long duration perforated patch recordings from spinal...
TRANSCRIPT
1
1 Innovative Methodology 2
3 4 5 6
Long duration perforated patch recordings from spinal 7
interneurons of adult mice 8
9 10
Andreas Husch, Nathan Cramer and Ronald M. Harris-Warrick 11 Department of Neurobiology and Behavior, Cornell University, Ithaca, NY, 14853 12
13 14 15 16 17 18
Running head: Perforated patch recordings from adult spinal interneurons 19 20 21 Contact information: 22 Andreas Husch 23 Department of Neurobiology and Behavior 24 Seeley Mudd Hall 25 Cornell University 26 Ithaca NY 14853 27 28 Fax: (607)-254-4316 29 Phone: (607)-254-4347 30 Email: [email protected] 31 32 Pages: 17 33 Figures: 5 34 35 Abstract 212 words 36 Introduction: 439 words 37 Discussion: 688 words 38
39
Articles in PresS. J Neurophysiol (September 7, 2011). doi:10.1152/jn.00673.2011
Copyright © 2011 by the American Physiological Society.
Articles in PresS. J Neurophysiol (September 7, 2011). doi:10.1152/jn.00673.2011
Copyright © 2011 by the American Physiological Society.
2
Abstract 40
It has been very difficult to record from interneurons in acute slices of the lumbar spinal cord 41
from mice older than three weeks of age. The low success rate and short recording times limit 42
in vitro experimentation on mouse spinal networks to neonatal and early postnatal periods, 43
when locomotor networks are still developmentally immature. To overcome this limitation and 44
enable investigation of mature locomotor network neurons, we have established a reliable 45
procedure to record from spinal cord neurons in slices from adult, behaviorally mature mice of 46
any age. Two key changes to the established neonate procedure were implemented. First, we 47
remove the cord by a dorsal laminectomy from a deeply anaesthetized animal. This enables 48
respiration and other vital functions to continue up to the moment the maximally oxygenated 49
lumbar spinal cord is removed, improving the health of the slices. Second, since adult spinal 50
cord interneurons appear more sensitive to the intracellular dialysis that occurs during whole cell 51
recordings, we introduced perforated patch recordings to the procedure. Stable recordings up 52
to 12 hours in duration were obtained with our new method. This will allow investigation of 53
changes in mature neuronal properties in disease states or after spinal cord injury, and also 54
allow prolonged recordings of responses to drug application that were previously impossible. 55
56
57
Keywords 58
Spinal cord, adult mice, perforated patch 59
60
3
Introduction 61
The spinal cord contains the Central Pattern Generator (CPG) networks that organize the motor 62
patterns for locomotion. These neuronal networks are composed of a set of interacting spinal 63
interneurons that generate the timing, phasing and intensity cues for the motoneurons to drive 64
rhythmic leg movements (Grillner and Jessell 2009; Kiehn 2006; Pearson 1993). As with other 65
neural networks, the output of the locomotor CPG depends on both the pattern of synaptic 66
connectivity and the intrinsic electrophysiological properties of the component interneurons. 67
Understanding these properties is a major area of current research, because they may change 68
in disease states or after spinal cord injury (SCI). 69
The intrinsic firing properties of spinal interneurons can be studied through a variety of 70
techniques. Blind patch recordings or visually guided recording of genetically tagged 71
interneurons have been used in the isolated largely intact spinal cord (Carlin et al. 2006; 72
Gosgnach et al. 2006; Kiehn and Butt 2003; Kjaerulff and Kiehn 1996; Lanuza et al. 2004; 73
Zhong et al. 2006; Zhong et al. 2007). However, this is only feasible in neonatal and early 74
postnatal animals: as the diameter of the spinal cord increases with age, adequate oxygenation 75
of the ventromedial region of the cord, where the locomotor CPG is located, decreases (Wilson 76
et al. 2003) leading to hypoxia-induced cell death. In principle, this can be avoided by studying 77
the firing properties of interneurons in slices of the cord, where greater control over the 78
extracellular environment can typically be achieved (Dougherty and Kiehn 2010; Wilson et al. 79
2010; Wilson et al. 2005; Zhong et al. 2010). However, this has also proven to be very difficult 80
in slices older than 1-2 weeks because changes in the extracellular matrix and the perineuronal 81
sheath seem to restrict access to the neuronal somata (Koppe et al. 1997; Milev et al. 1998), 82
making stable intracellular recordings more difficult (Morales et al. 2004). Consequently, even 83
though newborn mice are not able to walk in their first week of life (Clarac et al. 2004), most in 84
vitro studies of the cellular properties of spinal interneurons are conducted on P1-P5 neonates 85
4
(Dougherty and Kiehn 2010; Gosgnach et al. 2006; Lanuza et al. 2004; Zhang et al. 2008; 86
Zhong et al. 2010; Zhong et al. 2007). Since significant developmental changes occur between 87
