A recording chamber for small volume slice electrophysiology

Abstract

Electrophysiological recordings from brain slices are typically performed in small recording chambers that allow for the superfusion of the tissue with artificial extracellular solution (ECS), while the chamber holding the tissue is mounted in the optical path of a microscope to image neurons in the tissue. ECS itself is inexpensive, and thus superfusion rates and volumes of ECS consumed during an experiment using standard ECS are not critical. However, some experiments require the addition of expensive pharmacological agents or other chemical compounds to the ECS, creating a need to build superfusion systems that operate on small volumes while still delivering appropriate amounts of oxygen and other nutrients to the tissue. We developed a closed circulation tissue chamber for slice recordings that operates with small volumes of bath solution in the range of 1.0 to 2.6 ml and a constant oxygen/carbon dioxide delivery to the solution in the bath. In our chamber, the ECS is oxygenated and recirculated directly in the recording chamber, eliminating the need for tubes and external bottles/containers to recirculate and bubble ECS and greatly reducing the total ECS volume required for superfusion. At the same time, the efficiency of tissue oxygenation and health of the section are comparable to standard superfusion methods. We also determined that the small volume of ECS contains a sufficient amount of nutrients to support the health of a standard brain slice for several hours without concern for either depletion of nutrients or accumulation of waste products.

in vitro electrophysiology requires the surgical extraction of the neural tissue of interest from the whole organism and keeping it alive and healthy in artificial extracellular solution (ECS) for the duration of the experiment (Kerkut and Wheal 1981; Yamomoto and McIlwain 1996). The ECS supplies energy sources, various ions present in a typical extracellular environment, and other nutrients and is constantly oxygenated by pumping either pure oxygen or carbogen (a mix of oxygen and carbon dioxide) into the solution (Molleman 2003).

Unlike interface type chambers, also known as “Haas” type chambers (Haas et al. 1979), submerged brain slice electrophysiology setups allow for the constant immersion and superfusion of the tissue with oxygenated ECS during the actual electrophysiological recordings (Nicoll and Alger 1981). To do this, the recordings are performed in a small recording chamber mounted in the optical path of the microscope. This recording chamber is connected to an external storage container or a bottle holding fresh ECS via a tube (Molleman 2003). ECS is oxygenated in the external container and then fed into the recording chamber (= inflow) at a constant rate, typically several milliliters per minute. After the solution has passed over the tissue section, it is led out of the recording chamber through a second tube (= outflow) and into a waste container (White et al. 1978; Kerkut and Wheal 1981; Molleman 2003). With this type of setup and a flow rate of several milliliters of ECS per minute, one single slice experiment may require 1 liter of ECS or more, depending on the duration of the experiment. As long as plain ECS without additional pharmacological agents is used, such an open-circuit flow system works very well, and large quantities of ECS consumed are not a concern since the components of ECS are readily available and relatively inexpensive.

However, for experiments that require the addition of expensive and/or rare chemicals to be added to the ECS, using an open-circuit system that continuously discards ECS may be much too costly. When small volumes of solutions are needed, for example to use a valuable chemical sparingly, most investigators close the flow circuit, recycle the ECS, and reduce the total volume of solution that circulates in the flow system by choosing smaller holding containers, tubes, etc. With this standard approach we managed to reduce the total volume of solution in a closed circuit flow system to ∼7 ml.

To further reduce the total volume of ECS required for a brain slice experiment, we designed a new type of recording chamber that reduces the volume of the total required ECS solution to 1.0–2.6 ml. The key difference between this new chamber and closed-circuit flow systems is that our new design eliminates tubing and the holding container altogether. The ECS in the recording chamber is recirculated and oxygenated directly inside a small compartment within the microscope-mounted recording chamber.

Dissolved oxygen contents of the ECS in the chamber were measured to ensure that the liquid pump mechanism oxygenates the ECS sufficiently to support tissue health. We also performed a series of liquid chromatography and mass spectrometry analyses on samples of ECS that were either fresh or had been incubated with a brain slice in the chamber for several hours to test for potential depletion of nutrients or accumulation of waste products. The overall health of neurons in brain slices that were kept in the chamber for extended periods of time was assessed with images of the tissue and physiological recordings, confirming that the small-volume recording chamber is capable of maintaining the health of tissue sections for periods of time that are comparable to established superfusion systems.

MATERIALS AND METHODS

Chamber Design and Modeling of Fluid Flow

Technical drawings of several versions of the chamber were made using SolidWorks 2013 × 64 Edition (Dassault Systems SolidWorks, Waltham, MA), a computer-aided design (CAD) program (Figs. 1 and 2). During this step of the design process, the main goal was to create a recording chamber that can be mounted into the optical path of an upright microscope, for example, as a fixed-stage system. Additional criteria were that the chamber should provide sufficient space to hold a typical rodent brain slice and allow for sufficient maintenance of tissue health. Circulation and oxygenation of the chamber were achieved by designing a gas-propelled liquid pump. The principle of this type of pump is that gas that passes through a tube opening into fluid creates pressure gradients in the fluid surrounding the opening. Positive pressure will propel the fluid forward and away from the tube opening while negative pressure behind the tube opening will draw in new liquid. We designed the system such that the liquid drawn by negative pressure mixes with the gas and provides oxygenation of the liquid. Gas pressure is the main factor propelling the circulation of the bath solution, as well as providing oxygenation. For the gas-liquid pump to work properly, a certain minimum pressure is required. We designed the chamber such that the jet of liquid and gas produced by the pump does not impact the brain section directly but rather is reflected by two curved surfaces both on the outer rim of the chamber and on a central “peninsula” that softens the jet, disperses it, and ensures even and soft fluid flow across the part of the recording chamber that holds the brain section. Importantly, the second surface along this peninsula divides the single dispersed flow into two slower streams around the outer edge of the chamber and around the brain section and completes the circulation by combining both fluid streams and guiding them toward the fluid return that suctions fluid into the pump area (Fig. 3).

