Loose patch clamp membrane current measurements in cornus ammonis 1 neurons in murine hippocampal slices

Hippocampal pyramidal neuronal activity has been previously studied using conventional patch clamp in isolated cells and brain slices. We here introduce the loose patch clamping study of voltage‐activated currents from in situ pyramidal neurons in murine cornus ammonis 1 hippocampal coronal slices. Depolarizing pulses of 15‐ms duration elicited early transient inward, followed by transient and prolonged outward currents in the readily identifiable junctional region between the stratum pyramidalis (SP) and oriens (SO) containing pyramidal cell somas and initial segments. These resembled pyramidal cell currents previously recorded using conventional patch clamp. Shortening the depolarizing pulses to >1–2 ms continued to evoke transient currents; hyperpolarizing pulses to varying voltages evoked decays whose time constants could be shortened to <1 ms, clarifying the speed of clamping in this experimental system. The inward and outward currents had distinct pharmacological characteristics and voltage‐dependent inactivation and recovery from inactivation. Comparative recordings from the SP, known to contain pyramidal cell somas, demonstrated similar current properties. Recordings from the SO and stratum radiatum demonstrated smaller inward and outward current magnitudes and reduced transient outward currents, consistent with previous conventional patch clamp results from their different interneuron types. The loose patch clamp method is thus useful for in situ studies of neurons in hippocampal brain slices.


INTRODUCTION
The hippocampus is central to the regulation of memory, learning, emotion, and hormonal activity.Of its functional layers, the cornus ammonis 1 (CA1) is key to memory formation. 1It participates in a range of acute [2][3][4] and chronic clinical neuropathology. 5,6CA1 comprises different layers, each with distinct cellular organizations.The external plexiform layer contains pyramidal cell axons and hippocampal afferent fiber projections from the entorhinal cortex.The stratum oriens (SO) contains different interneuron types, including basket cells and orienslacunosum molecolare (O-LM) cells, and basal dendrites of pyramidal cells, from which branch their axon initial segments (AISs).The stratum pyramidalis (SP) contains the pyramidal cell somas.The stratum radiatum (SR) and stratum lacunosum/molecolare (SL-M) both contain apical dendritic arborizations of pyramidal cells and different interneuron types, including basket and ivy cells. 7,8 here introduce loose patch clamp methods to investigate particular features of membrane current in CA1 cells in hippocampal slices for the first time.This complements conventional tight patch electrode techniques proven to be invaluable in both on-and whole-cell electrophysiological single cell current-clamp [9][10][11][12] or voltage-clamp studies of neuronal cells in hippocampal tissue slices 13 or following isolation. 14,15e latter techniques utilize electrodes with relatively small tip diameters to achieve required high seal resistances. 16,17This approach permits both a whole-cell configuration requiring access to the internal cellular environment, or a cell-attached configuration leaving the membrane intact.
The whole-cell configuration permits membrane current measurement over the entire somal cell membrane, which is useful in pharmacological 18,19 and dynamic and action potential clamp studies of physiological actions of tetrodotoxin-sensitive Na + channels in CA1 pyramidal neurons 20 and of persistent Na + currents. 21,22It has also been useful for studying Ca 2+ currents. 19,23,24However, where there is cell membrane rupture to electrically access the intracellular environment, dialysis of often Ca 2+ -chelating pipette solution into the intracellular environment limits studies of intracellular calcium dynamics. 17,25Both this and perforated patch variants also do not permit multiple experimental seal formations and detachments.
Additionally, total current includes contributions not only from the cell soma but the dendritic tree through its cable properties.In one report, it was not possible to control potential during the early transient sodium current, likely due to currents generated at a distance from the soma in the axon or dendrites. 21On the other hand, with the small pipette diameters, on-cell techniques leaving intact membrane confines readings to relatively small localized membrane areas. 25The latter nevertheless proved useful characterizing both unit and summed Ca 2+ and Na + currents in apical dendrites of rat CA1 pyramidal neurons. 23ose patch clamp methods, first introduced to study relatively large diameter skeletal muscle fibers, 26 offer complementary advantages and limitations.Application of pipettes with larger tip diameters on cells without enzyme pretreatment permits current recordings from larger but defined somal membrane areas without membrane rup-ture.8][29] This reversibility of the loose patch seal also allows pairwise comparisons from the same patch, 26 overcoming problems of intercell variability in experimental maneuvers involving application and withdrawal of pharmacological agents. 30It also permits studies in successive different cells using the same electrode with consistent geometric and electrical properties. 31,32However, the larger pipette areas mean that the pipettes consequently do not form a tight seal around the membrane.The low resulting seal resistances permit significant current to pass between the gap between membrane and pipette tip.This must be compensated for, with the resulting requirement for bespoke recording and current delivery electronics distinct from that used in conventional tight patch clamping that can deliver larger compensation currents.Nevertheless, loose patch methods have been used to successfully study minimally perturbed cells in situ within other intact neuronal tissues. 33 introduce the loose patch clamp technique to study murine hippocampal pyramidal neurons in coronal brain slices.The method permitted investigation of activation and inactivation properties of their inward and outward current responses.We then examined the temporal limitations of the loose patch clamp in this preparation.We thereby demonstrated inward and outward current contributions, as well as determining current pharmacological, inactivation, and recovery properties.We next compared these responses from cells in the SO/SP junction with those obtained from other, SP, and SO and SR hippocampal strata known to contain either the somas of pyramidal neurons (SP), or other cellular regions or interneuron types (SO and SR).This confirmed recordings from the easily identified SO/SP and from the SP as reflecting features of a single, likely pyramidal cell type, as distinct from findings in the SO and SR.

