Assessment of Long-term Depression Induction in Adult Cerebellar Slices

Summary

In some gene-manipulated animals, using a single protocol may fail to induce LTD in cerebellar Purkinje cells, and there may be a discrepancy between LTD and motor learning. Multiple protocols are necessary to assess LTD-induction in gene-manipulated animals. Standard protocols are shown.

Abstract

Synaptic plasticity provides a mechanism for learning and memory. For cerebellar motor learning, long-term depression (LTD) of synaptic transmissions from parallel fibers (PF) to Purkinje cells (PC) is considered the basis for motor learning, and deficiencies of both LTD and motor learning are observed in various gene-manipulated animals. Common motor learning sets, such as adaptation of the optokinetic reflex (OKR), the vestibular-ocular reflex (VOR), and rotarod test were used for evaluation of motor learning ability. However, results obtained from the GluA2-carboxy terminus modified knock-in mice demonstrated normal adaptation of the VOR and the OKR, despite lacking PF-LTD. In that report, induction of LTD was only attempted using one type of stimulation protocol at room temperature. Thus, conditions to induce cerebellar LTD were explored in the same knock-in mutants using various protocols at near physiological temperature. Finally, we found stimulation protocols, by which LTD could be induced in these gene-manipulated mice. In this study, a set of protocols are proposed to evaluate LTD-induction, which will more accurately allow examination of the causal relationship between LTD and motor learning. In conclusion, experimental conditions are crucial when evaluating LTD in gene-manipulated mice.

Introduction

The synaptic organization of the elaborated neuronal networks of the cerebellar cortex, composed of PCs, molecular layer interneurons (basket and stellate cells), Golgi cells, PFs from granule cells, mossy fibers and climbing fibers (CFs), have been elucidated in terms of excitation/inhibition and divergence/convergence, and the well-organized circuitry diagram has suggested that the cerebellum is a “neuronal machine”¹, though there was previously no idea about purpose of this “machine”. Later Marr proposed that the PFs input to PCs constitute a triple layer associative learning network². He also suggested that each CF conveys a cerebral instruction for elemental movement². He assumed that simultaneous activation of PFs and CF would enhance PF-PC synapse activity, and cause long-term potentiation (LTP) of the PF-PC synapse. On the other hand, Albus assumed that synchronous activation of PFs and CF resulted in LTD at the PF-PC synapses³. Both the above studies interpret the cerebellum as a unique memory device, the incorporation of which into the cerebellar cortical network leads to the formation of the Marr–Albus model learning machine model.

Following these theoretical predictions, two lines of evidence suggest the presence of synaptic plasticity in the cerebellum. The first line of evidence was suggested by the anatomical organization of the flocculus; here MF pathways of vestibular organ origin and CF pathways of retinal origin converge on the PCs⁴. This unique convergence pattern suggests that a synaptic plasticity occurring in the flocculus causes the remarkable adaptability of the vestibulo-ocular reflex. Second, the recording of the PCs response in the flocculus and the lesioning of the flocculus also supported the above hypothesis^5,6,7. Furthermore, the PC discharge pattern during adaptation of a monkey’s hand movement⁸ supported the synaptic plasticity hypothesis, especially Albus’s LTD-hypothesis³.

To determine the nature of the synaptic plasticity directly, repeated conjunctive stimulation (Cjs) of a bundle of PFs and the CF that specifically innervates the PC in vivo was shown to induce LTD for the transmission efficacy of the PF–PC synapses^9,10,11. In the subsequent in vitro explorations using a cerebellar slice¹² and cultured PCs, conjunction of co-cultured granule cell stimulation and olive cell stimulation¹³ or conjunction of iontophoretically applied glutamate and somatic depolarization^14,15 caused LTD. The signal transduction mechanism underlying the LTD-induction was also intensively investigated using in vitro preparations^16,17.

Adaptations of the VOR and the OKR were often used for quantitative evaluation of gene-manipulation effects on cerebellar motor learning, because the vestibule-cerebellar cortex was proven to be the essential origin in the adaptive learning of the VOR^18,19,20 and the OKR^19,21 The correlation between failure of LTD-induction and impairment of behavioral motor learning has been taken as evidence that LTD plays an essential role in motor learning mechanisms²². These views are collectively referred to as the LTD hypothesis of motor learning, or Marr-Albus-Ito hypothesis^23,24,25,26.

Adaptive learning of eye movement was measured using similar protocols, while various experimental conditions were used to induce LTD in slice preparation^{27,28,29,30,31}. Recently, Schonewille et al.²⁶ reported that some gene-manipulated mice demonstrated normal motor learning, but the cerebellar slices did not show LTD, and thereby concluded that LTD was not essential for motor learning. However, the induction of LTD was only attempted using one type of protocol at room temperature. Hence, we used several types of LTD-inducing protocols under recording conditions at around 30 °C, and we confirmed that the LTD was reliably induced in the gene-manipulated mice by using these protocols at near physiological temperatures³².