Fig. 1.

Combining both slice superfusion and solution oxygenation directly in the recording chambers allows for a significant reduction of the total extracellular solution (ECS) volume required for an in vitro experiment. A: design of the small volume recording chamber showing an angled, near top view of the chamber. B: expanded view of the 3 parts that need to be assembled into a functional chamber. These are the main chamber (B1); the glass coverslip serving as the chamber's bottom (B2); and the gas inflow tube that needs to be inserted into one of the shorter side walls of the main chamber (B3). See materials and methods for the technical specifications of the parts. Scale bar = 24 mm.

Fig. 2.

Technical drawing of the chamber including all measurements in millimeters, and all design features with panels showing the chamber from all six sides. A: side view of the side opposite to the air inflow. B and D: side views of the longer side walls. C: side view of the side that contains the air inflow. E: top view. F: bottom view.

Fig. 3.

Fluid flow speeds varying over 3- to 4-fold range support the signature figure 8 flow pattern of the chamber while minimizing fluid velocity at and near the tissue section. Modeling the flow of solution in the chamber with different flow speeds at the solution outlet port (arrows in A and B). A: lowest fluid velocity (measured at the entrance to the chamber) that still produces the figure 8 liquid circulation, 0.3 ml/s. B: maximal fluid velocity that supports the figure “8” flow pattern. C: plot of various liquid velocities along a longitudinal cross section through the 30-mm long tissue placement compartment of the chamber, based on various initial fluid velocities at the fluid outlet port.

Once the design of a preliminary version of the chamber was completed, the liquid flow of ECS within the chamber was modeled (Girault and Raviart 2011). To do this, the SolidWorks 2013 Add-in Flow Simulation 2013 was used, which solves the Navier-Stokes equation using finite-element numerical analysis to calculate the range of ECS fluid velocities at which optimal circulation could be obtained. To simplify the modeling process, we assumed certain flow speeds of fluid (not air/fluid mix) at the outlet to the main chamber containing the brain section (Fig. 3, A and B, red arrows). Furthermore, “solid plastic” was chosen as the active surface parameter and (plain) water as the circulating solution. The modeling program also added a sealed top surface to the chamber and thus created an enclosed space in which water flow was modeled. Once the model was created, it was run with a range of fluid speeds to test how robust the signature fluid flow described above was with varying fluid speeds. We aimed for a chamber design in which the fluid circulation would be stable and function as designed for a range of fluid speeds yet produce uniform flow around the brain section area. Based on the results from the modeling process, the chamber design was modified, and fluid flow was tested again. This process was repeated several times until a chamber design was found that met all the criteria discussed above.

Rapid Prototyping

Once a promising chamber design was found, it was prototyped. Rapid prototyping was done on an Objet24 3-D printer running software version 24.1.1.5896 (Objet Geometries, Eden Prairie, MN) using Objet proprietary polymer material FULLCURE 835 VeroWhitePlus. The chamber design had an open bottom when prototyped (Fig. 1, A and B). This open bottom was then covered by a cover glass (Fisher Scientific cat. no: 12–548-5P, glass size 24 × 60 × 1 mm), which is attached and sealed around the edges of the chamber with clear silicone. During brain slice recordings, this glass bottom allows for illumination of the tissue section with transmitted light. Furthermore, the chamber has a small opening on one side, into which the gas delivery tube is inserted (object #2 in Fig. 1; peek capillary tubing 1/16″; OD, 0.020″; ID, 5″; Thermo Scientific product no: 37020).

During the design process, several chamber designs were prototyped and tested for functionality as described below. Based on the results, the design of the chamber was optimized further in SolidWorks, followed by a set of fluid flow modeling experiments on the optimized design, and further modification in SolidWorks. This process designing-modeling-prototyping-testing was repeated several times until the final design of the chamber was determined.

Measurements of Dissolved Oxygen

We used two independent methods to measure dissolved oxygen of the solution to test the efficiency of the pump mechanism to oxygenate the circulating ECS. The first method was a chemical kit for measuring dissolved oxygen (CHEMets Kit; Chemetrics Water Analysis System cat. no. K-7512; Chemetrics, Midland, VA). This kit uses the indigo carmine method (Wilkin et al. 2001; Gilbert et al. 1982) and is capable of measuring the dissolved oxygen at a range of 1–12 mg/l through a chemical reaction and a colorimetric readout. The second method was based on a handheld dissolved oxygen meter (MW 600, Dissolved Oxygen Meter; Milwaukee Instruments, Rocky Mount, NC). This meter uses the polarographic method (Clark and Sachs 1968; Clark 1970) and is capable of measuring dissolved oxygen ranges of 0.0–19.9 mg/l. For measurements that exceeded the range of the MW 600, the solution to be measured was diluted with (boiled) low oxygen water at various ratios, and the final dissolved oxygen concentration of the test solution was calculated from the measured values of the mixed samples and the mixing ratios. All measurements of dissolved oxygen performed with either method are reported as the average and means ± SE.

In Vitro Electrophysiology

The capability of the recording chamber to maintain healthy tissue sections over a period of several hours was tested. Specifically, we tested the capability of the chamber's aeration system to provide sufficient amounts of dissolved oxygen/carbon dioxide to keep the tissue alive and healthy over several hours. To this end, we maintained rodent brain slices for various periods of time in the recording chamber and performed patch-clamp recordings from neurons in these slices. In some experiments, synaptic afferent fiber bundles were stimulated electrically to increase the metabolic demands of the section. All animal procedures were approved by the University of Colorado School of Medicine Animal Care and Use Committee and strictly followed all applicable rules and regulations.

Slice preparation.