MATERIALS AND METHODS
All experimental procedures were approved by and conformed to guidelines of the ethical committee of the University of Surrey, Guildford, UK (NASPA-1819-25) All reagents were purchased from Sigma-Aldrich, unless otherwise stated.

Animals
Four-week-old C57BL/6 male mice (Charles River UK Ltd.) were maintained in the University of Surrey Biological Resource Facility under controlled conditions (ambient temperature 23 ± 2 • C, 12-h light/dark cycle) with food pellets and water supplied ad libitum.After a 1-week adaptation period to animal house conditions, animals were sacrificed by cervical dislocation (Schedule I, UK Animals [Scientific Procedures] Act 1986), and the brain was immediately dissected.

Tissue preparation
Following the decapitation of the animal, the skull was exposed by cutting and reflecting the surrounding skin.Two transverse incisions were made at the level of the occipital region, and the occipital region and interparietal bones were removed to expose the cerebellum.The calvaria was then opened by inserting the tip of a pair of fine scissors at the junction of the frontal and parietal bones.The remaining nerves and peduncles were removed, and the brain was harvested with a spatula and placed in ice-cold HEPES holding artificial cerebrospinal fluid (aCSF 34 containing 92 mM NaCl, 2.5 mM KCl, 30 mM NaHCO 3 , 1.25 mM NaH 2 PO 4 , 20 mM HEPES, 25 mM glucose, 10 mM MgCl 2 , 0.5 mM CaCl 2 ; pH adjusted to 7.4), which was constantly bubbled with a mixture of 95% O 2 /5% CO 2 .From these, 300-μm-thick coronal hippocampal slices were obtained using a micro-slicer (7000smz-2 Vibratome, Campden Instruments Ltd).Slices were incubated for 1 h at room temperature (20-25 • C) in HEPES holding aCSF, constantly bubbled with 95% O 2 /5% CO 2 .Tissue preparation and incubation of the slices were performed in HEPES holding aCSF.This limited neuronal damage resulting from edema 35 and cell swelling of the superficial neurons exposed to the mechanical stress of cutting procedures.
From each brain, four coronal slices were obtained from the regions of interest.From each slice, a single patch was obtained and studied in all the experiments, apart from those that compared currents from multiple hippocampal regions within the same slice.In the latter case, one patch was obtained for recording from each investigated region.Slices were maintained in HEPES holding aCSF, constantly bubbled with a mixture of 95% O 2 /5% CO 2 , for up to 5 h.

Loose patch clamp recording
The loose patch sharply contrasts from conventional patch clamp recording circuitry, requiring correction of larger seal leakage currents through <1 MΩ, as opposed to many GΩ in conventional patch, and consequent larger series resistance errors through the pipette length.Our adopted strategy 26 was adapted and validated in previous loose patch clamped amphibian, 36 mammalian skeletal 37,38 and cardiac atrial 39 and ventricular muscle. 40In Figure 1, the left arm of the custom-built circuit contains (a) the pipette tip and the seal that it makes with the patched membrane.It is the junction between the pipette resistance (R pip ) and the seal resistance (R seal ) that is to be "clamped."These connect (b) the back end of the pipette itself to both (c) a 10 MΩ resistor and (d) the inverting input of op amp 1. (e) The noninverting input of op amp 1 is connected to (f) a 10 KΩ resistor.
In this feedback configuration, op amp 1 adjusts its output voltage to minimize the voltage difference between its two inputs.The right-hand output of op amp 1 mirrors point (c) in this feedback configuration with the following resistors: (f) the 10 KΩ resistor, and the resistors (g) pipette/1000 of resistance R pip /1000, and (h) seal/1000 of resistance R seal /1000.Each of these resistors are thus 1/1000 of the resistance of the corresponding element on the opposite side of the circuit.Conse-quently, the current flowing through the 10 kΩ resistor will be 1000 times greater than the current from the loose patch pipette flowing through the 10 MΩ resistor.The junction (i) between pipette/1000, and seal/1000 is then the input to (j) the inverting terminal of clamping op amp 2. The latter is then compared to the input to (k), its noninverting terminal, from (l) the command voltage step.Again, op amp 2, adjusts its output voltage to minimize the voltage difference between its two inputs.This represents (m) the current required to voltage clamp the junction between pipette/1000, and seal/1000 and, therefore, the junction between pipette and seal resistors.
Manual calibrations and balancing prior to patch clamp measurements must, therefore, set the value of the Pipette/1000 and Seal/1000 resistors.A first, bath, mode, shorts out the Seal/1000 resistor, to enable pipette resistance measurement by applying a square-wave voltage clamp pulse, and adjusting Pipette/1000 to cancel the square-wave leakage current through the pipette resistance.In the second, patch, mode, the resistors are connected as before.On forming the patch, the Leak/1000 control is similarly adjusted to cancel the true leakage current through the seal resistance.Thus, in the actual operation of the system, the variable resistances of the compensating bridge circuit were adjusted to match the voltage drops across R pip and R seal so that the membrane patch is clamped to the command potential and the circuit output corresponds to the current flowing through the patch only.The pipette resistance (R pip ) was then compensated through its corresponding variable resistance in the bridge circuit, under guidance from the current evoked by an imposed square wave voltage clamp waveform viewed on an oscilloscope.In standard recording aCSF, the average R pip recorded was ∼ 180 kΩ.
Both the required circuitry and the initial calibrations of the loose patch thus differ from those of the conventional patch.Thus: (1) The loose patch corrects for much greater leak and consequently pipette currents.(2) The true membrane current is much smaller than the leak current.The latter must then be corrected before feeding this into the actual patch clamp circuit.(3) The values of the different resistors require optimizing for any given application and preparation against the different electrode tip diameters, leak currents, and pipette resistances.More detailed circuit features are summarized in Figure S1.