However, there remain some questions regarding the basic properties of conjunctive stimulation. The first is the relationship between the complex spike’s shape and the amplitude of LTD. Second, in conjunction with PF-stimulation and somatic depolarization, whether the number of stimuli used were necessary or not was elusive. In the present study, these questions were investigated using wild type (WT) mice.

Protocol

All experimental procedures were approved by the RIKEN committee on the care and use of animals in experiments. Mice were kept in the animal facility of the RIKEN Center for Brain Science under well-controlled temperature (23–25 °C) and humidity (45%–65%) conditions. Both male and female WT mice (C57BL/6, 3–6 months) were used.

1. Preparation of Solutions Used in the Experiments

NOTE: All solutions should be made in ultrapure water free of metals (resistivity > 18.2 MΩ) and other impurities (total organic carbon (TOC) < 5.0 ppb). Working artificial cerebrospinal fluid (ACSF) for slice-cutting and recording are made freshly on the day of experiment from a 10 times (x10) stock of ACSF. Bubble the solutions with 5% CO₂/ 95% O₂ gas mixture before use. The pH of ACSF is adjusted to 7.4 ± 0.1, and osmolarity is adjusted 315 ± 5 mOsm/kg by adding ultrapure water.

1. Prepare 10x stock of ACSF containing 1250 mM NaCl, 30 mM KCl, 12.5 mM NaH₂PO₄, and 260 mM NaHCO₃. This solution can be stored at 4 °C.
2. Prepare working ACSF containing 125 mM NaCl, 3 mM KCl, 2 mM CaCl₂, 1 mM Mg₂SO₄, 1.25 mM NaH₂PO₄, 26 mM NaHCO₃ and 20 mM glucose.
   1. First, add 1 mL of 2 M CaCl₂ solution followed by 1 mL of 1 M Mg₂SO₄ solution into around 800 mL of ultrapure water, to avoid precipitation. Then add 100 mL of 10x ACSF and glucose. Finally, make up to a total volume of 1,000 mL by adding ultrapure water.
3. Prepare 3.3% agar for brain handling. Dissolve 1 g of agar in 30 mL of 0.9% NaCl solution, and heat in a microwave until just boiling. Stir to mix, then pour it into a sterile 4 cm x 10 cm plastic box and allow to solidify. Store the agar plate (~8 mm thickness) in a refrigerator.
4. Prepare the internal solution.
   1. Prepare the K⁺-based internal solution containing 60 mM KCl, 60 mM K-gluconate, 0.3 mM EGTA, 4 mM MgCl₂, 4 mM ATP, 0.4 mM GTP and 30 mM HEPES (pH 7.2).
      NOTE: Low concentration (0.3 mM) of EGTA, a slow Ca²⁺-chelator, is added to chelate possibly contaminated Ca²⁺ in pure water, but this low concentration of EGTA in the internal solution never blocks the induction of LTD (Figure 3, Figure 4, Figure 5) during whole-cell recording. Measured osmotic pressure is 285 mOsm/kg.
   2. Prepare the Cs⁺-based internal solution containing 60 mM CsCl, 46 mM D-gluconate, 27 mM tetraethylammonium chloride (TEA-Cl), 0.3 mM EGTA, 4 mM MgCl₂, 4 mM ATP, 0.4 mM GTP and 30 mM HEPES (pH 7.2, adjusted using CsOH).
      NOTE: Cs⁺ blocks voltage-dependent K-channels, and improves space-clamp conditions at remote dendrites by increasing the length-constant. Measured osmotic pressure is 285 mOsm/kg.
   3. Prepare 200 µL aliquots of the solutions and store at -30 °C.