the neonatal period and the locomotor-mature age (2-4 weeks) (Jiang et al. 1999; Song et al. 88
2006), it is critically important to study mature interneurons in the investigation of spinal 89
locomotor networks. In addition, studies of adult neurons are essential to analyze their 90
responses to SCI, as these injuries occur most frequently in mature humans. Thus, animal 91
models of SCI should be performed with adults, to separate injury-induced changes in neuronal 92
properties from those evoked by interruption of the normal developmental schedule. 93
In this paper, we describe a method to record from identified spinal interneurons from 94
mice in slices of any age. Establishing this protocol will allow studies of the developmental 95
changes between neonatal stage and locomotor mature mice, and open up new possibilities to 96
investigate mid- and long-term changes after SCI on a cellular level. 97
98
Materials and Methods 99
Animals 100
Experiments were performed using adult (postnatal (P) day P43–P93) Chx10::CFP mice 101
generated by Drs. Steven Crone and Kamal Sharma at the University of Chicago. The animal 102
protocol was approved by the Institutional Animal Care and Use Committee at Cornell University 103
and was in accordance with National Institutes of Health guidelines. 104
105
Spinal cord slice procedure 106
Animals were deeply anesthetized with ketamine (1.5 mg/10g body weight) and xylazine (0.15 107
mg/10 g body weight). Nair (commercially available depilatory cream) was applied to remove 108
the fur in the dorsal area between the neck and the sacral region of the spine. The mouse was 109
placed on ice under a binocular microscope and administered pure oxygen (Fig. 1A). Access to 110
5
the sacral-mid-thoracic spinal column was gained from the dorsal surface (Fig. 1B), . A midline 111
incision along the thoracic vertebrae was made, and a dorsal laminectomy was performed 112
(Fig. 1C). To further cool the spinal cord, it was constantly superfused with ice-cold (0-4°C) 113
oxygenated (95% O2 and 5% CO2), glycerol-based modified artificial cerebrospinal fluid 114
(GACSF; Ye et al. 2006) which contained (in mM): 222 glycerol, 3.08 KCl, 1.18 KH2PO4, 1.25 115
MgSO4, 2.52 CaCl2, 25 NaHCO3, 11 D-glucose (~300 mOsm). With fine scissors, the ventral 116
and dorsal roots were transected. The cord was dissected from the lumbar spinal cord and 117
quickly transferred to a Sylgard-coated (Dow Corning, Midland, MI) petri dish containing ice-cold 118
GACSF and pinned with fine pins ventral side up. The ventral meninges and ventral roots were 119
rapidly removed by pulling the meninges with 2 fine forceps from anterior to posterior (Fig. 1D). 120
The cord was then turned over, and the dorsal meninges were removed. The spinal cord was 121
transferred into a small custom built chamber filled with 34°C low melting point agarose (A0701, 122
Sigma), which was immediately placed in ice-water slurry to rapidly cool and harden. The 123
agarose block containing the lumbar enlargement was trimmed to shape with a razor blade and 124
glued (Loctite 406, Henkel) on a metal vibratome disk. The disk was transferred into the 125
microtome buffer tray, which was filled with ice cold GACSF (kept cold with Microm CU65 126
cooling device, Thermo Scientific). 250 µm sections were sliced with a vibrating blade 127
microtome (HM-650 V; Thermo Scientific, advance speed: 0.9 mm/s; vibratory frequency: 100 128
Hz; vibratory amplitude: 0.7 mm; blade angle: -15°; blade: Feather Double-Edge Blade, Product 129
No.121-9, TED PELLA, Inc). The slices were immediately transferred into 35°C oxygenated 130
ACSF to recover for 45 minutes. ACSF contained (in mM): 111 NaCl, 3.08 KCl, 1.18 KH2PO4, 131
1.25 MgSO4, 2.52 CaCl2, 25 NaHCO3, 11 D-glucose (280 mOsm). The incubation chamber 132
was similar to the chamber used by Edwards et al. (1989). The slices were kept submerged 133
under the ACSF, while the solution was bubbled vigorously, but not so strongly as to lead to 134
damaging slice movements. It was essential for the steps from removal of the cord to slicing to 135
be performed as rapidly as possible for good slice viability, optimally within 10 min. Afterwards, 136
6
the slices were allowed to passively cool down to room temperature and recover for one hour. 137
The spinal neurons were visualized with a fixed-stage upright microscope (BX51WI, Olympus) 138
using a 60x water-immersion objective (LUMPLFLN 60XW, 1 numerical aperture, 2 mm working 139
distance, Olympus) with infrared-differential interference contrast (IR-DIC) and fluorescence 140
optics. 141
142
Perforated patch recordings 143
Slices were transferred to a recording chamber (~3 ml volume) and continuously superfused 144
with oxygenated ACSF at a flow rate of ~2 ml/min. To block most synaptic input in the slices, 145
neurons were isolated from rapid synaptic inputs with a combination of DL-2-Amino-5-146
phosphonopentanoic acid (AP-5, 10 µM; A5282, Sigma) and CNQX disodium salt hydrate 147