Bath setup and perfusion apparatus
At the end of the incubation time, a single coronal slice was placed in the bath chamber filled with 30 mL of standard recording aCSF (containing 124 mM NaCl, 2.5 mM KCl, 1.25 mM NaH 2 PO 4 , 24 mM NaHCO 3 , 5 mM HEPES, 12.5 mM glucose, 2 mM MgCl 2 , 2 mM CaCl 2 ; pH 7.3-7.4;T = 20-25 • C) (Figure 2A).The slice was secured on a Sylgard (Dow Chemical Co.) surface through a small chamber and a nylon-stringed harp (model RC-26, Warner Instruments).The bath was connected to a perfusion system driven by a pair of perfusion pumps (model 101UR, Watson-Marlow) that were both regulated to a rate flow of 4 mL/min to minimize mechanical disturbance and

Loose patch pipette manufacture and deployment
Loose patch electrodes (Figure 2B) were pulled from borosilicate glass capillary tubes (GC150-10; Harvard Apparatus) using a horizontal micropipette puller (Model P-97 Sutter Instrument Co.) using a twostep program to achieve a progressive taper.Following the second step, a square tip was achieved.This was confirmed by visualization under a microscope with a calibrated eyepiece reticle at 100× magnification.
Only pipettes with an even and homogenous tip with a 20-25 μm diameter were used.The pipette was inserted into a 45 • angled pipette holder (model Q45W-B15P, Warner Instruments) that was connected to an Ag/AgCl recording electrode.This was connected to the head stage of the loose patch amplifier and held at 45 • such that the pipette tip contacted the brain surface perpendicularly (Figure 2C).The pipette tip was immersed in the bath solution.This was drawn up into the pipette to contact its contained Ag/AgCl wire but leaving an air gap within the electrode shaft.The bath active ground offset potential was then adjusted so that the pipette electrode recorded zero current.
The pipette tip was positioned over the preparation while viewing through a dissection microscope and slowly lowered using a fine vertical micromanipulator (Prior Scientific Instruments Ltd.).Contact with the cell surface produced an increase in the resistance at the pipette tip indicated by a deflection in the oscilloscope trace.The R seal was compensated by adjusting the corresponding resistance in the compensating bridge circuit, and a 25-ms depolarizing step of (RMP [resting membrane potential] + 80) mV was applied to investigate patch viability.If the resulting response did not show a transient inward current followed by transient and sustained outward current components, the patching procedure was reattempted at an adjacent site.To optimize and stabilize the seal, gentle negative pressure was applied through a 1 mL syringe connected to the microelectrode holder, and R seal compensation was adjusted if required.Average R seal varied from patch to patch and ranged between 1.5 and 2.0 times the value of R pip (average ∼ 280 kΩ), in line with values obtained in previous studies on intact neuronal tissues. 41Viable patches were tested with clamp steps over a range of depolarizing voltages to obtain a family of current responses.
After seal acquisition, the bath was perfused with standard recording aCSF to maintain tissue viability by activating the perfusion pumps while monitoring R seal .Small changes (<5%) in R seal were compensated for by readjustment of the corresponding bridge resistance; larger changes signaled disruption of the patch, and these were accordingly discarded.The total duration of a recording from a single patch never exceeded 1 h.
Voltage clamp steps were delivered using an IBM-compatible computer.As the pipette clamped the extracellular face of the patch, the applied voltage steps produced membrane potential excursions of opposite sign to the conventionally expressed membrane potential and were relative to the RMP; they are accordingly referred to as such in this report.A series of membrane-depolarizing clamp steps was used to derive current-voltage curves reflecting channel activation.
Any residual uncompensated leak current was corrected for by a P/4 protocol, whereby four voltage clamp steps of opposite sign and a quarter in magnitude to the test pulse were delivered immediately after it.Since the P/4 pulses spanned voltages that would not activate any voltage-gated conductance, they yielded only leak currents, allowing their removal by summation with the test pulse current record.
Data were sampled at a 50 kHz digital sampling rate and filtered with a DC-10 kHz bandwidth, using a 10 kHz Bessel low pass filter.
The region of interest was visually identified.Membrane currents were recorded in voltage clamp mode, using a custom-made loose patch clamp amplifier.All recordings were carried out at room temperature (20-25 • C).The data were digitized and stored using a custom loose patch software.