2. Brain Dissection and Trimming

1. Chill and oxygenate two 50 mL beakers of ACSF on ice until the temperature is lower than 4 °C. Add 50 µL of tetrodotoxin (TTX, 1 mM) into one of the ice-cold beakers of ACSF and reserve it for slices cutting. To obtain mouse cerebellar slice reserving LTD-inducing ability, the addition of TTX to the normal ACSF is necessary.
2. Cool down the metal specimen tray by filling an ice-bath area of the slicing chamber with ice.
3. Pour 1 mL of isoflurane into an anesthetizing jar (~1000 mL) then place a mouse in it for 30–45 s. Ensure that the mouse is deeply anesthetized by confirming its inability to respond to mechanical stimulation.
4. Decapitate the mouse using surgical scissors. Hold the head and cut the superficial skin along the midline using an ophthalmological scissor. Pull the skin by holding with fingers to widely expose the skull’s surface.
5. Cut the skull horizontally along a line from the major spinocerebellar hole just above the ear and eye using an ophthalmological scissor. Cut the skull along a line above both eyes and remove to isolate the skull.
6. Cut the brain at the middle of cerebrum by using a scalpel, then isolate the caudal part of the brain including the cerebellum from the skull. Immerse it into an ice-cold beaker of ACSF. Usually, the total time from decapitation to immersion of the brain block into the pre-chilled beaker of ACSF should be less than 60 s.
7. Position of bubbling tubing should be adjusted so as not to stir the brain block in the beaker. Mechanical damage might cause swelling of the slice during recording. Leave it for at least 7 min and allow the brain to cool down.
8. To trim the brain block, cut a rectangular agar piece (2 cm x 2 cm) from a large agar plate (4 cm x 10 cm, stored at 4 °C) and put it on a filter paper to absorb the excess liquid.
9. Turn the agar piece upside-down on the filter paper, then place the agar piece on a filter paper on a pre-chilled metal specimen tray (16 cm x 20 cm). Pick up the brain block using a spatula and absorb excessive liquid around it with a piece of filter paper.
10. Mount the brain block onto the agar block using glue (medical cyanoacrylate instant adhesive). Make sure to attach the bottom (ventral side) of the brain block to the agar.
11. Cut out the right hemisphere with a blade. Be sure that the side of the cutting plane is as parallel as possible to the dendritic plane of the PC because this side is attached to the surface of the specimen tray. Cut and remove the other side of the hemisphere. Then, cut the brain between the superior and inferior colliculi, and cut off the spinal cord.
12. Glue the right side of trimmed cerebellum with the agar block onto the pre-chilled specimen tray. Spread excess glue around the cerebellum with the flat part of a spatula, to prevent excess glue from attaching to the cerebellar surface. Tilt the metal tray and pour ACSF in order to fix the glue and wash away the excess glue.

3. Brain Slicing

1. Orient the sample such that the dorsal side of the cerebellum is on the front side. Pour ice-cold cutting ACSF, containing 1 µM TTX, sufficient to immerse the cerebellum completely. Place a gas tube into the ACSF and start bubbling with O₂/CO₂ gas mixture.
2. Remove the arachnoid mater using a fine tweezer under binoculars. Cut the cerebellar peduncle with a blade, and remove the brainstem and agar block. Rotate the tray 180°, so that the dorsal surface of the cerebellum faces a razorblade.
3. Set the blade, and adjust the first cutting location. Set the vibratome slicing parameters to the following: amplitude to 5.5, frequency to 85 Hz, speed to 3–4, and slice thickness to 300 µm.
4. Transfer the cerebellar slice on a nylon-net into an acrylic incubator and immerse the slice completely into the oxygenated ACSF. The incubator should be placed in a water-bath that maintains a temperature at 26 °C.
5. Store the slices for at least 1 h to allow recovery from the damage during slicing.

4. Whole-cell Patch-clamp Recording

NOTE: A patch-clamp recording requires following equipment: an upright microscope with infrared differential interference contrast (IR-DIC) optics, a patch-clamp amplifier, data digitizer, digital stimulator, isolator, computer, software for data-acquisition and analysis, motorized manipulator, microscope platform, vibration isolation table, Faraday cage, solution heating system, peristaltic pumps and electrode puller.

1. Add picrotoxin (0.1 mM) to ACSF and resolve it using ultra-sonication for 3 min.
2. Perfuse a recording chamber with picrotoxin-containing, O₂-CO₂-saturated ACSF at rate of 2 mL/min. Maintain the temperature of the recording chamber at around 30 °C.
3. Make a recording electrode by pulling a borosilicate glass capillary with filament (outer diameter = 1.5 mm) using a puller with 4 steps. The tip-diameter should be around 1 µm.
4. Make a stimulating electrode by pulling the same capillary using the puller with 2 steps, then break to produce a fine tip by striking the tip against an iron block under a binocular microscope. The final diameter should be 3–5 µm.
5. Transfer the cerebellar slice to the recording chamber and fix it with a Pt-weight with nylon threads. Fill a stimulating electrode with ACSF.
6. For stimulation of the PFs, place the stimulating electrode on the surface of the molecular layer, around 50 µm away from the Purkinje cell layer.
7. For stimulation of the CF, place the stimulating electrode at the bottom of the Purkinje cell layer (steps 5.3, 5.4).
8. Filter K⁺-based or Cs⁺-based internal solution with a 0.45 µm filter. Use a micro-loader to fill a recording electrode with 8 µL of internal solution.
9. Apply a weak positive pressure to the recording electrode before immersing it into the ACSF. Its resistance should be 2–4 MΩ and the liquid junctional potential should be corrected.
10. Approach the healthy, bright cell body of the PC with the recording electrode. Push the surface of Purkinje cell slightly, stop applying positive pressure, next apply negative pressure until forming a giga-ohm seal. Then establish the whole-cell configuration using negative pressure.
11. Hold the membrane potential at -70 mV, and apply -2 mV pulse (duration, 100 ms) at 0.1 Hz to monitor input resistance, series resistance and input capacitance, continuously. Do not use the series resistance compensation. Discard data when the series resistance varies by more than 15%.