(CNQX, 10 µM; C239, Sigma) to block glutamatergic synapses, picrotoxin (10 µM) to block 148
GABAergic synapses, and strychnine (10 µM) to block glycinergic synapses. Perforated patch 149
recordings (PPRs)(Horn and Marty 1988; Rae et al. 1991) were made with thick walled, 150
unfilamented borosilicate glass (1.5 mm outer diameter, 1.0 mm inner diameter, PG52151-4, 151
WPI) on a vertical puller (PC-10, Narishige) with low resistances of 3-5 MΩ. The tip of the 152
pipette was first filled with intracellular solution containing (in mM) 135 K-gluconate, 10 KCl, 10 153
Hepes, 0.1 EGTA, 2 MgCl2 (adjusted to pH 7.2 with KOH, ~270 mOsm) by placing the pipette 154
tip side down into an 1.5 ml eppendorf cap filled with ~1.3 ml intracellular solution and applying 155
5-7 ml of negative pressure with a 10 ml syringe for 1 second. The pipette was then backfilled 156
with a combination of intracellular solution, Amphotericin B (A4888, Sigma) and pluronic F127 157
(P-2443, Sigma) (Herrington et al. 1995; Lovell and McCobb 2001). To prepare the solution, 1.2 158
mg Amphotericin B was dissolved in 20 µl dimethyl sulfoxide (DMSO, D8418, Sigma) and added 159
to 1 mg Pluronic acid F127 dissolved in 40 µl DMSO. The 60 µl Amphotericin B/pluronic 160
F127/DMSO-mix was added and vortexed in 1 ml intracellular solution. The ionophore mix was 161
stored at room temperature and replaced every hour as needed. The cells were approached 162
7
with application of very small positive air pressure (via the mouth) to the patch pipette to reach 163
cells up to several layers below the surface, but to avoid forcing the ionophore mix to the tip 164
which would impede seal formation. Close to the clearly visible membrane the positive pressure 165
was converted to a slight suction to obtain a gigaohm seal. The intact membrane patch in the 166
pipette as well as the eCFP signal in the neuron was observed during the experiment, to confirm 167
the stability of the perforated patch configuration. 168
169
Data acquisition and analysis 170
Current clamp recordings were made with a Multiclamp 700B amplifier (Molecular Devices) 171
controlled by Clampex (pClamp 9, Molecular Devices). Data were sampled at 10 kHz and low-172
pass filtered at 2 kHz. For action potential analysis, the membrane potential was adjusted to set 173
the firing frequency at 1 Hz or lower to elicit temporally isolated action potentials. The voltage 174
threshold for action potential generation was measured as the peak of the second derivative of 175
voltage with time during the rising phase of the action potential. The spike amplitude was 176
measured from the peak of the action potential to the peak afterhyperpolarization. The action 177
potential half-width was established at the voltage halfway from the spike threshold to the peak 178
of the action potential. To measure the membrane input resistance and rheobase, all neurons 179
were held below threshold at -60 mV with holding current (Ihold). Input resistance was estimated 180
by averaging the response to small hyperpolarizing current pulses (1s duration). The minimal 181
amount of current necessary for spike generation was defined as the rheobase. To measure the 182
spontaneous firing rate, the mean firing rate of twelve 10 second bins was averaged. 183
Data analysis was performed with Clampfit (Axon Instruments), Spike2 (CED), Igor Pro 6 184
(Wavemetrics), Graph Pad Prism (GraphPad Software inc.) and Matlab (MathWorks). Data are 185
given as mean ± standard deviation. To determine differences in parameter means unpaired or 186
paired t-tests were performed as appropriate. A significance level of 0.05 was accepted for all 187
tests. 188
8
189
Results 190
In earlier work, we used the standard neonatal procedure to attempt recordings from adult (P50 191
and older) interneurons. However, the recordings were very difficult to obtain, very unstable 192
and of short duration. Thus, we developed a new procedure which improved the protocol in 193
several ways to allow reliable recording from adult spinal interneurons of any age for long 194
durations. Two key changes made the difference. First, the dissection of the spinal cord from a 195
dorsal approach, in a physiologically functioning mouse increased the health of the slices and 196
the viability of the interneurons. The dorsal approach was previously described for spinal 197
motoneurons in adult rats by Carp et al. (2008). Second, we introduced the PPR technique 198
which avoids replacement of the intracellular contents with the electrode solution and enabled 199
us to make stable long-term recordings from the healthy neurons. 200
201
Comparison with the standard ‘neonate spinal cord slice procedure’ 202
To illustrate the benefits of the modified procedure we compared recordings from adult (P56-60) 203
V2a interneurons using the standard method (ventral dissection from decapitated animal and 204
whole cell recordings (Dougherty and Kiehn 2010; Wilson et al. 2010; Wilson et al. 2005; Zhong 205