Statistical analysis
Current-voltage relationships for activation and inactivation were fitted using custom scripts in the open-source R programming language.
Currents were normalized to the area under the pipette's tip, estimated from the diameter of the pipette, yielding the corresponding current densities, which are expressed as means ± SEM.

Post hoc histological staining
To confirm the electrode recording location, the electrode was lowered on the vertical axis to puncture the tissue.The preparation was then fixed in 10% neutral buffered formalin and embedded in wax using a Sakura VIP6 processor.The samples were sectioned using a HISTO core microtome (Leica) to 2-3 μm thickness.Each slice was then deparaffinized and immersed in 5% ethanol.The slides were stained with Luxol fast blue solution overnight at 37 • C.After washing in 95% ethanol and distilled water, each slide was differentiated in lithium carbonate solution for 3 min, followed by 70% ethanol for 5 min, and washed in distilled water.Each slide was then stained with cresyl violet for 10 min at 57 • C and again differentiated in 95% ethanol.Final dehydration was carried out in 100% ethanol (2 s for three times) and xylene (2 min for three times).Each slide was then cover-slipped and prepared for inspection.

Activation of inward and outward current in hippocampal pyramidal cells
The experiments first investigated the junction between the SP and the SO of the CA1 region, as this was readily identifiable and this could be confirmed histologically in the isolated hippocampal coronal slices.
This SO/SP junction region is known to contain both somas and AISs of pyramidal neurons, a relatively large cell type. 42Initial selection of the latter recording site maximized the likelihoods of recordings from either the pyramidal cell somas or AISs.In contrast, areas such as the SR and SO contain large varieties of smaller interneuron types in addition to pyramidal neuron dendrites. 43 3. To confirm and locate the precise region of each recording, the tissue was punctured at the recording site after each set of recordings by lowering the pipette in the vertical axis.This damaged the tissue locally, thereby marking out the electrode location, which was then identified post hoc by Luxol fast blue staining (Figure 4B).This approach was also employed in studies of all the remaining recording sites described below.

Limiting properties of the loose patch clamp
The present recordings employed pipettes with diameters (20-25 μm, mean = 22 μm) and tip resistances R pip (∼180 kΩ) that give seal resistances (280 kΩ) similar to those used in previous reports in neuronal cells studied in situ (15-30 μm, mean = 24 μm, ∼155 and 270 kΩ, respectively). 41However, seal resistances could be corrected for: small (<5%) variations in R seal were compensated for by readjustment of the corresponding bridge resistance, whereas larger changes signaled disruption of the patch.The currents showed similar features, some reflecting a requirement for progressively larger currents required for seal compensation.These expectedly showed some differences from previous studies in skeletal muscle using similar pipettes (20-25 μm and 150-180 kΩ) but achieving higher seal resistances (up to 2 MΩ), 38 likely reflecting the latter comprising a single, larger, uniformly oriented cell type.Thus, the maximum inward currents showed no indication of reversal within the range of the voltage steps examined (Figure 4C).
Nevertheless, the limiting characteristics of the loose patch clamp method in SO/SP cells could be assessed first by examining membrane currents elicited by depolarizing steps shortened through 15-, 2-, 1-, and 0.5-ms durations in the standard activation protocol used above (Figure 5A).The 15-ms step duration elicited the peak inward and transient and prolonged outward currents also shown in Figure 4. Shortening the step duration to 2 and 1 ms elicited persistent peak inward and transient, but not sustained, outward currents (Figure 5Ab,c).The corresponding current voltage plots suggested intact peak inward but reduced peak outward currents (Figure 5Ae,f).Further shortening to 0.5 ms reduced both inward and outward currents, respectively, to 39% and 2% of those observed with the 15-ms step duration (Figure 5Ad).current decays.This provided an approximate upper limit for the decay τ of the loose patch clamp current.

Separation of inward and outward loose patch current components
The observed inward and outward loose patch clamp currents could be separated into several subcomponents (Figure 6) that agreed with previous reports from other recording techniques and experimental systems.These manipulations employed extracellular NaCl replacement (Figure 6A) by N-methyl-d-glucamine (NMDG) (Figure 6Aa) and choline chloride (Figure 6Ab), which would act on Na + -dependent current components, and 4-aminopyridine (4-AP)-mediated transient 6Ba) and tetraethylammonium (TEA)-mediated transient (Figure 6Bb) and sustained K + current block alone (Figure 6Ba,b) and in combination (Figure 6Bc).In each such control procedure, the reference presence of activating currents was first established in a membrane patch in each separate preparation immersed in aCSF in the absence of a test agent.The bathing solution was then altered to one of Na + -substituted (Figure 6A) or a test solution containing a pharmacological agent (Figure 6B).Here, the volume (120 mL) of test solution washed through exceeded four times the bath volume, ensuring complete replacement.The pipette was then lifted from the seal and its internal solution was replaced with the test extracellular solution.The seal was then re-established by lowering the pipette with the same Comparing typical currents obtained with voltage steps from RMP to (RMP + 120) mV before (dark traces) and following (red traces), these test maneuvers confirmed that Na + replacement (Figure 6Aa,b) blocked peak inward, transient outward, but not prolonged outward current.This was consistent with expectations for effects on Na + and Na + -activated (SLACK and SLICK, encoded by KCNT1 and KCNT2, respectively) K + channels. 33The established transient K + current blocker 4AP (1 mM) indeed reduced transient outward (Figure 6Ba), and the established nonselective transient and sustained K + current blocker TEA (30 mM) indeed reduced both these observed outward current components (Figure 6Bb).Finally, a 4-AP (250 μM) and TEA-Cl (30 mM) combination totally blocked the outward currents, leaving only inward currents (Figure 6Bc).

