Ionic components of electric current at rat corneal wounds

Endogenous electric fields and currents occur naturally at wounds and are a strong signal guiding cell migration into the wound to promote healing. Many cells involved in wound healing respond to small physiological electric fields in vitro. It has long been assumed that wound electric fields are produced by passive ion leakage from damaged tissue. Could these fields be actively maintained and regulated as an active wound response? What are the molecular, ionic and cellular mechanisms underlying the wound electric currents?

Using rat cornea wounds as a model, we measured the dynamic timecourses of individual ion fluxes with ion-selective probes. We also examined chloride channel expression before and after wounding. After wounding, Ca2+ efflux increased steadily whereas K+ showed an initial large efflux which rapidly decreased. Surprisingly, Na+ flux at wounds was inward. A most significant observation was a persistent large influx of Cl−, which had a time course similar to the net wound electric currents we have measured previously. Fixation of the tissues abolished ion fluxes. Pharmacological agents which stimulate ion transport significantly increased flux of Cl−, Na+ and K+. Injury to the cornea caused significant changes in distribution and expression of Cl− channel CLC2.

These data suggest that the outward electric currents occurring naturally at corneal wounds are carried mainly by a large influx of chloride ions, and in part by effluxes of calcium and potassium ions. Ca2+ and Cl− fluxes appear to be mainly actively regulated, while K+ flux appears to be largely due to leakage. The dynamic changes of electric currents and specific ion fluxes after wounding suggest that electrical signaling is an active response to injury and offers potential novel approaches to modulate wound healing, for example eye-drops targeting ion transport to aid in the challenging management of non-healing corneal ulcers.

At human skin wounds, naturally occurring electric fields and currents were detected many years ago [1]–[4]. Modern techniques have confirmed and significantly extended our understanding of these endogenous electric fields. Electric currents of ∼10–100 µA/cm2 and fields of ∼40–177 mV/mm at wounds in humans and various animals have been measured in recent years with vibrating probes, micro-glass electrodes, micro-needle arrays, and bioelectric imagers [5]–[9].

Endogenously induced electric fields at wounds have emerged as an important signal that plays a critical role in wound healing [10]–[13]. Chemotaxis, haptotaxis (response to surface-bound molecules), contact guidance, free edges and other factors at wounds are directional cues for cell migration and wound healing. Recently, we demonstrated that electric signals in cell migration during wound healing are far more important than previously thought since they override other directional cues [6], [10], [14]. Applied electric fields have been shown to stimulate division, migration and differentiation of many types of cells: corneal epithelial cells [15]–[18], neuronal cells [19]–[21], fibroblasts [22] and skin epithelial cells [23]–[24] in a process called galvanotaxis or electrotaxis.

The mechanisms of how the electric currents are generated and maintained however remain very poorly understood and largely unstudied. It is generally assumed that the wound electric currents are caused by a short-circuit of the trans-epithelial potential difference (TEPD) at the wound site. In another words, a passive leakage of ions and other charged molecules. The TEPD is a common feature of most epithelia, including skin and corneal epithelia. In the cornea, ions are constantly being transported across the corneal epithelium, which generates and maintains the TEPD with the outer side negative [25]. Sodium ions are actively transported from the tears to the aqueous humor while chloride ions are pumped in the opposite direction, towards the tear side ( Fig. 1 ) [26]–[28]. Tight junctions between epithelial cells form a high resistance barrier to minimize ion leakage and maintain the TEPD. Damage to the corneal epithelium disrupts the tight junctional barrier and the TEPD collapses at the wound. The electrically positive TEPD under the surrounding intact epithelium thus drives a laterally-oriented electric field running from the intact cornea across the wound edge into the wound center, with the wound the negatively-charged cathode ( Fig. 1 ). Human corneal epithelial cells migrate to the cathode in vitro [15]–[18] and increasing cornea wound current pharmacologically enhances wound healing [6].