et al. 2010)) and the new method (dorsal dissection from living animal, P43-93 and PPR 206
recordings), and give examples of the stability of long term recordings only possible with PPR. 207
In the standard procedure deeply anesthetized animals were killed by decapitation, eviscerated, 208
and their spinal cords were ventrally dissected from the gut and isolated under ice-cold (4°C) 209
oxygenated (95%O2/5%CO2) modified sucrose based ACSF (in mM: 206 sucrose, 2 KCl, 26 210
NaHCO3, 1.25 NaH2PO4, 1 MgCl, 2 MgSO4, 1 CaCl2, and 20 d-glucose). Whole cell patch clamp 211
recordings (WCR)(Hamill et al. 1981) were performed with pipette solution containing (in mM): 212
138 K-gluconate, 10 HEPES, 5 ATP-Mg, 0.3 GTP-Li, and 0.0001 CaCl2 (pH 7.4 with KOH, 213
9
~240 mOsm). With 29 approaches to cells (3 preparations) using the standard neonatal slice 214
procedure, we were able to form 14 GΩ-seals, resulting in 11 successful whole-cell recordings 215
(WCRs). In comparison using the PPR and a dorsal spinal cord dissection, 26 pipette 216
approaches (12 preparations) resulted in 20 GΩ-seals, of which 11 led to stable recordings. 217
Using the new method, PPRs were performed as described in the Methods section. 218
After ensuring a GΩ seal, the perforation progress was monitored by the membrane potential 219
and the amplitude of the action potentials (Fig. 2A). The recorded membrane potential was 220
initially 0 mV, but over time the Amphotericin B diffused to the pipette tip and inserted into the 221
membrane. This enabled recording of the membrane potential after 10 ± 5 min (n = 23; 222
Fig. 2B). Initially the action potentials were small, as the initial very high series access 223
resistance (>>100 MΩ) causes low pass filtering of the measured signal (Fig. 2A). With further 224
perforation of the membrane, the series access resistance dropped to lower values (<100 MΩ); 225
after 35 ± 9 min (n = 23) the action potentials began to overshoot 0 mV. Over the period of 226
approximately an hour, the continuously dropping series access resistance eventually stabilized 227
at 15 to 25 MΩ. Thus, the neuron was ready for recording about one hour after GΩ seal 228
formation. 229
The PPRs lasted much longer than the traditional WCR measurements. Most of the 230
PPRs we conducted with the new method were discontinued by the investigator after about 5 231
hours, resulting in an average recording time of 5.3 ± 2.3 h ( n = 23; Fig. 2C). However, in cases 232
where we did not terminate the recording, it was possible to record as long as 12 hours from a 233
single spinal interneuron (Fig. 2C). In contrast to this, using the standard neonatal dissection, 234
the WCRs lasted only 35 ± 27 minutes (n = 11); this is consistent with our many adult WCR 235
recordings from other interneuron types (data not shown). 236
In addition to the relatively short recording duration using the old WCR approach, the 237
quality of the recording was poorer than with the PPRs. For example, using the standard 238
method, the initial membrane potential was more depolarized. In V2a neurons whose firing 239
10
frequency was low enough to accurately measure the membrane potential (AP frequency 240
≤ 1Hz), the potential was -40.3 ± 14.5 mV in WCRs (n = 5), while using PPRs it was in a more 241
hyperpolarized range with a smaller standard deviation (-54.1 ± 7.3 mV; n = 11; p = 0.02, t-test; 242
Fig. 3A). The depolarized membrane potential in WCRs was also reflected by significantly 243
greater negative current necessary to hold the neuron at -60 mV (p = 0.0045, t-test; Fig. 3B). 244
While in the WCR the Ihold at -60 mV was -34 ± 40 pA (n = 11), the PPRs only needed -245
6.7 ± 8.8 pA (n = 22; again less variability) to hold the neuron at -60 mV. This difference is in 246
part explained by the significantly lower input resistances measured in the neurons using the 247
WCR procedure (WCR, 0.8 ± 0.6 GΩ; n = 11. PPR, 1.6 ± 0.5 GΩ; n = 18; p = 0.002, t-test; 248
Fig. 3C). Reflecting the difference in input resistance, the rheobase, or current required to evoke 249
an action potential from -60 mV, was larger using the standard WCR procedure than the new 250
PP procedure (Fig. 3D). Both the rheobase (34 pA) and its variability (± 50 pA SD) were larger 251
using WCR than with PPR (7 ± 5 pA). 252
While the rate of spontaneous action potential firing in the new procedure was very 253
stable over hours, the spontaneous firing rate using the standard WCR procedure significantly 254
decreased during the first 10 minutes of recording (Fig. 3, E and G). In neurons where the initial 255
firing rate was higher than 0.5 Hz, the WCR showed a significant drop in firing rate from initial 256
(membrane breakthrough) 3 ± 2.7 Hz to 0.3 ± 0.8 Hz 10 minutes later (n = 9; p = 0.004, paired 257
t-test; Fig. 3E,G). The reduction in basal firing rate was often associated with either a variable 258
hyperpolarization (Fig. 3E) or a depolarization (data not shown) of the neuron. The 259