Coronal slices of brainstem were prepared from the auditory brain stem of six Mongolian gerbils (Meriones unguiculatus) ranging in age from postnatal days (P) 7 to 15. Animals were briefly anesthetized through isoflurane inhalation (IsoFlo; Abbott Laboratories, Abbott Park, IL) and decapitated. The brain was dissected out under ice-cold dissection Ringer containing the following (in mM): Ringer 1: 125 NaCl, 2.5 KCl, 1 MgCl₂, 0.1 CaCl₂, 25 glucose, 1.25 NaH₂PO₄, 25 NaHCO₃, 0.4 ascorbic acid, 3 myo-inositol, and 2 pyruvic acid (all chemicals from Sigma-Aldrich, St. Louis, MO), bubbled with 5% CO₂-95% O₂. Sections of 250 μm were cut on a vibratome (VT1000S; Leica Microsystems); transferred to an incubation chamber containing ECS containing the following (in mM): 125 NaCl, 2.5 KCl, 1 MgCl₂, 2 CaCl₂, 25 glucose, 1.25 NaH₂PO₄, 25 NaHCO₃, 0.4 ascorbic acid, 3 myo-inositol, and 2 pyruvic acid (all chemicals from Sigma-Aldrich); and bubbled with 5% CO₂-95% O₂. Slices were incubated in ECS for 30 min at 37°C and then cooled down to room temperature. All measurements were obtained within 4 h of slicing.

Whole cell recordings.

After incubation, slices were transferred to a prototype of the small volume recording chamber and continuously oxygenated with 5% CO₂-95% O₂ at a pressure of 70 mbar, delivered from a standard size 200 gas tank (Airgas USA, Denver, CO) through a standard two-stage carbogen gas regulator (VWR cat no. 55850388, Medical Oxygen Mix 50 PSIG; VWR, Radnor, PA). All recordings were done at room temperature. Previously characterized auditory brain stem neurons from the medial nucleus of the trapezoid body (MNTB; Borst and Soria van Hoeve 2012; Schneggenburger and Forsythe 2006; von Gersdorff and Borst 2002) were identified through a Zeiss Axioskop 2FS plus microscope with Dodt optics and an Achroplan ×40 water-immersion objective (NA 0.8; Zeiss, Oberkochen, Germany). The same microscope was also used to acquire the images of brain sections (see Fig. 6).

Fig. 6.

Electrical stimulation of brain slices did not deteriorate tissue health. Micrographs of 3 gerbil brain stem slices containing medial nucleus of the trapezoid body (MNTB) neurons. Images were taken at t = 0 (A and C), after 60 min of electrical stimulation (B, D, and E), and after 120 min of electrical stimulation (F). E and F also show a glass pipette that was used for additional patch-clamp recordings. The ages of the animals were postnatal day (P) 11 in A and B; P14 in C and D; and P15 in E and F. Scale bar = 20 μm for A–F.

Whole cell recordings were made with a HEKA EPC 10 double amplifier (HEKA Instruments, Lambrecht/Pfalz, Germany). Signals were filtered at 5–10 kHz and subsequently digitized at 30 kHz using Patchmaster Version 2 × 67 software (HEKA). Patch pipettes (2.4–4.0 MΩ) were pulled from 1.5-mm borosilicate glass (Harvard Instruments, Kent, UK) using a DMZ Universal Puller (Zeitz Instruments, Munich, Germany) and filled with internal solution (in mM): 113 K-gluconate, 4.5 MgCl₂, 9 HEPES, 5 EGTA, 14 Tris2-phosphocreatine, 4 Na₂-ATP, 0.3 Tris-GTP, and 1.5 CaCl₂, pH adjusted to 7.25 with KOH (all chemicals from Sigma-Aldrich).

Stimulation of synaptic inputs.

To increase the metabolic demands of neurons in the brain section, synaptic currents were elicited by midline stimulation of the excitatory calyceal input fiber bundle with a 5-MΩ bipolar stimulation electrode (matrix electrodes with 270-μm distance; FHC, Bowdoinham, ME). Stimuli were 100-μs-long square pulses of 20 V delivered with an STG 2004 computer-controlled four-channel stimulator (Multi Channel Systems, Reutlingen, Germany) and a stimulation isolation unit (Iso-flex; AMPI, Jerusalem, Israel). Stimulus trains consisted of 1-h long sequences of 20-Hz Poisson-distributed activity, which was restarted manually as soon as the sequence was completed (Hermann et al. 2007).

Cell Imaging at Physiological Temperature

Images of brain slices at 37°C (see Fig. 7) were obtained using a 3I Marianas system based on a Zeiss Axio Observer Z1 confocal microscope with attached CoolSNAP HQ2 camera. We used the wide-field imaging with dark field mode. This microscope was equipped with Okolab cage incubator with temperature, gas, and humidity control. We connected a carbogen gas tank to the built-in humidifier and delivered a humidified air mixture to our in-vitro chamber.

Fig. 7.

Tissue could be successfully maintained at physiological temperature. Micrographs of a gerbil brain stem slice containing MNTB neurons that was incubated at 37°C. A: t = 0; B: t = 120 min. As a control, the aeration mechanism was turned off in C and D. The age of the animal was P11. Scale bar = 20 μm for A–D.

Humidification

We used two approaches to increase the humidity during the electrophysiological recordings to reduce evaporation of ECS. First, we rehumidified the O₂/CO₂ mixture delivered to the chamber. We used a custom-built system that consisted of inlet and outlet tubes inserted into a standard 50-ml Polypropylene conical tube (Falcon; Corning Brand cat. no. 352070) through two holes in the lid of the tube. Gas from the carbogen tank was delivered via the inlet tube, passed through a bubbling stone into 20 ml of water within the Falcon tube, and exited as humidified gas via the outlet tube and into the recording chamber. The second method to increase the ambient humidity in the vicinity of the experimental set up was by using a household ultrasonic cool mist humidifier (Crane model EE-3186) and directing the humidifier's stream of moist air at the recording chamber.

Liquid Chromatography and Mass Spectrometry

Sample preparation.