Voltage-and time-dependent inactivation of peak inward and transient outward currents
The limiting features of loose patch clamp speed and current delivery thus permitted its use to demonstrate and approximate the magnitudes of inward or outward pyramidal cell membrane currents studied in hippocampal slice preparations. 33Although they precluded detailed characterization, particularly of the activation of rapid transient inward (RMP + 30) mV to 30% of its maximum value at (RMP + 80) mV, corresponding to persistent prolonged outward current (Figure 7Ac).The prolonged outward current inactivated by just 25% from its maximal value at (RMP + 120) mV and did not reach its maximal inactivation value within the pulse durations tested (Figure 7Ad).

Localization of pyramidal cell current components to specific CA1 layers
Further experiments extended the SO/SP junctional studies to other hippocampal regions.The SO/SP had formed a readily identifiable recording landmark in the slice preparations, which was confirmed in these experiments by post hoc histological staining.This contains both pyramidal cell somas and their AISs (Figure 8Bc). 42SO contains different interneuron types, including basket cells and O-LM cells, and basal dendrites of pyramidal cells, from which branch their AISs.SP contains pyramidal cell somas.SR and SL-M both contain apical dendritic arborizations of pyramidal cells and different interneuron types, including basket and ivy cells. 43 these cell types, only the largest size pyramidal neurons (soma diameter and area: 37.5 ± 14 SD [n = 88] µm and 948.75 ± 420.25 µm 2 SD [n = 88]) 44 and possibly basket cells (soma diameter and area: 15.6 ± 5.2 SD [n = 63] µm and 213 ± 30 µm 2 SD [n = 20]), 45,46 in contrast to the smaller remaining cells such as ivy cells (soma diameter and area: 10-30 [n = 2] µm and 87-145 µm 2 ), 47,48 are of a size amenable to recording by the relatively large diameter electrodes used by the loose patch clamp approach. 7,8Furthermore, basket cells are potentially distinguishable from pyramidal cells through their smaller peak inward currents than pyramidal cells (∼35 vs. 56 mS/cm 2 ; literature values for outward currents are not available). 49e maximum magnitudes of peak inward current and maximum, prolonged, and transient outward currents recorded in the SO/SP and SP were indistinguishable, consistent with their both arising from similar, likely somal, regions of the same pyramidal neurons (Table 1; see also Figure 8Cc).In contrast, the SR cells showed smaller peak inward, prolonged outward current, and maximum and transient outward current amplitudes than SO/SP, consistent with previous reports examining differences between pyramidal and basket cells (Figure 6Dc).The SO recordings showed smaller peak inward, maximum outward, and transient outward currents, though similar prolonged outward current amplitudes (Figure 8Ac), which could reflect previously reported differing pyramidal cell somal and AIS channel densities.sustained components.Of these, the first, here termed the transient outward current, displayed relatively early inactivation.Possible inactivation of the second, termed the prolonged outward, current was not investigated in the pulse lengths employed.The observed SO/SP currents resembled those from previous studies in conventionally patch clamped isolated pyramidal, as well in situ loose patch clamped hypothalamic neurons. 33These similarities included the transient inward 18,19,21 and at least three different outward components with fast (few ms), slow (up to 100 ms), or noninactivating kinetics, 15 attributed, respectively, to voltage-gated I Na and distinct voltage-gated I K subtypes.This comparison would identify the transient outward currents described here with the previously described fast-inactivating component, and the prolonged outward current with a combination of slowly and noninactivating components.
With increasing sizes of depolarizing voltage steps, the inward current components progressively increased.This contrasts with previous current voltage plots from conventional tight patch clamping or loose patch currents achieving higher seal resistances.These passed through a minimum and then tended toward a reversal potential.Nevertheless, these differences could reflect the progressively greater but increasingly incomplete current compensation for leak currents through the low seal resistances.Similar findings came with previous reports in loose patch clamped neuronal cells with similarly low seal resistances. 33

F I G U R E 1
Circuit features specific to the loose patch clamp method.

2
Experimental loose patch configuration.(A) Overhead view of the experimental setup with a coronal hippocampal slice kept in place using a nylon-stringed recording chamber placed in a 120 × 70 cm bath.The sample is interposed between input and output of the perfusion system.A reference electrode grounding the bath and a bath sensing electrode are shown.(B) Equivalent circuit of loose patch clamp electrode in contact with membrane.Compensation for the voltage error arising from currents flowing through the series combination of the pipette resistance (R pip ) and the seal resistance (R seal ) employed a bridge circuit in the custom-designed loose patch clamp amplifier.(C) Front view of the experimental setup showing the amplifier head stage connected to a 45 • angled electrode holder, in which the pipette was inserted.The head stage was inclined by a further 45 • so that the electrode tip would make perpendicular contact with the preparation during recording.The suction syringe is also displayed.maintain constant bath fluid levels.The bath ground-sensing electrode, a sheathed Ag/AgCl electrode in contact with the bath solution, was used to maintain the bath at ground potential.A nonchloridized plate electrode was used as the active ground current-passing electrode.The perfusing solutions were equilibrated for >1 h prior to experimentation and constantly bubbled throughout the experiment with a mixture of 95% O 2 /5% CO 2 .