hyperpolarization may arise in part from activation of KATP currents (see discussion). Using the 260
perforated patch method, the firing rate was very stable over time (Fig. 3F,H). The initial firing 261
rate when the APs became overshooting was 3.2 ± 1.1 Hz, and 10 min later was essentially the 262
same (3.5 ± 0.4 Hz; n = 7; p = 0.4, paired t-test; Fig. 3H). 263
264
Stability in recordings with adult slice preparation and perforated patch 265
11
In addition to the greater length of recording using PPR in adult spinal interneurons, the 266
excitability was generally very stable for many hours. The spontaneous firing frequency (with no 267
holding current) remained constant between 1 and 2 Hz for at least 4 hours (Figure 4A); 268
individual neurons had different initial firing rates, but maintained them at a similar frequency for 269
the duration of the recordings. Intrinsic parameters such as membrane input resistance, the IHold 270
at -60 mV, and the rheobase also remained constant over time. The amount of current needed 271
to hold the cells at -60 mV remained stable over the course of the recordings (-14.4 ± 13.6 pA to 272
-14.4 ± 11.1 pA after 4 hours; Fig. 4B). Once the series access resistance had stabilized, the 273
input resistance stayed approximately the same over many hours, reflecting the healthiness of 274
the recorded neuron and the noninvasive nature of the PPR. The mean input resistance after 275
stable perforation of 1.3 ± 0.5 GΩ was essentially the same 4 hours later (1.2 ± 0.4 GΩ; 276
Fig. 4C). The neuronal excitability, as reflected by the stability of the rheobase value, did not 277
change significantly over the first 4 hours (6.7 ± 2.6 pA after 1 hour vs. 6.6 ± 3.9 pA at 4 hours; 278
Fig. 4D). 279
Once the ionophore penetration into the membrane reached steady state, in most 280
recordings the properties of the action potentials (amplitude, half-width, and threshold) remained 281
stable over many hours (Fig. 5). Most of the experiments were terminated after about 4-5 hours, 282
but some showed stable firing properties for as long as 9 hours (Fig. 5A). For 11 V2a 283
interneurons, the AP properties were plotted over time. Means of 60 minute bins around 1, 2, 3 284
and 4 hours illustrate the stability of the AP properties (Fig. 5B-D). The AP amplitude remained 285
stable over many hours, averaging 72.2 ± 7.2 mV after 1 hour and 72.6 ± 3.6 mV after 4 hours 286
(Fig. 5B). The AP half-width also remained very stable, with virtually no change over time 287
(1.1 ± 0.3 ms after 1 hour to 1.0 ± 0.3 ms after 4 hours; Fig. 5C). Another extremely stable 288
parameter was the AP threshold, remaining around -36 mV for the duration of the recordings (-289
36.1 ± 4.5 mV after 1 hour; -36.2 ± 6.3 mV after 4 hours; Fig. 5D). 290
291
12
Discussion 292
A major current focus of our laboratory is to better understand how transection of the spinal cord 293
alters the intrinsic properties of neurons below the lesion site. In pursuing this goal, we found it 294
necessary to improve upon the procedures used to record intracellularly from adult spinal 295
interneurons in vitro. While the commonly used procedures enable reliable recordings from 296
neonatal tissue with relative ease (Carlin et al. 2006; Dougherty and Kiehn 2010; Gosgnach et 297
al. 2006; Lanuza et al. 2004; Wilson et al. 2005; Zhong et al. 2010), obtaining similarly stable 298
recordings in mature tissue (i.e. tissue from mice old enough to be developmentally mature and 299
with enough time post-injury for homeostatic changes in neuronal properties to occur) proved 300
problematic using the standard neonatal dissection and recording methods. By modifying the 301
neonatal approach, we have successfully overcome the “age barrier” in intracellular recordings 302
from spinal interneurons in vitro. While we made many changes to the procedure (outlined in 303
detail in Methods), there were two major changes that made the biggest impact: a dorsal 304
approach to dissecting the spinal cord and using perforated patch recordings. 305
In switching to a dorsal approach to the spinal cord dissection as previously described by 306
Carp et al. (2008) for spinal rat motoneurons, we were able to maintain normal blood flow to the 307
cord for as long as possible before it was finally removed from the animal. Ventral dissections, 308
which require initially killing the animal followed by evisceration prior to extracting the spinal 309
cord, induce hypoxia in the spinal cord much earlier in the dissection. Hypoxia is particularly 310
problematic in adult tissue where neurons appear to be less capable of recovering than similar 311
tissue from neonates (Haddad and Donnelly 1990). Another advantage of the dorsal approach 312
is that ice cold glycerol based GACSF (Ye et al. 2006) can be bathed over the exposed spinal 313
cord while it is intact and receiving normal blood flow. Cooling the tissue provides a 314
neuroprotective effect by reducing the rate of metabolism and slowing the effects of hypoxia 315
13
