|Home | About | Journals | Submit | Contact Us | Français|
Multi-well robotic planar patch-clamp has become common in drug development and safety programmes because it enables efficient and systematic testing of compounds against ion channels during voltage-clamp. It has not, however, been adopted significantly in other important areas of ion channel research, where conventional patch-clamp remains the favoured method. Here we show the wider potential of the multi-well approach with the capability for efficient intracellular solution exchange, describing protocols and success rates for recording from a range of native and primary mammalian cells derived from blood vessels, arthritic joints, and the immune and central nervous systems. The protocol involves preparing a suspension of single cells to be dispensed robotically into 4-8 microfluidic chambers each containing a glass chip with a small aperture. Under automated control, giga-seals and whole-cell access are achieved followed by pre-programmed routines of voltage paradigms and fast extracellular or intracellular solution exchange. Recording from 48 chambers usually takes 1-6 hr depending on the experimental design and yields 16-33 cell recordings.
In 1991 Bert Sakmann and Erwin Neher were awarded the Nobel Prize in Physiology or Medicine for their discoveries on the function of ion channels, which were made possible by their development of patch-clamp technology1. The technology revolutionized electrophysiological studies of cells and ion channels. It enabled resolution of unitary current events through single ion channel proteins as well as recording of macroscopic ionic currents across the membranes of a plethora of previously unexplored cell types from the animal and plant kingdoms, with options for simultaneous membrane potential control (voltage-clamp), fast kinetic resolution, and intracellular dialysis. The field of electrophysiology was transformed from one that was focused mostly on voltage-clamp recordings only from a restricted selection of large cells (i.e. that are amenable to two-electrode impalement) or to observations of membrane potential fluctuations in the absence of voltage-clamp (i.e. by cell impalement with a single sharp electrode). The vast expansion in the volume of high quality data on the properties of the many hundreds of ion channels controlling ion movement and electrical signaling in cells is testament to the seminal nature of Sakmann’s and Neher’s discoveries and the developments made by many scientists who worked with them or followed their example2.
However, along with recognition of the power of patch-clamp technology came recognition of the slow data output and need for highly specialized operators, aspects that have blocked or hindered implementation for a substantial body of investigators. Other techniques can be used to measure ion channel activity and may deliver higher data throughput and require less specialized operators. Rubidium is, for example, used to measure K+ channel activity because many types of K+ channel are permeable to this ion3. Fluorescent dyes are used to detect changes in membrane potential caused by changes in ion channel activity4-5. Nevertheless, these methods lack many important advantages of patch-clamp, including capabilities for voltage-clamp, intracellular dialysis, high kinetic resolution and quantification of ionic selectivity1. An alternative is enabled by incorporation of ion channels into artificial lipid bilayers6, permitting voltage-clamp but not providing information on ion channels within the cellular context, which is so often essential for how the ion channels behave. Driven by appreciation of patch-clamp as a gold standard for ion channel studies and the desire for greater data output in therapeutic drug discovery and safety programmes, major efforts were made to produce robotic patch-clamp systems7-9.
A particularly successful robotic approach is based on planar patch-clamp chips and this is the basis for the technology described here. With this method, the glass pipette of conventional patch-clamp recording is replaced by a thin sheet of flat glass with a small aperture in the middle (the planar chip). Conventional patch-clamp involves the maneuvering, under a microscope, of the fine tip (1-2 μm) of the patch pipette onto a cell, which commonly has a diameter in the region of 2-50 μm. In contrast, the planar method involves the robotic delivery of a suspension of cells to the chip, with one cell attaching to the aperture randomly by the application of negative pressure underneath the chip; the intracellular solution is underneath the chip, and there is no microscope or micromanipulator (e.g. Fig. 1a). Various forms of the method have been successful. In some cases, numerous apertures are made in the chip and 384-well throughput is achieved (www.moleculardevices.com), but without many of the important advantages of conventional patch-clamp. In other cases, systems have been developed for 2, 4, 8 or 16-well recordings whilst retaining advantages of the conventional system (Fig. 1a-d and see below). During this development period, interest in the multi-well systems has resided mainly in the pharmaceutical sector, with focus on cell lines induced to over-express ion channels9. While this is important, it does not affect or relate to significant elements of the large ion channel community in the basic or academic research sectors where studies of endogenous ion channels and electrical phenomena in native cells remain common and important.