Our own and other studies have shown that wound electric currents have a dynamic time course and appear to be actively regulated [6]–[8]. We thus hypothesize that wound electric currents are an actively modulated wound response mechanism, i.e. wound tissues actively generate and regulate the wound electric currents. We set to investigate the ionic and molecular mechanisms of the endogenous wound electric currents using a cornea wound model. The corneal epithelium has multiple functions including mechanical protection from pathogens and refraction of light onto the lens and retina [29]. Appropriate optical refraction is dependent on a smooth corneal surface and the maintenance of appropriate stromal thickness. As well as the endothelium, the corneal epithelium also contributes to the maintenance of corneal transparency by preventing stromal swelling [30]. The integrity of the epithelial surface is therefore essential for normal vision. Most corneal epithelial wounds are repaired promptly by the migration of epithelial cells at the wound margin towards the center of the wound [31]–[33]. The initial response in corneal wound healing is the migration of surrounding epithelial cells into the wound bed, which is likely to be guided by the electrical signal [6], [10].

Here, we used ion-selective probes to define the ionic mechanisms of wound electric currents. We used pharmacological manipulation of ion transport to alter ion flux, and also fixed the tissues to make sure the fluxes happen only in live tissues. We also determined changes in distribution and expression of chloride transport molecules following wounding. By defining the electrophysiological properties of the cornea we will be able to understand the mechanisms of generation of the endogenous wound electric fields and offer new approaches to manage wound healing and corneal injury. Therapeutic modulation of the corneal wound ion flux in order to enhance the endogenous electric field might represent a promising new strategy towards promoting wound healing in patients with chronic non-healing lesions.

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All procedures were approved by the University of California, Davis, Institutional Animal Care and Use Committee (protocol # 07-13078).

Using ex vivo corneas allows for easier pharmacological manipulation and ion substitution of the bathing solution. We used this model as the predominant experimental system for comparision with our extensive previous data on rat and human donor cornea wound electric currents [6], [34]. The rats (Sprague-Dawley, male, weight 200–224 g) were euthanized by CO2 inhalation and cervical dislocation. The eyes were enucleated and kept in cold BSS until use.

A chamber was manufactured to hold isolated eyes for wounding and ion flux measurements. This was a 9 cm Petri dish (Falcon; BD Biosciences, San Jose, CA, USA) with two loops of nickel-chromium wire glued in so that the eye is held immobile but can be rotated and/or tilted to give full access to the whole cornea ( Fig. 2A, B ). The diameter of each loop is 5–8 mm. The loops hold the eyeball in place so the cornea can be accessed from different angles. Using a scalpel or a trephine, a wound can be made while the cornea is viewed under a dissecting microscope at ×20–×40. After taking measurements of the unwounded cornea, a wound was made under a dissecting microscope (model SMZ-10, Nikon Instruments Inc., Melville, NY, USA) by scraping the cornea with an ophthalmologic scalpel (Beaver, 22.5°, 4 mm; BD Co., Franklin Lakes, NJ, USA) to remove approximately 2 mm2 of corneal epithelium.

Borosilicate glass capillaries without filament (10 cm long, 1.5 mm outer diameter, 1.12 mm inner diameter) were purchased from World Precision Instruments (WPI, Inc., Sarasota, FL, USA; cat #TW150-4). Silanization solution I (cat # 85126), sodium ionophore II cocktail A (cat # 71178), chloride ionophore I cocktail A (cat # 24902), calcium ionophore I cocktail A (cat # 21048), potassium ionophore I cocktail A (cat # 60031) and hydrogen ionophore I cocktail A (cat # 95291) were obtained from Sigma-Aldrich (St. Louis, MO, USA). During measurements, corneas were immersed in an artificial tear solution (BSS+; Alcon Laboratories, Inc., Fort Worth, TX, USA) which contained (mM): 122.18 NaCl, 5.1KCl, 1.05 CaCl2.2H2O, 0.98 MgCl2.6H2O, 2.96 Na2HPO4, 25 NaHCO3, 5.11 D-Glucose and 0.3 glutathione disulphide.