F I G U R E 3
Pulse protocols used in the study.(A, a) Pulse protocol used for voltage dependence of current activation, imposing two voltage steps over a time course of 30 ms before restoring RMP.A 5-ms hyperpolarizing prepulse (to voltage V 0 ) at (RMP -40) mV relieved any residual Na v inactivation at the RMP.This was followed by a variable test pulse (to voltage V 1 ) of 15-ms duration starting at (RMP -40) mV and altered in (RMP + 10) mV increments until a maximum test voltage of (RMP + 120) mV was reached.The RMP was finally restored after 21 ms of recording time (to voltage V 2 ).(A,b) To investigate the properties of the clamp in time, the length of the step to voltage V 1 was varied to a 2-ms (2.), 1-ms (3.), and 0.5-ms (4.) duration.(B).Pulse protocol for the study of current inactivation consisted of a single depolarizing step at (RMP + 100) mV of 34-ms duration (at voltage V 2 ), imposed subsequently to a variable test pulse (to voltage V 1 ) of 10-ms duration starting at (RMP -40) mV and altered in (RMP + 10) mV increments until a maximum test voltage of (RMP + 120) mV.(C) To explore time-dependent recovery from inactivation, a 4.5-ms duration conditioning step to voltage V 1 = (RMP + 120) mV was applied.The holding potential was then restored to V 2 = RMP for 4.5 ms.The subsequent test steps to V 3 = (RMP + 120) mV were imposed after different time intervals, ∆T, between 10 and 38 ms in 2 ms increments.Finally, a hyperpolarizing step at (RMP -40) mV of 9-ms duration was applied at T = 41 ms (at voltage V 4 ).A standard activation protocol examined membrane responses to applied depolarizing steps (Figure3A).Patches were first held at the RMP for 1 ms.All voltage excursions are made from this background resting potential (see, e.g., Refs.30 and 38).A prepulse of 5-ms duration to V 0 = (RMP − 40) mV was then imposed to minimize background Na V channel inactivation.This was followed by successively larger 15ms duration depolarizing test voltage steps starting at (RMP -40) mV.Their amplitude was successively increased in 10 mV increments up to V 1 = (RMP + 120) mV.The membrane potential was then finally restored to V 2 = RMP after 21 ms of recording time.The depolarizing steps produced a first detectable response at around (RMP + 40) mV (Figure4A).This consisted of a transient inward followed by an outward current.The latter included two distinct, transient and prolonged, components.The inward current reached its maximum amplitude (mean: 59.26 ± 1.48 pA/μm 2 , N 1 = 30, N 2 = 120, n = 120) after ∼1 ms following the imposition of voltage steps to the most depolarized voltage of V 1 = (RMP + 120) mV; this was followed by a rapid decay.The subsequent outward currents transiently rose to a maximum amplitude (mean: 103.16 ± 2.33 pA/μm 2 ), named here the maximum outward current, at ∼2.5 ms after imposition of that voltage step.The outward current amplitude then decayed to a stable prolonged value, the prolonged outward current, with an average maximum amplitude of 45.61 ± 0.98 pA/μm 2 .A transient outward current could be determined from the difference between the peak of the outward current and the amplitude of the prolonged outward current.This had an average maximum amplitude of 57.54 ± 1.87 pA/μm 2 .

4
Activation properties of inward currents and outward currents of murine neurons in the CA1 SO/SP junction under loose patch clamp.(A) Typical currents elicited by progressively depolarizing voltage steps are shown for a typical patch (R seal : 260 kΩ, pipette diameter: 22 μm).Initially, an inward current develops in response to membrane depolarization.This is followed by an outward current showing two distinct components: transient and prolonged.We defined the maximum outward and prolonged outward currents as the respective peak and final steady state value of the outward current phase, and their difference as the maximum transient outward current.(B) Diagram in (a) shows the electrode location relative to the hippocampal layers.This was confirmed through post hoc staining with cresyl violet (b).(C) Mean currents plotted against voltage excursion; the resulting current−voltage curves, expressing current density normalized to the pipette diameter, are shown in (a), (b), (c), and (d), respectively, for peak inward, maximum outward, prolonged outward, and transient outward currents.N = 120 patches, from 30 independent experiments.Average R seal : 300 kΩ.RMP, resting membrane potential.N 1 = 30, N 2 = 120, n = 120.

Second, repolarizing loose
patch clamp steps terminating the depolarizing steps imposed hyperpolarizing steps to a range of membrane potentials extending to a membrane hyperpolarization of (RMP -80) mV.A four-step protocol (Figure5Ba, top panel) first held patches at the RMP for 5 ms, then applied a hyperpolarizing 10-ms duration prepulse to voltage V 0 .A 15-ms duration depolarizing step was next imposed to voltage V 1 = (RMP + 140) mV.A subsequent 15-ms duration step repolarized the membrane to a range of levels V 2 .Successive sweeps altered V 2 in (RMP -10) mV increments to a final value of (RMP -80) mV.Finally, a 10-ms duration hyperpolarizing was made to a voltage V 3 = (RMP -40) mV.The repolarizing steps elicited monotonic decays (Figure4).Their overall time course would reflect the recovery properties of both clamp-and voltage-dependent ion currents (Figure5Ba, bottom panel).