We have found that robotic multi-well planar patch-clamp technology can generate high quality publishable data10-12 (and illustrative data shown in this article) from native cells and cells in primary culture, as well as confer advantages and be enabling compared with conventional patch-clamp. Systematic data production has been possible for a reasonably large number of cells on one day of recording. Therefore, comparison of two experimental conditions has been more reliable than previously, reducing problems due to day-to-day variation in cells and giving greater statistical power. In some cases, we have seen significant increase in performance and studies have been enabled that would previously have been abandoned. The micro-fluidic system uses small volumes, reducing wastage of agents that may be expensive or in limited supply. Although there is possibility for operator intervention during recordings, the reduced tendency for intervention increases the likelihood of generating reliable, objective, outputs. Routine intracellular perfusion opens possibilities for studying intracellular regulation and pharmacology of channels, studies that were previously available only to highly skilled and dedicated experimenters. Use with transiently transfected cells is possible. One potential concern is the cost of planar chips in the context of a typical research grant and during development of methods for a previously unexplored cell type. Another is that there is reduced sealing success rate in the absence of seal enhancer solutions. For a detailed comparison of multi-well planar and conventional patch-clamp methods the reader is referred to Table 1. We expect that the remaining challenges will largely be overcome or accepted. Users may decide that the lesser flexibility compared with conventional patch-clamp (e.g. absence of the outside-out patch configuration) is outweighed by the advantages, especially when the primary need is whole-cell recording. Lastly, multi-well planar technology has the potential to widen access to patch-clamp and generate a platform for achieving international agreement on the properties of ion channels.
A wide range of single cells can be used for multi-well patch-clamp recording (for examples, see Reagents list). There are few specific limitations to the choice of cell, though we recommend that there is prior knowledge of the suitability of the cells for conventional whole-cell patch-clamp recording before embarking on multi-well planar patch-clamp studies. Cells that are contractile (e.g. cardiac muscle cells), very large (e.g. Xenopus laevis oocytes) or that have long processes (e.g. hippocampal neurones) present special challenges and we would advise first obtaining experience with other cell types (e.g. as shown by our illustrative examples). Single cell preparation techniques are similar to those generally used for many different studies of freshly-isolated cells or cells maintained in culture media. Homogeneity in the cell population would often be a feature the user would want to achieve to an acceptable level, although we suggest a method that might enable studies of heterogeneous cell populations (see below). Clumps of cells and pieces or slices of tissue are unlikely to be suitable, but we have not specifically investigated such preparations. The technique is not suitable for whole organ or in vivo studies.
Whole-cell patch-clamp recording has good signal-to-noise ratio and can resolve current amplitudes of a few pico (10-12) Amperes (pA). It is, however, easier to obtain reliable data when current amplitudes are larger (e.g. 0.05-5×10-9 A, nA). Problems with the quality of voltage-clamp may arise with large currents and these need to be addressed by adoption of suitable compensatory circuitry, as in conventional patch-clamp recordings13. If the ionic current of interest cannot be resolved by conventional whole-cell patch-clamp recording, it is unlikely to be detected by multi-well planar patch-clamp. We have investigated multi-well planar patch-clamp for studies of whole-cell (macroscopic) currents but the reader should be aware that single channel events are detectable in planar cell-attached patch-clamp recordings8 and it might be possible to generate a configuration similar to the inside-out patch if the whole-cell membrane is permeabilised.
We recommend that users follow one of our suggested protocols at least for the first time. On the basis of data obtained, users can optimise experimental conditions. There are various aspects to optimise, including: cell preparation; aperture of the hole in the chip; ionic solutions during seal formation and sampling data; suction parameters to form seals and break-through to the whole-cell configuration; and voltage paradigms.
We recommend that users follow published voltage protocols and ionic solutions for their ion channel of interest because these can have critical effects on the determination of ion channel properties and especially on quantitative information about such properties.
The duration of each recording depends substantially on the ion channel and cell type under investigation. In whole-cell configuration users might normally expect recordings to be up to 1 hour long.