Borosilicate glass capillaries were heat-pulled using a Sutter P-97 electrode puller (Sutter Instrument Company, Novato, CA, USA) with the following settings: heat 470, pull 13, velocity 15, delay 1, to give tips 3–4 µm in diameter. Electrodes were heated in an oven (Model 10; Quincy Lab Inc., Chicago, IL, USA) at 200°C overnight to dehydrate and then rendered hydrophobic by addition dropwise of about 3 ml of silanization solution. Electrodes were kept in the oven until all the silanization solution had vaporized. After cooling, electrodes were stored in an electrode storage jar (WPI; cat # E215) inside a glass dessicator (Fisher Scientific, Pittsburgh, PA, USA; cat # 08-595D) containing dessicant (Drierite; W.A. Hammond Drierite Co. Ltd, Xenia, OH, USA). To prepare an electrode for use, it was first back-filled using a disposable plastic Pasteur pipette (Phenix Research Products, Candler, NC, USA; cat # PP-137030) heat-pulled to a fine filament, with a 1 cm length of solution containing 100 mM of the ion to be measured. For example, NaCl is used if sodium or chloride are to be measured, CaCl2.2H20 for calcium, KCL for potassium, pH 7 buffer for protons. Pressure was applied via a 3 ml syringe with attached silicon tube to push the air-bubble at the tip out, forcing the liquid to the tip. The electrode was then tip-filled with a small amount (50–100 µm length) of ionophore specific for the ion to be measured by attaching the electrode to a micro-positioner and dipping the electrode tip into a droplet of ionophore under a dissecting microscope. Reference electrodes were glass capillaries as above, cut into 5 cm lengths and briefly fire polished in a bunsen burner flame. They were filled with 3 M of NaCl, K-acetate or KCl, with 2% agar. Agar was added to the solution of choice and brought to the boil on a hotplate to dissolve the agar.

The hot solution was drawn into the electrode with a plastic Pasteur pipette and the electrode dropped into cold KCl, NaCl or K-acetate solution. Reference electrodes prepared thus can be stored under appropriate solution in Petri dishes for several weeks prior to use. Reference electrodes must not contain the ion to be measured.

Reference electrodes were mounted in a straight microelectrode holder with a gold 2 mm male connector and Ag/AgCl pellet (WPI; cat # MEH3S) and mounted on a 3-dimensional micro-positioner (Newport Corporation, Irvine, CA, USA). Measuring electrodes were mounted in a straight microelectrode holder with a gold 1 mm male connector and Ag/AgCl wire (Warner Instruments, Holliston, MA, USA; cat # QSW-A15P) and attached to a headstage mounted on a computer-controlled electronic 3-dimensional micro-positioner (Newport). The measuring electrode and reference electrode were connected to an Ion Amp V3.0 amplifier (BioCurrents Research Center, Woods Hole, MA, USA).

As well as static measurements of ion concentration ( Fig. 2C ; see section Selective ion flux measurements below), electrodes were also used in self-referencing mode, where the electrode moves between two points 30 µm apart at low frequency (0.3 Hz) near the wound edge [35], [36]. The electrode movement was controlled by a computer running IonView32 software (BioCurrents Research Center) controlling the electronic 3-dimensional micro-positioner. The electrode pauses at each position (‘near’ and ‘far’) and the output is recorded to computer ( Fig. 2D, E ). If an ion flux is present, the electrode detects a difference in ion concentration between its two positions. The ion flux can be calculated using the formula: J = Cu(dc/dx) where C is the ion concentration in the solution, u is the ion mobility, and dc is the concentration difference over distance dx. Ion flux data are presented in pmol/cm2/sec or nmol/cm2/sec. Data were recorded onto computer using IonView32 software.

Electrodes were calibrated in standard solutions that bracketed the concentration of ion present in the BSS solution ( Fig. 2F ). For example, BSS solution contains ∼150 mM Na+, so the sodium electrode was calibrated in 10, 100 and 200 mM NaCl. Plotting the logarithm of the molar concentration against the output of the amplifier (in mV) usually gave a linear fit with an R2 value of at least 0.99 ( Fig. 2G ; R2 value of 1.0 indicates perfect fit of a linear trendline), and the formula describing the line was used to convert the mV readings from the amplifier during sample measurements into actual ion concentrations from which ion flux can be calculated.