5. Induction of LTD

1. Stimulate the molecular layer with a pulse (duration, 0.1 ms). Identify the PF-excitatory postsynaptic currents (EPSC) by applying a double pulse stimulus (interspike interval (ISI) of 50 ms). The PF-EPSC should show paired-pulse facilitation and gradual increase in amplitude relative to increase in stimulation intensity.
2. Record the test response of the PF-EPSC by applying a single pulse at 0.1 Hz. Adjust the intensity of the stimulus so that the evoked EPSC amplitude is around 200 pA. Avoid contamination of current through the voltage-dependent ionic channel.
3. Stimulate the CF at the bottom of the Purkinje cell layer, and identify the EPSC elicited by the CF activation (by applying a double pulse stimulus). The CF-EPSC should show paired-pulse depression and an all-or-none manner according to the increase in stimulation intensity. For LTD induction, a single stimulus should be used.
4. LTD-inducing protocol 1
   1. Using an electrode containing K⁺-based internal solution under current-clamp conditions, apply a single PF-stimulus and a single CF-stimulus simultaneously at 1 Hz for 5 min (300 pulses) (Figure 1A).
5. LTD-inducing protocol 2
   1. Using an electrode-containing K⁺-based internal solution under current-clamp conditions, apply double PF-stimuli (ISI of 50 ms) and single CF-stimulus as the second PF-stimulus is coincident with CF-stimuli at 1 Hz for 5 min (Figure 1B).
6. LTD-inducing protocol 3
   1. Using an electrode containing Cs⁺-based internal solution under voltage-clamp conditions, apply a double PF-stimulus (ISI of 50 ms) and a single depolarizing voltage-step (-70 to 0 mV, 50 ms) to the soma at 1 Hz for 3 min, so that the second PF-stimulus is equivalent to the beginning of the depolarizing voltage step (Figure 1C).
7. LTD-inducing protocol-4
   1. Using an electrode containing Cs⁺-based internal solution under voltage-clamp conditions, apply the PF-stimuli (5x at 100 Hz) and a single depolarizing voltage-step (-70 to 0 mV, 50 ms) to the soma at 0.5 Hz for 3 min, simultaneously (Figure 1D).

Results

Four protocols were used in this study to induce cerebellar LTD. In the first two protocols (protocol 1 and 2), the conjunction of the PF-stimulation and the CF-stimulation was applied under current-clamp conditions. In the other two protocols (protocol 3 and 4), somatic depolarization was substituted for the CF-stimulation under voltage-clamp conditions. Voltage-traces or current-traces during conjunctive stimulation were compared (Figure 2).

Conjunction of 1 PF-...

Discussion

Differences among the four protocols

In LTD-inducing protocols 1 and 2, Cjs 300 times at 1 Hz is sufficient to induce cerebellar LTD. Stimulation frequency of the CF seemed to be in a physiological range, because the complex spike firing rate in alert adult mice (P60) was reported to be 1.25 Hz³⁶. However, the CF stimulation alone did not cause long-term plasticity in the PF-CF synapse, as used in protocols 1 and 2 (Figure 4,

Materials

Name	Company	Catalog Number	Comments
Amplifier	Molecular Devices-Axon	Multiclamp 700B
Borosilicate glass capillary	Sutter	BF150-110-10
Digitizer	Molecular Devices-Axon	Digidata1322A
Electrode puller	Sutter	Model P-97
Isoflurane	FUJIFILM Wako Pure Chemical	26675-46-7
Isolator	A.M.P.I.	ISOflex
Linear slicer	Dosaka EM	PRO7N
Microscope	NIKON	Eclipse E600FN
Peristaltic pump	Gilson	MP1 Single Channel Pump
Picrotoxin	Sigma-Aldrich	P1675
Pure water maker	Merck-Millipore	MilliQ 7000
Software for experiment	Molecular probe-Axon	pClamp 10
Software for statistics	KyensLab	KyPlot 5.0
Stimulator	WPI	DS8000
Temperature controller	Warner	TC-324B
Tetrodotoxin	Tocris	1078