Figure 5Bb illustrates results
Figure 5Bb illustrates results obtained from a typical pulse protocol involving two repolarizing voltage steps to (RMP + 100) and (RMP -40) mV.Single exponential fits were used to approximate the time course of these current relaxations.The initial repolarizing step to (RMP + 100) elicited a decay tail of overall time constant τ = 2.95 ms.The repolarizing step to (RMP -40) mV produced a faster decay of τ = 1.94 ms. Figure 5Bc plots the time constants against voltage.These decreased with hyperpolarization, tending toward a limiting minimum τ that reflects the loose patch clamp and the most rapid membrane

5
Loose patch clamp responses from the neurons of the murine hippocampal SO/SP junction to varying pulse protocols to explore clamp characteristics.(A) Responses obtained to progressively shortened pulse durations to V 1 in a typical patch (R seal : 240 kΩ, pipette diameter: 24 μm).(a) Current components observed with 15-ms duration depolarizing steps.(b) Persistence of peak inward and maximum outward current amplitude but not prolonged outward current with 2-ms duration depolarizing steps.(c) Persistence of maximal value of peak inward current even at large depolarization steps (RMP + 100 to RMP + 160) mV but reduced maximum outward phase amplitude with 1-ms duration depolarizing steps.(d) Current response that entirely develops after the cessation of the stimulus and fails to achieve maximal values in all the current components with 0.5-ms duration depolarization steps.(e, f) Current-voltage curves for peak inward (e) and maximum outward currents (f).N = 5 patches, from five independent experiments.Average R seal : 260 kΩ.(B) Loose patch clamp currents at the end of depolarizing voltage steps.Patches were first held at the RMP for 5 ms.Hyperpolarizing 10-ms duration prepulses to voltage V 0 were then applied, followed by a depolarizing 15-ms duration step to a single V 1 = (RMP + 140) mV.The membrane was subject to repolarizing 15-ms duration steps to voltage V 2 altered through successive sweeps in (RMP -10) mV increments to a final value of (RMP -80) mV.A final hyperpolarizing step of 10-ms duration altered the voltage to V 3 = (RMP -40) mV.Panel (a): family of currents for a typical patch (R seal : 280 kΩ, pipette diameter: 22 μm).Single exponentials were fitted to each tail at each voltage.Panel (b): two illustrative sets of results.At (RMP + 100) mV, the current following the small repolarizing step shows an overall time constant τ of 2.95 ms.At (RMP -40) mV, the current following the strongly repolarizing step shows an overall time constant τ of 1.94 ms.(c) Plots of mean ± SEM fitted time constant as a function of voltage.The fitted time constant of the tail monotonically increases with increasing depolarization.n = 5 patches, from five independent experiments.Average R seal : 270 kΩ.RMP, resting membrane potential.
currents, it was possible to assess voltage dependences of inward and outward current inactivation by sustained voltage steps and its recovery (Figure 7A).Patches were first held at the RMP for 1 ms.A 5-ms duration prepulse stepped the voltage to V 0 = (RMP − 40) mV.Depolarizing 10-ms duration conditioning steps then altered membrane voltage to V 1 in 17 successive sweeps by (RMP + 10) mV increments between (RMP − 40) mV and (RMP + 120) mV (Figure 3B).These conditioning steps to V 1 elicited the previously characterized inward, transient, and prolonged outward currents.A final 34-ms duration step was made to a fixed voltage V 2 = (RMP + 100) mV lasting to the end of the total 50 ms record length.The step to V 2 elicited inward and transient outward currents reduced in amplitude to extents dependent upon the prior voltage prepulse to V 1 .The peak inward current showed a relatively steep inactivation function falling between (RMP + 10) mV and (RMP + 50) mV (Figure 7Ab).The maximum outward current showed a more gradual inactivation function falling between

Figure
Figure7Bexemplifies time courses of recovery from the inactivation of inward and outward current components following the restoration of the membrane potential.The pulse protocol first held the potential at the RMP for 1 ms from the beginning of the recording period.A 4.5-ms duration step to V 1 = (RMP + 120) mV then elicited a maximal current response, followed by inactivation.The membrane potential was then restored to V 2 = RMP for 4.5 ms.A subsequent set of steps to V 3 = (RMP + 120) mV was imposed at successively increasing 10-38 ms time intervals in ∆T = 2 ms increments (Figure3C).These steps elicited both inward and outward currents whose maximum amplitude, normalized to their corresponding values in the V 1 step, reflected current recovery from inactivation with time.Finally, in all sweeps, the voltage was stepped to a hyperpolarizing level of V 4 = (RMP -40) mV, applied at T = 41 ms, for 9 ms.Full recovery was then achieved for both peak inward current (88% after 15-ms, 97% at 27 ms) and maximum outward current (92% after 15-ms, full recovery at 24-ms) and was characterized by an exponential time course.
outward I inactivation Test pulse (mV relative to RMP)