The following controls should be included when studying effects of compounds or other substances on channel activity: (i) Solution exchange with the solvent used to dissolve the test substance; and (ii) Solution exchange using the same solution to explore if fluid movement has any effect on channel function. When the experiment involves cell transfection, controls might include cells transfected with DNA vector or scrambled siRNA.
Commercial systems available for multi-well planar patch-clamp include PatchXpress 7000A (www.moleculardevices.com), QPatch (www.sophion.dk) and Patchliner (www.nanion.de). The protocol described here is for the Patchliner, which is provided with a Tecan robot (www.tecan.com) and HEKA (www.heka.com) patch-clamp amplifier and software. In principle, multi-well systems other than the Patchliner may be suitable for the types of studies we describe in this article but we have not investigated this matter directly and are not aware of published evidence to confirm use in this context. There are differences in the design, manufacturing and storage of the planar chips that may lead to differences in performance with different cell types.
Data analysis routines are similar to those developed for conventional patch clamp, with various software options and levels of automation available. Data can be converted for use with common analysis and presentation packages, including Excel (office.microsoft.com) and Origin (www.originlab.com).
choose the medium relevant to the cell line(s) in use.
Prepare the appropriate intracellular solutions (50 ml) in deionised water, filter (0.22 μm) and measure the osmolarity and the pH. The solutions can be aliquoted (4 ml aliquots) and stored at -20 °C.
▲ CRITICAL STEP: To reduce the likelihood of degradation, ATP should be added to the internal solution only on the day of the experiment and then the osmolarity and pH re-checked.
▲ CRITICAL STEP All recording solutions should be filtered. The internal solution is particularly important and should be filtered with a 0.22 μm pore diameter filter. The osmolarity of recording solutions is measured directly using a freezing-point osmometer and adjusted if necessary with mannitol or sucrose, non-permeating molecules, to 290 mOsM for all internal solutions and between 290 and 310 mOsM for all external solutions, unless otherwise stated. External osmolarity should always be higher than the internal.
Prepare extracellular solutions (500 ml) in deionised water, filter and measure the osmolarity and the pH. The solutions can be stored at 4 °C for up to 5 days. Solutions should be warmed to room temperature (20±2 °C) before use.
Prepare solutions in deionised water, filter and measure the osmolarity and the pH. The solutions can be stored at 4 °C for up to 5 days. Solutions should be warmed to room temperature (20±2 °C) before use.
Pre-programmed protocols (Trees, e.g. Fig 1e) and pulse generator files are available. The Editor window allows the user to load, create and modify Trees for sealing a cell, achieving and maintaining the patch clamp configuration and performing patch clamp experiments.
The electrodes are manufactured from Ag/AgCl coated steel and need to be regularly chloridated in bleach, controlled via a pre-programmed Tree that takes 45 min to complete. Electrodes need to be re-chloridated about once a week and are generally replaced every 2 months.
At the start of all experiments, planar chips are filled with intracellular seal enhancer solution and the extracellular solution that will be used later during data collection. Cells are dispensed into the extracellular chamber and giga-seals formed, usually with the aid of the extracellular seal enhancer solution. Typically, following seal formation and before establishment of the whole-cell, solutions are exchanged for those appropriate to the cell type under investigation (Fig 1e).
The Patchmaster software has integral online analysis capabilities allowing user-defined levels of automation.
1| Load a pre-programmed PatchControl©HT Tree. A simplified Tree is shown for illustrative purposes (Fig. 1e).
2| Select Edit Mode to access COMMAND folders that can be edited and inserted in the Tree using the generic drag-and-drop function.
3| To program the robot to collect solutions at the desired volume, dispense them at the selected speed and activate the appropriate pulse generator file in Patchmaster, the ‘define experiment sequence’ of the Tree is selected and edited (Fig. 1e, Step 1). Using the dialogue box, compound positions are set (e.g. position C2 seen in Fig 1d). Possible speeds of solution application range from 1 to 857 μl/s, with speeds of 4 μl/s (intracellular solution) and 10 μl/s (extracellular solution) being typical. For ligand-gated ion channels, ≥30 μl/s is advised for extracellular solution application.
4 | Highlight the ‘set compound positions’ branch of the Tree (Fig. 1e) to select the rack settings for solutions/compounds, specify the compound name and stipulate the concentration. Define these in the drop down menu using the dialogue option. Once the data acquisition is activated (see step 9) an Excel file is generated which logs the compound used, its concentration and the time of application, and allows for subsequent automated analysis (e.g. IC50 determination).