Samples were stored at −70°C before analysis. Samples were thawed on ice, and aliquoted into autosampler vials for liquid chromatography and mass spectrometry analysis. Methanol extracted vs. “neat” samples showed no differences; therefore, samples were injected “neat” with no preparation.

Liquid chromatography.

Samples were analyzed using normal-phase chromatography on an Agilent 1290 series pump using a Phenomenex Kinetex HILIC, 2.6-μm, 100-Å (2.1 × 50 mm) analytical column and an Agilent Zorbax Eclipse Plus-C8 5-μm (2.1 × 12.5 mm) narrow bore guard column. One microliter of sample was injected onto the column with mobile phase A (50% ACN with pH 5.8 ammonium acetate) and mobile phase B (90% ACN with pH 5.8 ammonium acetate) at a flow rate of 0.6 ml/min. Gradient elution was as follows: 0–2 min 100% B, 2–2.1 min 100-90% B, 2.1–8.6 min 90-50% B, 8.6–8.7 min 50-0% B, and 8.7–14.7 min 0% B, followed by column reequilibration. Autosampler tray temperature was 4°C and column temperature was 20°C.

Mass spectrometry.

A nontargeted analysis was performed by generating extracted ion chromatograms for each compound and comparing the resulting area under the curve to determine whether nutrient abundances were significantly different in the various samples (Moco et al. 2007; Dettmer et al. 2007). This nontargeted analysis determined potential differences in the abundance of a nutrient ingredient caused by the incubation of the brain slice using relative quantitation (Armstrong et al. 2007; Hori et al. 2011; Evans et al. 2014).

All samples were run in full scan mode on an Agilent 6520 Quadrupole Time-of-Flight (Q-TOF; Agilent Technologies, Santa Clara, CA) with the following parameters: negative ionization mode with ESI source, mass range: 50–1700 m/z, scan rate: 2.22 spectra/s, gas temperature: 300°C, gas flow: 10.0 l/min, nebulizer: 30 psi, skimmer: 60 V, capillary voltage: 4,000 V, fragmentor: 120 V, and reference masses: 119.036 and 966.007 (Agilent reference mix). Vitamin C, pyruvic acid, myo-inositol, and glucose were confirmed using mass spectrometry/mass spectrometry at 10, 20, and 40 eV with a 500 ms/spectra acquisition time, 4 m/z isolation width, and 1-min delta retention time. Additionally, the instrument's response to solutions with a range of known concentrations of glucose (0.01 to 100 mM), pyruvic acid (0.02 to 20 mM), myo-inositol (0.03 to 30 mM), and ascorbic acid (0.04 to 40 mM) was tested, confirming that the sensitivity of the instrument was in the required range.

Data extraction and analysis.

Spectral data from full scan mass spectrometry was extracted using Profinder software (Agilent Technologies) to perform mass and retention time alignment. Spectral peaks were exported into Mass Profiler Professional (Agilent Technologies) to subtract out sample preparation and instrument blanks. Statistical analysis was performed using ANOVA with Benjamini Hochberg false discovery rate multiple testing correction ≤0.05 and fold change analysis >1.2.

Compound identification.

Compounds were annotated based on exact mass and isotope ratios using an in-house database comprising Metlin, Lipid Maps, KEGG, and HMDB. To confirm compound identifications for glucose, myo-inositol, vitamin C, and pyruvic acid, the tandem mass spectrometry peaks were matched to METLIN and NIST14 MS/MS spectral libraries.

Supplemental Information

A complete set of technical drawings with all measurements is available for download at www.slicechamber.org. A patent application regulating the commercial use for this device is pending (application no. US 14/483,883, publication no. US20150072372 A1).

RESULTS

Our main goal was to design a chamber for electrophysiological recordings that could provide oxygenation and circulation of the ECS within the microscope mounted recording chamber and thus eliminate external holding containers and connecting tubes. This approach has allowed us to significantly reduce the total volume of ECS required to support ECS flow. The resulting chamber is a recording chamber consisting of 1) a compartment into which a typical rodent brain slice can be placed, arrested under a standard slice hold down (“harp”), and viewed on an inverted or upright microscope; 2) an efficient circulation system that can superfuse the section with ECS and support oxygen and nutrients at rates that are sufficient to maintain tissue health; 3) an air mixing compartment that can oxygenate/carbogenate the ECS with an efficiency that is sufficient to maintain tissue health; and 4) with a total required fluid volume of the entire system ranging between about 1 and maximally 3 ml. We accomplished these objectives by creating a self-contained chamber (Fig. 1A) consisting of a single object that is 3D printed (Fig. 1B, object 1), with areas dedicated to solution inlet, oxygenation, circulation, and outlet and a second area for tissue section placement. Additional elements (Fig. 1B) include a standard microscope cover glass to be attached and sealed to the bottom portion of the chamber (Fig. 1B, object 2), and a tube for gas delivery (Fig. 1B, object 3), to be inserted into the chamber through a predesigned port. Importantly, we eliminated external containers, bottles, and tubes as well as the need for an external pump or gravity system to circulate the solution.

As described in materials and methods, the design process of this chamber was iterative, meaning that we designed preliminary versions of this chamber in SolidWorks, modeled the solution flow in these chambers, optimized the design accordingly, and prototyped designs that were promising based on the modeling data. These prototypes were then experimentally tested for fluid flow, oxygenation capabilities, and general usability, after which further modifications in the design were performed, followed by further modeling, prototyping, and experimental testing. In this report, we only present the data obtained from modeling the flow dynamics and from experimentally testing the final version of the chamber (Fig. 1); data obtained from testing preliminary designs that were ultimately rejected are not presented here. Figure 2 shows technical drawings of the final chamber design.

Fluid Flow Modeling Determines a Signature “Figure 8” Pattern of Fluid Flow in the Chamber

Fig. 8.