7
Inactivation properties and recovery from inactivation shown by peak inward current and maximum outward current recorded in murine hippocampal SO/SP junction neurons under loose patch clamp.(A) Typical recordings in response to inactivation pulse protocol beginning from the RMP for a typical patch (a) (R seal : 220 kΩ, pipette diameter: 22 μm).The currents were quantified and plotted against the voltage excursion for the conditioning voltage step, by selecting an appropriate time window (I min in the 14-18 ms interval for peak inward current; I max in the 16-22 ms interval for maximum outward current; and I max in the 30-36 ms interval for prolonged outward current).The resulting current-voltage curves, expressing current density normalized to the pipette diameter, are shown for peak inward current (b), maximum outward current (c), and prolonged outward current (d).N = 15 patches, from 15 coronal slices selected from 15 brains.Average R seal : 230 kΩ.(B) Currents illustrating recovery from inactivation of peak inward current and maximum outward current following the restoration of the membrane potential.(a) Typical recordings are shown in response to time-dependent recovery pulse protocol for a typical patch (R seal : 250 kΩ, pipette diameter: 24 μm).The currents were plotted against time intervening between termination of the conditioning and imposition of the test pulse.The resulting current-time curves, expressing current density normalized to the pipette diameter, are shown in (b) and (c), respectively, for peak inward current and maximum outward current recovery.N = 15 patches, from 15 independent experiments.Average R seal : 240 kΩ.RMP, resting membrane potential.

F
I G U R E 8 (A-D) Comparisons of membrane currents from different strata of the CA1 hippocampal region.A first set of currents was recorded from neurons in the SO/SP junction in response to standard activation protocol (R seal : 290 kΩ, pipette diameter: 22 µm) (B).From here, the pipette was lifted and moved to the center of the SP (R seal : 300 kΩ, pipette diameter: 22 µm) (C), where a new set of currents was acquired.Two additional patches were symmetrically acquired in the SO (R seal : 310 kΩ, pipette diameter: 22 µm) (A) and the SR (R seal : 270 kΩ, pipette diameter: 22 µm) (D).(E) Mean (± SEM) maximum currents from all patches plotted against voltage excursion, giving current-voltage curves.Peak inward (a), maximum outward (b), prolonged outward (c), and transient outward (d) currents are displayed.The currents derived from the SO/SP junction and the SP neurons are similar in amplitude and waveform, suggesting that such recordings reflect electrical activity in the CA1 pyramidal neuron somas.In contrast, recordings obtained in the SO and SR show lower amplitude currents and an outward phase dominated by the prolonged component.N = 5 patches (for region) from five independent experiments.Average R seal : 290 kΩ.RMP, resting membrane potential.TAB L E 1Maximal amplitude current densities for current components in different hippocampal recording sites.
Furthermore, assessments made of clamp speed or time constant first demonstrated that reducing the duration of even large pulses to between (RMP + 100) and (RMP + 160) mV down to even 2 ms continued to elicit peak inward and maximum outward current amplitude.Further pulse duration shortening to <1 and <2 ms progressively compromised those inward and outward currents.These limiting characteristics in the voltage clamping time course agreed with time constants obtained with membrane repolarizations to progressively positive membrane potentials.Simple assessments of the limitations of the loose patch clamp technique outlined in the Introduction, as applicable in this specific preparation, thus suggested that it could provide indications of the existence of neuronal membrane current contributions and their relative magnitudes in hippocampal slices in situ.The pulse protocols do involve expressing membrane voltage relative to the resting potential rather than absolute membrane potentials.In addition, loose patch electrodes make smaller membrane seal resistances, limiting their current compensation at large voltage excursions and determination of reversal potentials of the inward currents.33,51Although it would not quantify inward current kinetics in detail, it also remains useful in determinations of current inactivation properties.The latter measurements demonstrated distinct inactivation voltage dependences and time dependences of their recovery in the inward and transient and prolonged outward currents.Nevertheless, cells could be studied, without the enzymatic treatments following isolation that are required in conventional patch clamp studies, in relatively unperturbed tissue preparations in situ, superfused with physiological solutions.The method further permitted multiple use of the same recording pipettes for comparing results obtained at different recording sites.This feature is potentially useful in protocols involving applications and withdrawals of pharmacological agents.Finally, single loose patch clamp electrodes could similarly be used repetitively to compare recordings from multiple recording sites.Here, findings from the easily identified SO/SP were compared with those from the SP, containing pyramidal neuron somas, the SO their dendrites, and the SR, the latter containing further, different interneuron subtypes.The SP recordings showed currents with similar characteristics consistent with both SO/SP and SP currents being derived from pyramidal cell somas.In contrast, the SO showed smaller peak inward, maximum outward, and transient outward currents, though similar prolonged outward current amplitudes.The SR showed smaller peak inward and prolonged outward current and maximum outward current and smaller transient outward current amplitudes than SP and SO, differences consistent with a distinct, possibly basket, SR cell type.49In conclusion, this study introduces intact tissue loose patch clamping as a technique for the study of ion currents produced by voltagegated channels on the membrane of hippocampal neurons.It confirmed features previously demonstrated by the conventional patch clamp technique.It would offer a complementary technique in allowing in situ recordings in minimally perturbed cells within brain slices, as opposed to isolated cells.The loose patch clamp permits multiple recordings both between recording sites and before and following pharmacological solution changes.