associated with the remainder of the dissection (Erecinska et al. 2003). Thus, this step 316
optimizes the capacity of the tissue to recover after the slicing process. 317
Even with healthier tissue slices, we found that our whole cell recordings were not as 318
stable as we routinely achieved in neonatal tissue, and obtaining adequate seals was also more 319
difficult. It is likely that spinal interneurons from older mice are more susceptible to dialysis of 320
the intracellular milieu that occurs with whole cell recordings. In particular, we suspect that 321
rundown in intracellular ATP levels leads to activation of the ATP-dependent KATP potassium 322
current, hyperpolarizing the neuron and dramatically altering its intrinsic properties and overall 323
health of the neurons (Seino and Miki 2003). This contributes to the poor health of adult spinal 324
cord neurons, but, since not all neurons hyperpolarized during the ten minutes after 325
breakthrough of the patch, additional consequences of intracellular dialysis must also occur. To 326
overcome this sensitivity to intracellular dialysis, we switched to perforated patch recordings 327
which minimize the exchange of solutions between the pipette and the recorded neuron (Horn 328
and Marty 1988). Using Amphotericin B as our ionophore, we were able to obtain long lasting 329
recordings during which the intrinsic properties of the recorded neurons remained remarkably 330
stable. Indeed, the recordings were most often terminated not because of a decline in the health 331
of the cell but at the discretion of the experimenter. 332
Together, the dorsal dissection combined with the use of perforated patch recordings 333
(and other minor methodological changes; see Methods for complete details) enable long 334
lasting and highly stable recordings from spinal V2a interneurons of developmentally mature 335
mice (as old as 93 days). The ability to perform these recordings is an important step for future 336
studies to understand how spinal networks generate particular behaviors (such as locomotion) 337
and the progression of disease/injury states which can only be investigated in relatively old 338
preparations. In addition, because the recordings are extremely stable, it is possible to discern 339
even subtle effects of neuromodulators on the intrinsic properties of neurons, which were 340
previously difficult to measure due to the rundown of various intrinsic currents in whole cell 341
14
recordings. We have not tested our procedure on other interneurons or motoneurons, but in our 342
slices these cells also appear healthy and should be amenable to perforated patch recordings. 343
We believe that the ability to record from mature spinal interneurons provides an additional and 344
valuable tool in understanding fundamental questions of how the nervous system generates 345
behaviors and studying the neuronal consequences of disease or injury. 346
347
15
Acknowledgements 348
We thank Dr. Kamal Sharma at the University of Chicago for kindly providing 349
Chx10::CFP mouse line and Drs. Shelby Dietz and Bruce Johnson for valuable 350
comments on earlier versions of the manuscript. We also thank V. Patel and C. Benton 351
for outstanding technical assistance. This work was supported by NIH grant NS057599 352
to RHW and DFG grant HU 1963/1-1 to AH. 353
354
355
16
References 356
Carlin KP, Dai Y, and Jordan LM. Cholinergic and serotonergic excitation of ascending 357
commissural neurons in the thoraco-lumbar spinal cord of the neonatal mouse. J Neurophysiol 358
95: 1278-1284, 2006. 359
Carp JS, Tennissen AM, Mongeluzi DL, Dudek CJ, Chen XY, and Wolpaw JR. An in vitro 360
protocol for recording from spinal motoneurons of adult rats. J Neurophysiol 100: 474-481, 361
2008. 362
Clarac F, Brocard F, and Vinay L. The maturation of locomotor networks. Prog Brain Res 143: 363
57-66, 2004. 364
Dougherty KJ, and Kiehn O. Firing and cellular properties of V2a interneurons in the rodent 365
spinal cord. J Neurosci 30: 24-37, 2010. 366
Edwards FA, Konnerth A, Sakmann B, and Takahashi T. A thin slice preparation for patch 367
clamp recordings from neurones of the mammalian central nervous system. Pflugers Arch 414: 368
600-612, 1989. 369
Erecinska M, Thoresen M, and Silver IA. Effects of hypothermia on energy metabolism in 370
Mammalian central nervous system. J Cereb Blood Flow Metab 23: 513-530, 2003. 371
Gosgnach S, Lanuza GM, Butt SJ, Saueressig H, Zhang Y, Velasquez T, Riethmacher D, 372
Callaway EM, Kiehn O, and Goulding M. V1 spinal neurons regulate the speed of vertebrate 373
locomotor outputs. Nature 440: 215-219, 2006. 374
Grillner S, and Jessell TM. Measured motion: searching for simplicity in spinal locomotor 375
networks. Curr Opin Neurobiol 19: 572-586, 2009. 376
Haddad GG, and Donnelly DF. O2 deprivation induces a major depolarization in brain stem 377
neurons in the adult but not in the neonatal rat. J Physiol 429: 411-428, 1990. 378
17
Hamill OP, Marty A, Neher E, Sakmann B, and Sigworth FJ. Improved patch-clamp 379