Chromatograms of compounds in ECS using mass spectrometry analysis. A: liquid chromatography-mass spectrometry (LC-MS) total ion chromatogram showing overlap of chamber buffer at T0 (n = 3), T10min (n = 3), and T3hr (n = 3). Levels of glucose and myo-inositol showed no statistically significant changes after 3 h (ANOVA, P = 0.455 and P = 0.367 respectively). B: LC-MS extracted ion chromatogram for ascorbic acid. Ascorbic acid was detected at T0 but it was below the instrument detection limit at T10min and T3hr. C: LC-MS extracted ion chromatogram for pyruvic acid. Pyruvic acid showed no statistically significant change over time, ANOVA, P = 0.183.

Modeling of the solution flow in the chamber resulted in defining the lowest possible (Fig. 3A) and the highest possible (Fig. 3B) fluid flow rates that support proper circulation of the liquid in the chamber. Between outlet flow rates of 0.3 ml/s (Fig. 3A) and 1 ml/s (Fig. 3B), the solution flows along the chamber walls with the outlet stream splitting into the upper and lower basins, producing a signature “figure 8” shaped circulation. Modeling data suggest that fluid flow rates between 0.3 and 1 ml/s, measured at the outlet port (Fig. 3, A and B, red arrows), can support appropriate circulation. The graph in Fig. 3C summarizes the velocities of the modeled liquid flow at assumed points along the midline of the slice placement compartment resulting from various flow rates at the outlet ports. Note that these flow rates seem high along the chamber wall (24.3 to 87.2 mm/s maximum for 0.3 to 1 ml/s outlet flow rate), because the diameter of the outlet port is very small (2.5 × 2.5 mm) and solution exiting this port is immediately dispersed over a much larger surface area. The flow rates are noticeably slowed down (2.1 to 7.3 mm/s maximum for 0.3 to 1 ml/s outlet flow rate) in the fluid current in the region where the section is placed (approximately between position 12 mm and position 22 mm along the midline axis), preventing potential mechanical disturbances to the tissue during recordings. For comparison, we calculated that in an open circuit flow system the laminar flow rate along the chamber is 2.56 mm/s for an input liquid volume of 12 ml/min and this flow rate is similar to that around the brain section in our new design. These modeling data suggest that the chamber design supports appropriate superfusion and circulation of the bath solution and the maintenance of proper fluid circulation for inlet flow rates that vary over at least a three- to fourfold range. Fluid movements around the tissue section are much slower than near the fluid outlet and the edges of the chamber. This design allows for efficient support of the section with fresh oxygenated ECS while at the same time keeping fluid currents near and above the section to a minimum.

Measurements of the Relationship between Gas Pressure and Liquid Flow Suggest a Three- to Fourfold Range of Flow Speeds that Maintain the Flow Pattern

Since fluid flow and circulation in the chamber, the metrics used in the modeling studies described above, depend on the pressure of the gas entering the chamber and the resulting pump efficiency, we measured the range of gas pressures that supported the figure 8 fluid movement described above. Gas pressure was measured and regulated by a standard commercial two-stage regulator (VWR cat no. 55850388) attached to a standard size 200 carbogen tank. The accuracy of the readout of this regulator was additionally confirmed with a HMG01 pressure meter (ATP Instruments). We define the optimal pressure as the pressure of gas needed to circulate and maximally oxygenate the fluid volume in the chamber. We found the optimal gas pressure to be 65–70 mbar, measured at the outlet of the second stage of the regulator, just before the gas enters the tube inserted in the chamber. This pressure is appropriate to propel fluid in the circular motion and oxygenate water to the levels described in the section below.

To compare the flow rates obtained with certain gas pressures to the flow rates calculated above, we recorded the flow of a small floating particle (a piece of yarn) along the walls of the chamber and analyzed the flow speed of this particle when propelled with various gas pressures. Gas pressure was adjusted until the resulting flow speed corresponded to flow speeds described above. All pressure measurements and applicable volumes are summarized in Table 1. Note that gas pressures between 30 and 110 mbar, a three- to fourfold range, support proper fluid circulation in the chamber and that different air pressures supported different total fluid volumes (Table 1).

Table 1.

Gas pressures and supported total fluid volumes that produce the figure 8 circulation pattern

Air Pressure at Inflow, mbar	Supported Total Fluid Volumes, ml
30	1
50	1.5
70	1.5, 2.0
110	1.7, 2.5, 2.6

Input gas pressure values required to optimally circulate liquid in the chamber, the range of gas pressures that support the figure 8 solution flow, and total volumes of extracellular solution that can be used in the chamber to support flow at these gas pressures are listed. Note that the optimal solution volume depends on the gas pressure delivered.

Dissolved Oxygen Measurements Suggest that the Oxygenation Mechanism Can Achieve Sufficient Saturation with Oxygen

We measured the efficiency of the recording chamber in enriching the solution it contains with dissolved oxygen. We used the CHEMets colorimetric measurement kit in the first set of our experiments. We measured the dissolved oxygen content of boiled water to obtain a baseline. During the boiling process, most of the oxygen dissolved in water escapes, allowing the boiled water to cool back to room temperature in a tightly closed container largely maintains a low oxygen content. The bottled water measured a dissolved oxygen content of 3.3 ± 0.17 mg/l (n = 3). Deoxygenated water was then filled into the recording chamber and bubbled with carbogen supplied via the water pump mechanism for 10 min. With the colorimetric method we found that after 10 min of water circulating in the chamber, the dissolved oxygen content was at 12 mg/l (average 12.0 ± 0 mg/l, n = 3), suggesting that the solution had reached 176% of the atmospheric dissolved oxygen equilibrium at the laboratory's particular altitude (6.8 mg/l at an altitude of ∼1,500 m, ∼620 mmHg at 24°C; Radtke et al. 1998). As a control experiment, deoxygenated water was placed in the recording chamber and allowed to passively oxygenate with atmospheric oxygen (no bubbling). After 14 min of passive oxygenation, this solution only reached 5.3 ± 0.17 mg/l (n = 3). Thus active oxygenation with the water pump method saturated or even supersaturated the solution with dissolved oxygen, while passive oxygenation resulted in less than half of the dissolved oxygen content compared with the active method.