First, the eye was mounted in the chamber (containing BSS saline at room temperature) under the dissecting microscope and allowed to equilibrate to the new temperature. The reference and measuring electrodes were placed in the bath. With the electrode at reference (bath) position approximately 1 cm from the cornea, the background value was recorded for a few seconds. The electrode was then moved into measuring position about 30 µm from the intact, unwounded cornea surface for 10–15 seconds until the new value was stable. The electrode was then moved back to reference position and the recording stopped when the stable background value was reached. The cornea was then wounded as above and measurements made at the left wound edge ( Fig. 2C ). This position was chosen since we previously showed that the current at the wound edge was significantly greater than at the wound center, indicating substantial ion pumping and/or leakage at the cut edge of the corneal epithelium [6]. This gives a single ‘peak’ (see Fig. 2C insert on right) and the difference between the background and the wound value in mV can be converted into the difference in ion concentration using the equation derived by plotting the calibration readings ( Fig. 2G ). Wound edge measurements were made at various time-points after wounding to give timelapse data of wound ion concentration change over time. We also did some long time-lapse experiments up to 6 hours.

In self-referencing mode, the oscillating electrode was first placed near the intact cornea to measure unwounded ion flux, and after wounding placed near the wound edge and measurements made in that position for up to 90 min ( Fig. 2D, E ).

Aminophylline (cat # A1755) and ascorbic acid (cat # A5960) were obtained from Sigma-Aldrich (St. Louis, MO, USA). We previously showed that drugs which affect ion flux also alter cornea wound electric current and rate of wound healing [6], [34]. Here we tested the effect of aminophylline (10 mM) and ascorbic acid (1 mM) on calcium, sodium and chloride ion flux. Eyes were kept in BSS solution at room temperature with drug for 20 minutes before beginning experiments. We found that the presence of aminophylline interfered with chloride ionophore I cocktail A, perturbing the electrode responses and giving unpredictable results. Therefore, for chloride flux measurements in the presence of aminophylline, we used a different ionophore: chloride-selective liquid ion-exchanger microelectrode cocktail A (Sigma-Aldrich; cat # 24899).

Ion-selective micro-electrode data are expressed two ways. First we show the concentration difference between the reference position in the bath solution, and the measuring position close to the cornea surface or wound edge (see Fig. 2C ). For example, Figure 3A shows calcium concentration at unwounded cornea (time 0) and at wound edge (wound edge has a relatively higher concentration). We then present data of actual ion flux, calculated from data recorded with the electrode in self-referencing mode as it oscillates close to the cornea surface (see Fig. 2D, E ). For example, Figure 3B shows calcium flux before and after wounding, showing a steady increase of calcium efflux (outward ion flow) after wounding. For each ion measured, the concentration difference data, and ion flux data, correlated. For example, higher ion concentration at wound edge = ion efflux (outward flow). Note that in all timelapse graphs ( Figs. 3 – 7 ), the measurements made at intact cornea surface before wounding are shown at time zero.

Rat cornea tissue were fixed with 4% paraformaldehyde (Sigma-Aldrich) in phosphate-buffered saline, pH 7.4 (PBS; Sigma-Aldrich) for 20 min. Cryo-sections were cut and collected on gelatin-coated slides. Non-specific binding sites were blocked with PBS containing 2% bovine serum albumin (BSA; Sigma-Aldrich), 10% donkey serum plus 0.3% Triton X-100 (Sigma-Aldrich), for 30 min at room temperature. A monoclonal antibody against CLC2 (1∶200, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA; Cat # sc-16430), was used to label the protein (1 hour at room temperature). After washing in PBS, the samples were incubated with Cy3 conjugated secondary antibody (1∶200, Jackson Immuno Research Laboratories Inc., West Grove, PA, USA) and phalloidin-FITC (1∶100, Sigma-Aldrich) for 1 h at room temperature. Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories Inc., Burlingame, CA, USA; cat # H-1200). Flourescent signals were visualized on a fluorescent microscope (Axiovert 100; Zeiss, Germany).

