Diabetic retinopathy (DR), a leading cause of blindness in American adults, is caused by neovascularization, the unwanted new growth of retinal blood vessels triggered by poor circulation and the resulting reduction in metabolic oxygen (see Figure 1). Several laboratory animal studies indicate that retinal oxygenation can be partially or completely restored by breathing pure oxygen.1–3 Other researchers4 have shown that humans who breathed oxygen before intraocular pressure elevation sustained their vision longer than if they simply breathed air. Despite evidence that metabolic oxygen can be beneficial for certain ophthalmic diseases such as DR, it has been hard to implement technology to provide oxygen directly to the tissue, where it is needed.
We recently developed5 a prosthesis with the potential to provide metabolic oxygen to retinal tissue in a physiologically neutral manner. This builds on state-of-the-art science and engineering in artificial sight, as well as significant progress in developing multielectrode-array retinal prostheses for patients with age-related macular degeneration and retinitis pigmentosa.6,7
Figure 1. Schematic of the eye. Poor circulation of blood in the retina, a complication of diabetes, can lead to diabetic retinopathy.
Vitreous humor, the eye's liquid medium, is mostly water. It would, therefore, be useful if one could simply electrolyze water. The anode and cathode reactions would be
while the combined reactions are
Molecular hydrogen is not a safety issue in this environment, and it is metabolically inert. However, the vitreous humor contains about 150mM of sodium chloride (NaCl) in solution. The dissolved chloride ions (Cl−) are problematic because they can be oxidized to chlorine,
When Cl− participates in the electrochemistry, the overall reaction is
This last reaction is known as the chlor-alkali process. It forms caustic soda (NaOH) and chlorine (Cl2), both of which are eye irritants. The metabolic-prosthesis technique selectively electrolyzes physiological saline to molecular oxygen and hydrogen, while simultaneously suppressing the formation of Cl2and assuring the pH (a measure of the acidity or basicity of a solution) remains within an acceptable range.
When a metal electrode is immersed in an electrolyte such as NaCl, the concentration of Cl− ions at the electrode-electrolyte interface is much smaller, almost zero, than in bulk solution. The reason for this behavior is the opposing electrostatic force between the mobile charge carriers in the metal (electrons) and in the electrolyte (Na+ and Cl−).8 In physiological saline, the degree of Cl− exclusion is somewhat less than 1nm and greater than the van der Waals diameter of water. The characteristic distance is known as the Debye length.8 A very small number of Cl− ions chemisorb onto the metal electrode. Selective physiological saline electrolysis is accomplished by charge-limited electrolysis (see Figure 2).
Figure 2. (a) Electrode-saline interface illustrating water molecules, sodium (Na+), and chloride (Cl-) ions. Cl- ions are largely excluded from the region defined approximately by the Debye length. (b) Electrochemical production of oxygen. Inset: pulsing protocol. (©IEEE 2009, used with permission.)
The shiny silver sphere in Figure 2(a) is one of the platinum electrodes. The first row surrounding the electrode is primarily composed of water molecules. Solvated Na+ ions are in close proximity. Further removed from the interface, by approximately one Debye length, are the Cl−ions. The inset in Figure 2(b) shows the repetitive pulses of positive charge that are injected into the anode. If the amount of positive charge is stoichiometrically equal to or less than the first one or two water rows adjacent to the electrode-electrolyte interface, the Cl− ions do not participate in the electrochemistry. However, as soon as positive charge is injected into the anode, the Cl− ions start moving towards the interface. If all of the positive charge is consumed by the time they arrive, no chlorine is formed. The charge distribution's relaxation time at the interface is about 6ms. The time between pulses for our experiments was 10ms, so equilibrium was reestablished following each pulse. The data of Figure 2(b) indicates that the pulse duration can be as long as 200μs before chorine begins to form.
Figure 3. One embodiment of the three-electrode pH clamp. (a) The ocular cavity contains a cathode and an anode. The external electrode (behind the ear) is a second cathode. (b) Schematic of the glass apparatus that mimicked the electrode arrangement of (a). (c) Effect of electrolysis on 2mM phosphate-buffered saline. (©IEEE 2009, used with permission.)
Real vitreous humor is not simple saline. It contains a complex mixture of organic compounds like ascorbate, lactate, and hyaluronic acid. To assure no pH drift in the intraocular cavity, we developed a pH clamp (see Figure 3) based on a three-electrode electrolysis technique. An anode and cathode are implanted into the intraocular cavity, and a second cathode is implanted behind the patient's ear. The anode reaction produces protons, making the local aqueous environment more acidic. The cathode reaction consumes protons, which make the local pH more basic. Although the oxidation of Cl− ions is suppressed by the pulsing technique, the complex nature of the vitreous humor might still produce pH drift. The three-electrode pH clamp can maintain pH within predetermined values. Figure 3(a) shows the overall electrode placement. Figure 3(b) represents a glass cell that mimics the physical dimensions of the human eye with two electrodes in the intraocular cavity and a third electrode implanted behind the ear, a technique that was introduced in the first artificial retina implant.
The main idea is that the intraocular pH can be set to any desired level by choosing the appropriate pair of electrodes. For example, if the pH is too acidic, oxygenation can be accomplished by extending the pulse duration into the chlor-alkali zone and choosing the electrode pair in the intraocular cavity. This combination—Figure 3(c)—elevates pH. Conversely, if the intraocular pH is too basic, the chosen intraocular anode will be combined with the behind-the-ear cathode to lower the pH. In a real-world surgical-implant device, pH, O2, and other measurements would be made by microsensors and electrode pairs chosen by microswitching circuitry. We emphasize that our research5 is a proof of principle. Considerably more developmental work needs to be done to reach the device level. For the immediate future we will study chlorine-free oxygen production in chemical formulations of vitreous humor and perform laboratory studies with animals. The work will then move on to the design and development of oxygenation devices for implantation studies in animals, with the long-term goal of human studies.
We acknowledge support from the Department of Energy's Office of Biological and Environmental Research and the National Academies Keck Foundation Initiatives Smart Prosthetics seed grant program.
Chemical Sciences Division
Oak Ridge National Laboratory
Oak Ridge, TN
Elias Greenbaum is a corporate fellow and leader of the Molecular Bioscience and Biotechnology research group. In collaboration with Mark Humayun, he founded the Department of Energy Office of Science artificial retina project.
Mark S. Humayun
Keck School of Medicine
University of Southern California (USC)
Los Angeles, CA
Doheny Eye Institute
Los Angeles, CA
Mark Humayun holds the Cornelius J. Pings Chair in biomedical sciences at USC. He is professor of ophthalmology, cell and neurobiology, and biomedical engineering at the Keck School of Medicine, the Doheny Eye Institute, and the USC Viterbi School of Engineering.