5| Select the ‘add cells’ folder and set the pressure (for attracting the cell) to between -50 and -150 mB depending on the characteristics of the cell membrane. Adjust the pressure to be maintained to -50 mB once a cell seals on to the aperture of the chip chamber. Pressure is applied from below the chip to bring a cell onto the aperture.
6| Within the ‘check for cell contact’ folder, adjust the pressure settings in the drop down menu to improve the seal or encourage a cell onto the aperture. Set the pressure to a desired value for a given time (e.g. -100mB for durations such as 10 s) and then program it to return to atmospheric pressure for the same duration. Adjust the settings to repeat this process multiple times with increasing pressure (e.g. -10 mB incremental increase) up to a maximum -200 mB or until the desired seal resistance is obtained. Pressure settings in this folder will need to be optimised by the operator to suit the cell type depending on the type and/or characteristics of the cell membrane.
7| To improve the seal further, highlight the ‘monitor seal’ branch command within the ‘add seal enhancer solution’ folder and in the drop down menu adjust the pressure and voltage settings and also the condition parameters to give an increase-release pressure cycle. These settings will need to be optimised for different cell types/characteristics. Pressure pulses are typically applied every 5 s, with incremental increases in pressure. Once the desired seal is attained, the pressure is typically maintained at -50 mB. The condition for a high-quality seal is typically set to ≥1 GΩ and, if 1 GΩ is not achieved, recordings are continued if the resistance is >0.6 GΩ, but otherwise timed out and the chamber disabled. The operator defines the thresholds for proceeding with data sampling.
▲ CRITICAL STEP It is helpful to use seal enhancer solutions to increase the chance of giga-seal formation. The extracellular seal enhancer solution is a high Ca2+-containing solution, whereas the intracellular solution contains fluoride (see REAGENT SETUP). Seal enhancer solutions are usually replaced once a giga-seal is achieved and before gaining whole cell access (Fig 1e Tree ‘exchange solutions’).
8| Select the ‘check for whole cell’ folder (Fig 1e) and set the pressure in the drop down menu to apply suction pulses, typically from -50 mB up to -350 mB in -50 mB increments. These pressure changes are applied individually to the different cells and the settings depend on the seal resistance, membrane capacity and series resistance. In addition, adjust the size of the high voltage pulses to 600-800 mV (“zap”), which can be applied to help rupture the patch of membrane, thus establishing the whole-cell configuration. Chip chambers where cells have not sealed or achieved the whole cell configuration are disabled at this stage. The operator can adjust pressure, voltage and condition settings according to the cell type/characteristics.
9| Insert the appropriate commands into the ‘whole-cell experiment’ folder in the correct order to correspond to the experiment sequence defined at the beginning of the Tree. These commands will activate Patchmaster data acquisition software and direct the robot to apply compounds according to the experimental protocol. Also adjust the time in the ‘time step’ branch of this folder to control the duration of data acquisition and the exposure time of the cells to different compounds (Fig. 1e).
▲ CRITICAL STEP The ‘maintenance’ command (Fig. 1e) is routinely inserted in the Tree, immediately following any extracellular solution application command, to minimise risk of solutions leaking into the electrical hardware and to wash the robot pipette arm. These features also mean that there are no external solution volume constraints due to limited waste chamber size and no mixing of compounds due to contamination by the pipette. The maintenance commands are especially important when applying several compounds consecutively (e.g. when constructing dose-response curves).
10| Highlight the ‘End’ folder and adjust the settings in the drop down menu to move the chip wagon to the next chamber and then, if desired, to the next chip. Furthermore, set the program here to re-activate any chip chambers disabled in the previous run. The chip cartridges contain 16 chambers each with a planar glass chip embedded in it. These chambers are arranged in two rows of eight (Fig. 1b, c). For Patchliner Quattro, the first chip site consists of the first four chambers; two in row one and two in row two (i.e. there are four sets of four chambers in one cartridge). For Patchliner Octo there are 2 chip sites (i.e. there are two sets of eight chambers in one cartridge).
11| Save Trees in a user-defined folder. Once a Tree is optimised for a particular cell/channel type it is suitable for further use with the same cell/channel type without additional modification.