techniques for high-resolution current recording from cells and cell-free membrane patches. 380
Pflugers Arch 391: 85-100, 1981. 381
Herrington J, Solaro CR, Neely A, and Lingle CJ. The suppression of Ca2+- and voltage-382
dependent outward K+ current during mAChR activation in rat adrenal chromaffin cells. J Physiol 383
485 ( Pt 2): 297-318, 1995. 384
Horn R, and Marty A. Muscarinic activation of ionic currents measured by a new whole-cell 385
recording method. J Gen Physiol 92: 145-159, 1988. 386
Jiang Z, Rempel J, Li J, Sawchuk MA, Carlin KP, and Brownstone RM. Development of L-387
type calcium channels and a nifedipine-sensitive motor activity in the postnatal mouse spinal 388
cord. Eur J Neurosci 11: 3481-3487, 1999. 389
Kiehn O. Locomotor circuits in the mammalian spinal cord. Annu Rev Neurosci 29: 279-306, 390
2006. 391
Kiehn O, and Butt SJ. Physiological, anatomical and genetic identification of CPG neurons in 392
the developing mammalian spinal cord. Prog Neurobiol 70: 347-361, 2003. 393
Kjaerulff O, and Kiehn O. Distribution of networks generating and coordinating locomotor 394
activity in the neonatal rat spinal cord in vitro: a lesion study. J Neurosci 16: 5777-5794, 1996. 395
Koppe G, Bruckner G, Brauer K, Hartig W, and Bigl V. Developmental patterns of 396
proteoglycan-containing extracellular matrix in perineuronal nets and neuropil of the postnatal 397
rat brain. Cell Tissue Res 288: 33-41, 1997. 398
Lanuza GM, Gosgnach S, Pierani A, Jessell TM, and Goulding M. Genetic identification of 399
spinal interneurons that coordinate left-right locomotor activity necessary for walking 400
movements. Neuron 42: 375-386, 2004. 401
Lovell PV, and McCobb DP. Pituitary control of BK potassium channel function and intrinsic 402
firing properties of adrenal chromaffin cells. J Neurosci 21: 3429-3442, 2001. 403
18
Milev P, Maurel P, Chiba A, Mevissen M, Popp S, Yamaguchi Y, Margolis RK, and 404
Margolis RU. Differential regulation of expression of hyaluronan-binding proteoglycans in 405
developing brain: aggrecan, versican, neurocan, and brevican. Biochem Biophys Res Commun 406
247: 207-212, 1998. 407
Morales E, Fernandez FR, Sinclair S, Molineux ML, Mehaffey WH, and Turner RW. 408
Releasing the peri-neuronal net to patch-clamp neurons in adult CNS. Pflugers Arch 448: 248-409
258, 2004. 410
Pearson KG. Common principles of motor control in vertebrates and invertebrates. Annu Rev 411
Neurosci 16: 265-297, 1993. 412
Rae J, Cooper K, Gates P, and Watsky M. Low access resistance perforated patch recordings 413
using amphotericin B. J Neurosci Methods 37: 15-26, 1991. 414
Seino S, and Miki T. Physiological and pathophysiological roles of ATP-sensitive K+ channels. 415
Prog Biophys Mol Biol 81: 133-176, 2003. 416
Song ZM, Hu J, Rudy B, and Redman SJ. Developmental changes in the expression of 417
calbindin and potassium-channel subunits Kv3.1b and Kv3.2 in mouse Renshaw cells. 418
Neuroscience 139: 531-538, 2006. 419
Wilson JM, Blagovechtchenski E, and Brownstone RM. Genetically defined inhibitory 420
neurons in the mouse spinal cord dorsal horn: a possible source of rhythmic inhibition of 421
motoneurons during fictive locomotion. J Neurosci 30: 1137-1148, 2010. 422
Wilson JM, Hartley R, Maxwell DJ, Todd AJ, Lieberam I, Kaltschmidt JA, Yoshida Y, 423
Jessell TM, and Brownstone RM. Conditional rhythmicity of ventral spinal interneurons 424
defined by expression of the Hb9 homeodomain protein. J Neurosci 25: 5710-5719, 2005. 425
Wilson RJ, Chersa T, and Whelan PJ. Tissue PO2 and the effects of hypoxia on the 426
generation of locomotor-like activity in the in vitro spinal cord of the neonatal mouse. 427
Neuroscience 117: 183-196, 2003. 428
19
Ye JH, Zhang J, Xiao C, and Kong JQ. Patch-clamp studies in the CNS illustrate a simple new 429
method for obtaining viable neurons in rat brain slices: glycerol replacement of NaCl protects 430
CNS neurons. J Neurosci Methods 158: 251-259, 2006. 431
Zhang Y, Narayan S, Geiman E, Lanuza GM, Velasquez T, Shanks B, Akay T, Dyck J, 432
Pearson K, Gosgnach S, Fan CM, and Goulding M. V3 spinal neurons establish a robust and 433
balanced locomotor rhythm during walking. Neuron 60: 84-96, 2008. 434
Zhong G, Diaz-Rios M, and Harris-Warrick RM. Intrinsic and functional differences among 435
commissural interneurons during fictive locomotion and serotonergic modulation in the neonatal 436
mouse. J Neurosci 26: 6509-6517, 2006. 437
Zhong G, Droho S, Crone SA, Dietz S, Kwan AC, Webb WW, Sharma K, and Harris-438
Warrick RM. Electrophysiological characterization of V2a interneurons and their locomotor-439
related activity in the neonatal mouse spinal cord. J Neurosci 30: 170-182, 2010. 440
Zhong G, Masino MA, and Harris-Warrick RM. Persistent sodium currents participate in fictive 441
locomotion generation in neonatal mouse spinal cord. J Neurosci 27: 4507-4518, 2007. 442
443
444
445
20
Figure Legends 446
Fig. 1. Adult mouse spinal cord dissection. A: Overview of the deeply anaesthetized mouse 447