Expression level of CLC2 mRNA was determined using quantitative real time polymerase chain reaction (qPCR). PCR primers were designed using Primer3 online. Human CLC2 mRNA sequence was downloaded from the National Center for Biotechnology Information (NCBI) GenBank. A region that matched the consensus sequence for CLC2 was chosen for PCR primers. Primers were chosen to be 20 bp in length, with a GC content of 45–60%, no long repeats of a single base, and 150–500 bp of products length. We further confirmed lack of cross-reactivity between the 3′ end of the CLC- primers and other known genes by BLAST (basic local alignment search tool) searches. The optimized primer oligos used were synthesized by Eurofins MWG Operon (London, UK). Primary antibody against CLC2 was obtained from Santa Cruz Biotechnology, Inc. Human corneal epithelial cells were cultured in supplemented hormonal epithelial medium (DMEM/F12; Invitrogen, Carlsbad, CA, USA) in 10 cm tissue culture dishes at 37°C and 5% CO2 until confluent. Scratch wounds were made with a sterile plastic pipette tip.

We have previously described in detail the vibrating probe technique for non-invasively measuring endogenous electric current [7], [37]. Briefly, the probe is an insulated stainless-steel needle (WPI; cat # SSM33A70) with a platinum ball electroplated to the tip. The probe, mounted on a 3-dimensional micromanipulator (model H; Line Tool Co., Allentown PA, USA), is vibrated in solution about 50 µm from the cornea surface by a piezoelectric bender (made in-house) at a precise frequency (usually ∼150–200 Hz). If an electric current is present due to ion flux, the charge on the platinum ball fluctuates in proportion to the size of the current. The probe is connected to a lock-in amplifier (model SR530; Stanford Research Systems, Sunnyvale, CA, USA) that locks on to the probe's specific frequency signal. The probe is calibrated in a current density of 1.5 µA/cm2 at the start and end of experiments. Eyes were treated as above for 20 min in drug at room temperature before measurements. DIDS (4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid disodium salt hydrate) was obtained from Sigma-Aldrich (cat # D3514).

Data analysis, graphs, linear trendlines and statistical calculations were done with Microsoft Excel. Data are expressed as mean ± standard error of the mean (SEM). Differences between mean values were compared using Student's t test.

We present the data from ion-selective measurement in two ways: 1) as the ion concentration compared to background: higher or lower at the cornea surface or wound edge; and 2) as ion flux, calculated from data recorded with the probe in self-referencing mode oscillating close to the wound edge.

We first measured calcium concentration at corneal wounds. Calcium concentration at unwounded cornea surface was only slightly higher than background level in the artificial tear solution (0.015±0.002 mM). After wounding, the calcium concentration increased within the first 20 min. The level then remained stable at around 0.07 mM ( Fig. 3A , showing only up to 90 minutes). In longer experiments, this value was maintained, remaining significantly higher than the unwounded value for up to six hours after wounding (P<0.04).

We then determined calcium flux at corneal wounds using the electrode in self-referencing mode to calculate the actual calcium ion flux. We saw a small calcium efflux in unwounded corneas. After wounding, the calcium efflux increased steadily until 20 min, stabilizing at a maximum value of 5 pmol/cm2/sec, that was significantly greater than unwounded flux (P<0.02) ( Fig. 3B ).

We previously showed that pharmacological drug aminophylline significantly increased cornea wound current and enhanced wound healing [6]. Here, aminophylline (10 mM) had no effect on calcium flux ( Fig. 3B ), suggesting that aminophylline-induced enhancement of cornea electric current is not due to increased calcium flux. Amiloride, which inhibits the Na+/Ca+ exchanger and voltage-gated Ca+2 channels [38], also had no effect on calcium flux (data not shown). To confirm that calcium efflux was due to active transport, we measured calcium concentration for 30 minutes to confirm normal efflux and then fixed the cornea for 30 min in 4% paraformaldehyde. After fixation and washing, the calcium concentration had dropped to virtually zero ( Fig. 3C ).

Potassium concentration at unwounded cornea surface was virtually the same as the background level in the artificial tear solution. However, immediately after wounding there was a rapid and large increase of potassium concentration at the wound edge from ∼0 mM to over 1 mM ( Fig. 4A ). This high level of potassium then dropped quickly, reaching a lower, stable value at ∼0.5 mM from 20–90 min after wounding which remained significantly higher than unwounded values. Similarly, wounding also induced a rapid increase in K+ efflux of 150–200 pmol/cm2/sec. This large efflux of K+ however was short-lived, starting to decrease 20 min after wounding, and reached a significantly lower level of ∼50 pmol/cm2/sec ( Fig. 4B ).