■ PAUSE POINT The user can pause at this point before proceeding with cell preparation and recording.
12| To prepare cell suspensions of T lymphoblasts, neutrophils and Jurkat cells (i.e. non-adherent cells), follow option A. For adherent cell types, follow option B.
▲ CRITICAL STEP Cell confluency of 60-80 % is critical. If cells are allowed to grow to a higher confluency, widespread contact among cells occurs and cells cluster and form unhealthy cytoplasmic granules.
▲ CRITICAL STEP Use only healthy cells. The health of the cell impacts on the quality of the seal formed between the cell membrane and the planar chip, which ultimately influences the quality of the recording. Once cells are prepared in suspension there is no opportunity to select the healthiest cells unless fluorescent cell-selection is incorporated (Anticipated Results).
13| View cells under a light microscope to ensure they are single, non-clustered, cells with smooth membrane edges (e.g. Fig. 2).
14| Select the appropriate NPC©-16 chips based on our experience or user optimisation. Typically, our chip (aperture) resistances are 1-2 MΩ (smooth muscle cells), 2-3 MΩ (astrocytes), 3-5 MΩ (lymphoblasts, Jurkat cells, synoviocyte, SH-SY5Y cells and HEK 293 cells) or 5-8 MΩ (neutrophils). NPC©-16 chip resistance can be manufactured within a range from 1-8 MΩ with an accuracy of ±0.5 MΩ for resistances between 1-4 MΩ and ±1 MΩ for resistances between 5-8 MΩ, making them suitable for a variety of cell types of different sizes. Place three chips on the chip wagon (Fig 1d).
15| Place the suspended cells from step 12 in the cell hotel (Fig 1d) and activate the hotel using the robot drop down menu option in the software. Set the cell hotel pipette to aspirate the cells every 30 s not only to avoid clumping and sedimentation, but to increase the length of time that the cells remain viable once prepared in suspension. Based on trypan blue exclusion and ionic current recording we find that cells retain good viability for at least 3 h after they have been prepared in suspension, although some deterioration in success rates has been observed after 2 h.
16| Place the solutions in their appropriate positions according to the ‘define experiment sequence’ settings of the chosen Tree.
17| Select the ‘initialise robot and electronics’ folder (Fig 1e) and activate initialisation by clicking the Start arrow at the bottom of the Editor Window. This folder only needs to be activated once at the beginning of each day of experiments. It generates a new data file within the HEKA software and sets all amplifier and robot parameters to reasonable starting values. The robot is initialised and the pipette is washed.
18| Activate a newly modified Tree by highlighting the ‘start’ folder (Fig 1e) and clicking the start arrow at the bottom of the Editor Window. This will activate the robot and the Tree, loop back to the start at the end of each run and continue this process until all NPC©-16 chips on the chip wagon have been used. At this point the robot will stop.
▲ CRITICAL STEP Whole-cell patch-clamp experiments are performed at room temperature (20±2 °C).
19| Close the PatchControl©HT software by selecting ‘quit’ from the menu in the Editor window. Select the ‘save and exit’ option when closing down Patchmaster to ensure that data files are saved properly.
■ PAUSE POINT Data can be analyzed anytime after recording.
20| For data analysis, use the standard export and import functions of the HEKA software (www.heka.com). The data analysis is as time-consuming as for conventional experiments, although for standard IC50 fits an automated export function (e.g. IGOR-Pro software) can be activated during data acquisition.
21| To display the data for immediate presentation, set up the online analysis using instructions in the Patchmaster manual and open the online analysis. Highlight a sweep or a series of sweeps in the replay window. Raw data traces will be displayed in the oscilloscope window. Select the overlay sweep and overlay series functions in the oscilloscope window and activate the scan option. The scan function generates a digital display of amplitude, time or voltage measurements from the raw data trace. At the same time, time-course or current-voltage plots can be displayed in the online analysis window. Create a screen shot of windows of interest to provide an easy way of presenting data visually without first undertaking lengthy analysis.