The indigo carmine method used above only allows for measurements of dissolved oxygen up to 12 mg/l, a value that was reached with 10 min of active oxygenation. Under natural conditions when the solution is oxygenated with standard atmospheric air, the dissolved oxygen content of solutions rarely if ever exceeds this value. However, when ECS is bubbled with either pure oxygen or carbogen, which contains 95% oxygen, the oxygen partial pressure is significantly higher than the 0.21 atmospheres that are contained in atmospheric air, and thus the saturation point for dissolved oxygen can be higher than 12 mg/l (= supersaturation).

We therefore repeated the dissolved oxygen measurements with a second, polarographic method, where a platinum electrode (cathode) measures a flux of oxygen across an oxygen permeable Teflon membrane (Clark and Sachs 1968). We used a Milwaukee Instruments MW 600 oxygen meter, which has a range of 0 to 19.9 mg/l dissolved oxygen, and repeated the experiments described above. This second set of experiments produced similar results to those observed with the first method (average ± SE = 10.6 ± 0.9 mg/l; n = 3), when the solution was oxygenated for 10 min. We also confirmed that oxygenation with the water pump method enriched the solution with oxygen faster than simple exposure to atmospheric oxygen would (5.6 ± 0.2 mg/l) after 18 min of exposure to atmospheric oxygen.

However, when the water pump mechanism was allowed to run for longer than 10 min, the oxygen content in the solution increased further and reached ∼18.7 mg/l after 50 min (Fig. 4). Thus the water pump mechanism was capable of producing a dissolved oxygen concentration at 275% of the value that is considered saturated for the altitude (1,500 m above sea level). In brain slice experiments using more traditional methods, investigators also achieve such “supersaturation” of ECS with oxygen when they insert a gas dispersion tube into a bottle or holding container with ECS and bubble the ECS with either pure oxygen or almost pure oxygen. Therefore, we also measured the dissolved oxygen content of solution that was oxygenated with more standard methods. Specifically, we bubbled water in a 0.5 l glass bottle (Pyrex no. 1395) with an inserted standard dispersion device (Robu Glas Filter cat no. 1810, pore size 3, Hattert, Germany) at 138 mbar for 30 min. Subsequently, the solution was allowed to flow from this bottle through a set of tubing (Tygon R 3603, ID: 0.8, OD: 2.4 mm, cat. no. T3601-13; Saint-Gobain Performance Plastics, Aurora, OH) via gravitational force into a recording chamber, and dissolved oxygen of this solution was measured. We found that the dissolved oxygen content in this type of setting was 29.8 mg/l, suggesting that oxygen supersaturation levels were higher when solution was bubbled with a standard gas dispersion tube, compared with oxygenation in the chamber with the water pump technique.

Fig. 4.

The oxygenation mechanism within the chamber is capable of oxygenating the solution in the chamber to supersaturation values and is nearly as effective as standard “bubbling” of solution in an external container with gas dispersion tubes. Dissolved oxygen (DO) concentrations in solution samples that were collected from the chamber after having been in the chamber and exposed to oxygenation for various amounts of time. The initial t = 0 value represents the oxygen concentration of deoxygenated (boiled) water. Dashed line indicates dissolved oxygen in water at equilibrium with atmospheric oxygen.

Electrophysiological Recordings Suggest Stable Physiological Properties Over Prolonged Incubation Times

To confirm the suitability of our recording chamber for brain slice recordings, we performed a number of patch-clamp recordings from auditory brain stem neurons. Specifically, we performed current-clamp recordings from neurons in the MNTB in the mammalian auditory brain stem. This nucleus was chosen for two reasons. First, it has been extensively studied and its neural responses are well characterized (von Gersdorff and Borst 2002; Schneggenburger and Forsythe 2006; Borst and Soria van Hoeve 2012). Second, auditory neurons, especially in the lower auditory system, are known for their high level of metabolism (Clarke and Sokoloff 1994), which is presumably due to their chronic and high rates of action potential firing. Since high metabolic rates are directly related to high oxygen consumption rates, neurons in such metabolically active areas are suitable model systems to test the oxygen delivery capability of our novel recording chamber. We obtained whole cell recordings and then performed current-clamp recordings on 13 MNTB neurons in brain slices that have been maintained and oxygenated in our recording chamber for various amounts of time between 8 min and 3 h. Figure 5, A and B, shows sample traces of two current-clamp recordings from two MNTB neurons into which current of various amplitudes was injected. Both negative and positive current injections elicited typical responses in these neurons, including the opening of I_h channels (Koch et al. 2004), spiking (Taschenberger and von Gersdorff 2000), and a phasic spiking pattern of MNTB neurons (Forsythe and Barnes-Davies 1993), suggesting that all recordings were performed from healthy neurons. The traces in Fig. 5A were recorded from a neuron in a brain slice that was incubated and carbogenated in the novel chamber for 60 min but only a few min after break-in. The traces shown in Fig. 5B were from a neuron in a brain section that was incubated in the chamber for 180 min.

Fig. 5.

Incubating brain slices in the chamber for prolonged periods of time maintains tissue health, even when afferent fibers are chronically stimulated to trigger action potential firing. The figure shows current-clamp traces recorded from neurons in slices that were incubated in the chamber for various amounts of time (A and B) and had been electrically stimulated for prolonged periods of time (C–F). The traces show responses to current injections of −200 to +800 pA in 100-pA steps (A and B) or current injections of −300 to +800 pA in 100-pA steps (C–F) for a duration of 300 ms. The recordings were performed after tissue sections had been incubated in the chamber for 60 min (A) and 180 min (B) without electrical stimulation. The recordings shown in C and E were obtained directly after establishing a whole cell configuration. Subsequently, the recordings were held for 75 min (D) or 24 min (F) while afferent fibers were stimulated. In the case of E and F an intact direct synaptic connection was verified.

