides, all subsequent work was undertaken using extracts of Korean Red Ginseng having determined the spectrum of ginsenosides present and their specific levels. 

Incubation with ginsenosides improved hydraulic transport, doubling the conductivity of Bruch's membrane across the age range examined (Fig. 5). The age-plots were translated toward the right by approximately 19 years (Fig. 4B), so that the conductivities observed after ginsenoside treatment were equivalent to those of donors 19 years younger. If the same shift was to occur in AMD patients, it would delay the onset of RPE detachments and slow the progression of the disease. Diffusional transport of albumin across Bruch's was also improved significantly following incubation with the ginsenosides (P < 0.001). Age-related transport curves were shifted to the right by approximately 15 years (i.e., the diffusional status of Bruch's was equivalent to that of a donor 15 years younger). 

This in vitro study using donor human Bruch's preparations has shown that ginsenosides can improve the transportation characteristics of the membrane. In light of the decaying transport curves of Figures 3 and 8, any improvement in transportation would be considered to effectively slow the aging process. Because age is the dominant risk factor in AMD, and if these results could be translated to the in vivo situation, then slowing the age-related changes in Bruch's membrane would be expected to reduce the age of onset and progression of the disease. From the results of the present study, ginsenosides would appear to be useful candidates for assessing the potential for therapeutic intervention in AMD. 

In summary, ginsenosides improved both hydraulic and diffusional transport across donor Bruch's membrane delaying the aging process by 19 and 15 years, respectively. Thus, improvements were elicited in both the waste disposal (removal of toxic metabolites) and nutritional supply pathways allowing increased transport of carrier mediated entities, such as vitamin A, essential metals, lipids, and antioxidants. Improved supply routes across Bruch's are expected to strengthen the antioxidant machinery within the RPE and perhaps confer longer beneficial effects than the 15 to 19 year shifts of aging curves. The direct targeting of transport systems in Bruch's would appear to be a better option for therapeutic intervention rather than simply relying on the current supplementation strategies that appear to be ineffective in slowing the progression of AMD. 

Further work is now required to assess the efficacy of individual ginsenosides, to obtain dose response curves for effects on release of specific lipid classes and on release of MMPs. Results from such studies are required to assess whether a single or a mixture of ginsenosides should be used for developing a therapeutic formulation. 

Acknowledgments

The authors thank Ann Patmore, BSc (supported by NIHR), for undertaking the electron microscopy. 

Supported by grants from GBioMix Ltd., (JeonJu, Korea), the Department of Health through the award made by the National Institute for Health Research to Moorfields Eye Hospital NHS Foundation Trust (London, UK) and UCL Institute of Ophthalmology for a Biomedical Research Centre for Ophthalmology (London, UK). The views expressed in this publication are those of the authors and not necessarily those of the Department of Health. 

Disclosure: Y. Lee, P; A.A. Hussain, P; J.-H. Seok, None; S.-H. Kim, None; J. Marshall, None 

Footnotes

 Open Access Article: no