Steps 1-11, typical modification of pre-programmed experimental routine: ~30 min
Step 12-13, preparation of cells: ~ 20-30 min
Step 14-16, commencing the experiment: ~10 min
Step 17-19, patching: ~ 1-6 h
Step 20-21, data analysis: ~ 1-10 h
Experiments have been performed using a planar system with capacity for up to 8 simultaneous recordings and incorporating micro-fluidics for rapid, unlimited, extracellular and intracellular solution exchange (Fig. 1). Experiments were performed alongside conventional patch-clamp experiments that have been routine in our labs for up to 20 years. Validation of the system was achieved with cell lines over-expressing ligand-gated, receptor-operated and voltage-gated ion channels (Fig. 3a-d). Although the cells need to be released into suspension to make the recordings, the data are similar to those obtained by conventional patch-clamp when cells are often attached to a glass coverslip (Fig 3d).
While validating the planar system we were engaged in a separate project using conventional patch-clamp to study ion channels of rabbit and human fibroblast-like synoviocytes, which are relevant to arthritis. The success rate in making these conventional recordings from the cells was poor (e.g. Table 3). The cells attach to coverslips, spread out as thin structures and show irregular shapes (Fig. 2a). Despite preconception that such cells would not perform well with planar technology, the success rates were high, both in terms of giga-seal formation and whole-cell recording (Table 3). Whole-cell series resistance values were remarkably consistent, suitably low compared with membrane resistance, and independent of cell size (Fig 4a, Table 3). Similar success occurred with primary-cultured proliferating human vascular smooth muscle cells (Fig 4b), which are also troublesome in conventional studies (Table 3). An illustrative recording shows reversible block of constitutive non-selective cationic current by gadolinium ions, and subsequent block by 2-aminoethoxydiphenylborate (2-APB) (Fig 4c). Recordings were long in duration and stable during solution exchange (Fig 4c). We were encouraged to explore other cell types. Similar high performance occurred with rat astrocytes and, albeit at lower success rates, human lymphoblasts or neutrophils (Table 3). Current-voltage relationships (I-Vs) were similar to those observed during conventional recordings (Fig 4d). Original traces show the signal-to-noise ratio and characteristics expected of whole-cell current recordings (Fig 4e).
Whole-cell access could be attained by physical rupture of the cell-attached patch or by permeabilisation with amphotericin B (Table 3, Fig 4e). Amphotericin B limits intracellular dialysis to small ions, making recordings more physiological and reducing run-down of channel activity. Physical rupture allows intracellular application of exogenous substances. With the robotic system, not only could substances be applied intracellularly, but complete intracellular solution exchange was easy and efficient (Fig 4f). Multiple successive exchanges could be achieved (e.g. Fig. 3c).
The planar method of patch-clamp has special features that need to be handled by the investigator. For example, some membrane processes may be adversely affected by the protocol required to produce the cell suspension, or by seal enhancer solutions. Therefore, independent assessment of key ion channel properties should initially be obtained for comparative purposes, either from the published literature or new conventional patch-clamp recordings. Secondly, if transient transfection is important for the experiments (e.g. for RNAi studies), high efficiency of the transfection may be necessary because of the random process by which a cell reaches the aperture in the chip. Several electroporation or lipid transfection methods provide such high efficiency. Contamination of transfected cell populations with non-transfected cells will tend to increase the variance within the test data group, increasing the need for greater numbers of cell recordings prior to statistical analysis; that is, transfection efficiency less than 100 % may be acceptable if the recording success rate is high enough. Sorting of transfected cells before recording (e.g. by FACS) is an option for increasing the proportion of transfected cells, but may carry with it disadvantages such as reduced total cell count and expense. Users should be aware that a fluorescent cell (e.g. expressing green fluorescent protein) can be detected on the aperture of the chip by viewing the underside of the cartridge on a fluorescence microscope after ionic current recording. By this means, ionic current data can be categorized into test (transfected cells) and control (untransfected cells) groups, enabling all ionic current data to be used and avoiding wastage of chips. Such an approach also has potential to eliminate data from unhealthy cells if a fluorescent live-dead cell indicator is incorporated into the experiment, or to categorize data from heterogenous cell populations.
The work was primarily supported by grants from the Wellcome Trust. PS was supported by an Overseas Research Student Award (UK) and YB by the Egyptian Ministry of Higher Education. We thank Nanion Technologies for good technical support.
COMPETING INTERESTS STATEMENTS
The authors declare no competing financial interests.