To increase the metabolic demands of the tissue section, we stimulated the afferent excitatory fiber bundle that innervates the MNTB with ongoing electrical stimulation. We placed a bipolar stimulation electrode into this fiber bundle near the midline (Taschenberger and von Gersdorff 2000; Hermann et al. 2007) and stimulated the fiber bundle electrically with 20 V/20 Hz Poisson distributed stimulus trains for several hours. In connected cells, the afferent stimulation elicits synaptic release in the calyx of Held, a type of giant synapse known for a very high level of transmitter release, which leads to action potentials in the postsynaptic neuron in the majority of events. We recorded from 11 MNTB neurons from slices that were stimulated with this method and confirmed an intact direct synaptic connection in four of these neurons. Figure 5, C and D, shows current-clamp recordings from a MNTB neuron, which was patch clamped at t = 0 (Fig. 5C), followed by 75 min of electrical stimulation at 20 Hz, which were occasionally interrupted for a set of current-clamp recordings. The recording obtained after 75 min of stimulation, just before the cell was lost, is shown in Fig. 5D. While the recording obtained at t = 0 shows a higher number of action potentials in response to the current injection than the recording at 75 min, many other physiological parameters such as threshold, resting potential, or spike amplitude were very similar between the two recordings, suggesting that the neuron was still healthy after 75 min of whole cell patch-clamp and electrical stimulation. Figure 5, E and F, shows recordings from a MNTB neuron that received a confirmed intact synaptic input through the calyx of Held. Similar as in Fig. 5, C and D, this recording was held for a prolonged time period while the axonal input of this calyx (and other calyces) was stimulated with ongoing 20-Hz Poisson stimulation. Similar as it was the case for the other recordings, prolonged incubation and stimulation did not alter the neuron's firing significantly.

Imaging of Brain Slices Suggests Little Change of Overall Slice Health During Prolonged Incubation and Stimulation

An alternative method to assess the health of neurons in a brain slice is to image the section under Nomarski or Dodt optics (Allen et al. 1969; Dodt et al. 2002) to assess their size, shape, and intactness of their cell membranes. Figure 6 shows three images from MNTB brain sections that were imaged with Dodt optics at t = 0 (Fig. 6, A and C), followed by 1 h of 20-Hz electrical stimulation as described above, followed by the acquisition of a second image (Fig. 6, B, D, and E). In some cases, electrical stimulation was performed for longer time periods (Fig. 6F, 120 min). In some of these experiments, single neurons in these sections were patch clamped to collect the physiological data (glass electrode shown in the Fig. 6, E and F). A comparison of the images at the beginning and the end of the stimulation period reveals little difference, suggesting that the overall slice health did not deteriorate significantly during the incubation/stimulation period.

Many researchers perform patch-clamp recordings at or near physiological temperature to eliminate the risk of temperature related artifacts in their data. While our chamber does not include an active heating mechanism to maintain ECS at physiological temperature, it should be compatible with existing heating mechanisms used already by researchers. To test for the chamber's capabilities to maintain slice health at physiological temperature, we installed the chamber in an environmental chamber mounted on a 3i Marianas inverted spinning disk microscope and imaged brain sections over time with this machine. The environmental chamber was adjusted to 37°C and the humidity to 100%, and brain sections were imaged repeatedly over a period of 2 h. Figure 7 shows images from such a section at t = 0 (Fig. 7A) and t = 120 min (Fig. 7B). A comparison of these images again suggests that there was no significant deterioration of overall slice quality during the test period.

By contrast, when aeration with the gas mixture was off, tissue health deteriorated rapidly (Fig. 7, C and D). Figure 7C shows an image of a tissue section at t = 0. This section was incubated in the same way as the section shown in Fig. 7, A and B, except that the aeration was off. After 60 min the tissue health had degraded to the point where it was difficult to even see neurons in the section (arrows), suggesting that the aeration provided through the water pump mechanism plays a significant role in maintaining tissue health.

Liquid Chromatography and Mass Spectrometry of ECS Samples Suggests Negligible Depletion of Nutrients or Accumulation of Waste Products

Most existing perfusion systems use significantly larger volumes of ECS than the chamber described here, exposing the brain section to an overall larger amount of nutrients and diluting possible cellular waste products to a larger degree. To test whether the extremely small ECS volumes used here result in changes of the ECS caused by the metabolic activity of the brain section, we performed liquid chromatography and mass spectrometry analyses on ECS samples that were incubated in the chamber with a brain section for various amounts of time. Specifically, we compared the composition of fresh ECS, solution that was used to incubate a brain section in the chamber for 10 min, and solution that was used to incubate a section for 3 h. We observed no significant changes in glucose, myo-inositol, and pyruvic acid over the 3-h period (Fig. 8, A and C). In contrast, to the ascorbic acid that was already not measurable in the ECS after 10 min of incubation and remained absent for the 3-h incubation time. However, ascorbic acid has been reported to be unstable in oxygenated solutions and oxidize initially to dehydroascorbic acid, which then oxidizes further to 2,3-diketogluconic acid with a halflife of ∼6 min (May 2012). Thus the most likely reason why ascorbic acid decreased was that it had undergone a series of oxidation steps and degraded, rather than being metabolized by the brain tissue.

On the other hand, prolonged incubation of brain slices with low volumes of ECS did not yield any measurable waste products. One advantage of using liquid chromatography and mass spectrometry is that this analysis does not require any assumptions by the investigator, i.e., it is able to detect compounds that an investigator did not explicitly look for. Our samples did not contain any identifiable metabolic waste products, even after the 3-h incubation period, suggesting that even prolonged incubation periods do not compromise the solution.