receiving pure oxygen by a mask. The spinal column was exposed from the dorsal surface by 448
removal of the overlying fat and muscle. B: Detailed view of the spinal column after removing 449
skin and muscles covering it. C: Exposed intact spinal cord after performing the partial dorsal 450
laminectomy. D: After removing the spinal cord including the lumbar enlargement, the cord was 451
pinned in an ice cold Sylgard Petri dish filled with glycerol based GACSF. The meninges were 452
removed by pulling with two forceps towards the posterior end of the cord. 453
454
Fig. 2. Perforated patch recording (PPR) on a spinal V2a interneuron of a seven week old 455
mouse. A: Top trace: current clamp recording (53 minutes duration) showing the process of 456
perforated patch formation after sealing onto a V2a interneuron. See Results for description. B: 457
Time course of the perforation process. The data points show the time for individual neurons to 458
reach stable membrane potential (VM), and overshooting action potentials (APAmp). C: Duration 459
of recordings is much longer using the new PPR method than the standard whole cell recording 460
(WCR) method. Most PPRs were voluntarily terminated after 4-5 hours, so the average 461
recording duration is artificially shortened. 462
463
Fig. 3. Recording quality is superior using the new perforated patch recording (PPR) method. A: 464
The mean membrane potential in standard whole cell recordings (WCRs) was more depolarized 465
with a higher variability than the membrane potential with the newly introduced perforated patch 466
recordings (PPRs). B: The holding current Ihold to maintain the neuron at -60 mV was 467
significantly more negative and variable in standard WCRs than using the new PPR. C: The 468
membrane input resistance with the standard approach whole cell recordings was significantly 469
lower compared to the perforated patch recordings. D: The rheobase was significantly larger 470
21
and more variable using the standard approach than the new approach. E: V2a neuron 471
recording using the standard WCR approach. During the first 10 minutes, the spontaneous firing 472
rate decreased and the membrane potential hyperpolarized. F: V2a neuron recording using the 473
new PP approach. Firing frequency was typically very stable. For each panel, the lower 474
recordings are made at the points of the stars in the upper recordings. G: Using the standard 475
WCR approach, the firing frequency dropped significantly after 10 minutes, and most neurons 476
fell silent. H: Using the new PP approach, the firing rate did not change significantly after 10 477
minutes. 478
479
Fig. 4. Recording stability is superior using the new PP method. A-D: Using the new PP 480
approach, the AP firing rate (A), IHold at -60 mV (B), input resistance (C) and rheobase values 481
(D) remained very stable for at least 4.5 hours. Symbols represent 11 individual interneurons 482
per plot, with 3 to 4 data points measured from each neuron during the first 4.5 hours of 483
recording. The average ± standard deviation graphs represent 60 minute bins. 484
485
Fig. 5. Stability of action potential properties. A: Examples of spontaneous action potentials 486
(APs) in a single V2a interneuron at different recording times between 0.5 and 9 hours after 487
recording began. Insets show averaged AP waveforms (averaged by aligning peaks from 10-12 488
APs) at a higher sweep speed. B-D: AP amplitude, half-width and threshold remained very 489
stable for at least 4.5 hours. Symbols represent 11 individual interneurons per plot, with 3 to 4 490
data points measured during the first 4.5 hours of recording. The average ± standard deviation 491
graphs represent 60 minute bins. 492
Figure 1: Adult mouse spinal cord dissection
C
B
A
D
oxygen mask
spinal column
fat
muscle
spinal cordmeninges
1.5 cm
5 mm
Figure 2: Perforated patch recordings in adult mouse spinal INs
5 min
2 s
0 mV
20 mV
-60 mV
50’
1 s
1’ 8’ 12’ 35’
1’
8’
12’
35’
50’
88’ 177’120’ 210’
A
B CPerforation progress
Stable VM APAmp
0
20
40
60
Tim
e af
ter s
eal (
min
)
Recording duration
PPR WCR0
2
4
6
8
10
12
Rec
. dur
atio
n (h
)
Figure 3: Recording quality
A B
C WCR PPR WCR PPR
WCR PPR WCR PPR
20 mV
1 s
1 min
WCR PPR
** **
E F
D
Firing WCR
0
2
4
6
8G
AP
firin
g (H
z)
10 minInitial
Firing PPR
0
2
4
6
8
AP
firin
g (H
z)
10 minInitial
H
0
1
2
3
Rin
put
-80
-60
-40
-20
0
VM(m
V)-150
-100
-50
0
50
I hold (p
A)
0
50
100
150
I rheo
base
(pA
)
* **
** *
**
Figure 4: Perforated patch recording stability
2.5
2.0
1.5
1.0
0.5
0.04321
Recording time (h)
-50-40-30-20-10
0
4321Recording time (h)
16
12
8
4
04321
Recording time (h)
6
4
2
0AP
firin
g (H
z)
Recording time (h)
A
C
B
D
4321
Rin
put
I hold (p
A)
I rheo
base
(pA
)
Figure 5: AP stability over recording time
2 ms
90
80
70
60
50AP
ampl
itude
(mV
)
4321Recording time (h)
1.6
1.2
0.8
0.4
0.0
AP
wid
th (m
s)
-50
-40
-30
-20
AP
thre
shol
d (m
V)
4321Recording time (h)
4321Recording time (h)
B C D
0.5 h 3 h
6 h 9 h
20 mV
500 ms
A