DISCUSSION

We present a new design for a self-contained recording chamber for very small volumes of ECS (1.0–2.6 ml total) that has applications in brain slice electrophysiological experiments. The key difference between this chamber and more traditional systems is air-propelled circulation of the liquid within the chamber itself, thus eliminating the need for external fluid storage containers and connecting tubing. The chamber design includes the following: a gas inlet port, a fluid circulation mechanism, and the slice placement compartment. These compartments are continuous yet separated from each other by fluid dynamics such that rapid influx of pressurized gas does not disturb the tissue section under investigation.

We demonstrated that active oxygenation of the solution is provided through a gas tube that also serves in propelling liquid circulation. We found that the oxygenation mechanism of our chamber can oxygenate an aqueous solution to the saturation point of dissolved oxygen in atmospheric conditions and can even “supersaturate” the solution. Compared with the more commonly used bubbling, i.e., oxygenation of solution in an external container with gas dispersion tubes, we found that the more conventional methods can “supersaturate” the solution even further than the internal mechanism of our chamber but it is unclear how this difference is relevant for tissue health (Colt 1983; Erecińska and Silver 2001; Turner et al. 2007). Healthy looking slices and the ability to perform patch-clamp recordings from brain slices that have been maintained and in some cases electrically stimulated in this chamber for several hours suggest that the oxygen concentration in the chamber may be sufficient to keep neurons alive and healthy as measured in their ability to fire action potentials.

Other studies have used a variety of methods to assess tissue health such as measuring hippocampal oscillations (which only occur in healthy sections; Hajos and Mody 2009; Hajos et al. 2009) or measuring the availability of oxygen to cells and specifically mitochondria (e.g., Turner et al 2007). The main goal of determining dissolved oxygen contents and tissue health in our chamber was to ensure that tissue health in our chamber is sufficient to perform physiological recordings while reducing the fluid volume as much as possible. We think we have achieved this goal, since our data suggest that health of sections incubated in our chamber is comparable or only slightly less good than in more traditional superfusion systems. If there is a difference in tissue health, it appears to be not significant at least not for experiments lasting up to 3 h (the longest period that was tested).

Another concern with using extremely small volumes of ECS is the potential depletion of nutrients such as glucose and/or the potential accumulation of harmful waste products. Specifically, nutrients contained in our ECS are as follows: glucose to maintain glycolysis in the absence of a circulatory system; myo-inositol to support homeostasis and slice metabolism (Fisher et al 2002); or pyruvate and ascorbic acid as a neuroprotectants (Gramsbergen et al. 2000; Fink 2008; Halliwell 1992). Our mass spectrometry analysis suggests that these nutrients do not deplete and do not convert into potentially harmful metabolites. The exception is ascorbic acid, which oxidizes within minutes of exposure to oxygen and becomes ineffective (May 2012).

We suggest that this chamber may be a useful tool for brain slice recordings that involve the addition of expensive chemicals or pharmacological agents to the bath solution. For example, one motivation to develop this chamber in our laboratory was a series of experiments involving caged-glutamate compounds. These compounds have to be added to the bath at distinct concentrations and are relatively costly. The cost per experiment or per day of experimentation depends on the total volume of ECS enriched with caged glutamate that is being used during such an experiment; the smaller the total required volume, the more cost effective the experiments are.

Other than saving costs by using smaller amounts of expensive chemicals per experiment, this chamber also keeps the fluid levels of ECS in the bath constant. Most perfusion systems that rely on external tubes to add and remove ECS continuously use suction to remove ECS from the chamber. A typical challenge with using such a suction system is to keep the bath fluid level constant, avoid vibrations due to changing fluid levels, and avoid electrical noise often associated with suction. By recirculating a constant amount of fluid within the chamber itself, all these potential problems do not apply.

One challenge of working with this chamber is related to general problems of working with very small volumes of liquid and liquid evaporation. Since the liquid volume in the recording chamber is already very small, any evaporation will be noticed. We also found that the gas mix itself (the standard 5% CO₂-95% O₂ mix) contributes to the evaporation, as the gas in the pressurized tanks is dehumidified when the tanks are filled by the manufacturer to prevent internal corrosion. We found that the evaporation effects resulting from dry pressurized gas can be prevented by rehumidifying the gas from the tank before allowing it to enter the recording chamber by passing the dry air through water using a standard bubbling rod or stone. By contrast, we addressed the evaporation into a laboratory room with dry air by placing a household humidifier into the vicinity of the chamber and directing the mist stream at the chamber.

While the chamber in its current form does not allow for the independent heating of the ECS, it is possible to use it with some existing heating systems to obtain recordings at physiological temperature. However, many common heating systems such as inline heaters or external water bath systems are incompatible with this chamber since it does not depend on a constant fluid influx from external pipes.

It would also be challenging to exchange solutions in the chamber during an ongoing patch-clamp recording, as is sometimes done to wash in pharmacological agents. However, these types of experiments inherently need to use larger fluid volumes such that an experimenter would probably choose an open flow system over this self-contained chamber.

GRANTS

This work was supported by National Institute on Deafness and Other Communication Disorders Grant RO1-DC-011582 (to A. Klug). Some imaging experiments were performed in the University of Colorado Anschutz Medical Campus Advanced Light Microscopy Core supported in part by National Center for Research Resources Colorado Clinical and Translational Sciences Institute Grant UL1-RR-025780.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: A.D., N.R., T.C.L., and A.K. conception and design of research; A.D., K.D.Q., C.I.C.-Q., and T.C.L. performed experiments; A.D., K.D.Q., C.I.C.-Q., N.R., T.C.L., and A.K. analyzed data; A.D., T.C.L., and A.K. interpreted results of experiments; A.D., K.D.Q., T.C.L., and A.K. prepared figures; A.D. and A.K. drafted manuscript; A.D., K.D.Q., N.R., T.C.L., and A.K. edited and revised manuscript; A.D., K.D.Q., C.I.C.-Q., N.R., T.C.L., and A.K. approved final version of manuscript.