031119-Dynamic balance of metabotropic inputs causes dorsal horn neurons to switch functional states by tangshuming


									  Dynamic balance of metabotropic
inputs causes dorsal horn neurons to
       switch functional states
   Dominique Derjean1, 3, Sandrine Bertrand1, 3, Gwendal Le
   Masson1, Marc Landry1, Valérie Morisset2 & Frédéric Nagy1

           Nature Neuroscience 6, 274 - 281 (2003)

•   The dorsal horn of the spinal
    cord is the first central relay for
    inputs from primary nocieptor
    sensory fibers. Relay deep
    dorsal horn neurons (DHNs)
    integrate both innocuous and
    nociceptive inputs and are of
    fundamental importance for
    plasticity in information
    processing and transfer.
•   Individual neurons have characteristic and
    different intrinsic membrane properties
    that affect how they respond to same sensory
•   “Intrinsic membrane properties” refer to a
    neuron’s electrical properties in isolation, in the
    absence of synaptic inputs.
•   In normal state, the deep DHNs show three kinds
    of intrinsic membrane properties:
    (a) Tonically firing neurons
    (b) Plateau neurons
    (c) Oscillatory neurons
•    Intrinsic membrane properties are not immutable.
    Rather, neuromodulatory substances released
    from neighboring neurons that act via second
    messenger systems often alter one or more of the
    voltage and time-dependant currents in a neuron
    and change a neuron’s intrinsic membrane
    properties for seconds, minutes or hours.
• Derjean et al. in this issue show that the spinal
  cord neurons that receive information from the
  sensory neurons activated by painful stimuli are
  not simple faithful followers of the train of action
  potentials fired by the sensory neurons. Instead,
  the spinal cord neurons display complex
  membrane properties that transform the sensory
  signal. As a consequence, short-term
  modulation of these membrane properties could
  contribute to modifications of pain sensitivity.
                               Three firing modes: a balance of
1. Activation of group-I
glutamate receptors
(mGluRs) by the
                                   metabotropic controls
agonists ACPD .(Fig1b)

2. applied the mGluR1
antagonist 4-CPG on
those neurons with
spontaneous plateau
properties. (Fig1g)

*This indicates  that the
active properties of the
deep DHNs are under a
sustained basal
modulatory control.

3. activating GABAB
receptors with the
agonist baclofen .(Fig1C)

4. superfusion of the
GABAB antagonist
CGP55845 .(Fig1E)
*This shows that the
metabotropic GABAB
system exerts a tonic
inhibitory control over the
active properties of deep
Evidently, it is the balance
of permissive and
suppressive modulatory
systems that controls the
active properties of deep
•   To test whether these antagonistic modulatory pathways share a common cellular
    mediator, Authors targeted a family of channels linked to a variety of metabotropic
    receptors: the G-protein dependent potassium channel Kir3. Membrane currents
    modified by bath-application of GABAB receptors or mGluR agonists were analyzed
    using a voltage ramp and a subtraction procedure and on synaptically isolated DHNs.

• Current–voltage curves were determined using
  voltage ramps from -55mV to -155 mV,
             Antagonistic regulation of a Kir3 current

                                                                                      GABAB receptors agonists
                                                                                                                        mGluR agonists

                                                                                                                                       Erev, -99.4 ± 2.2 mV
                                       Outward current
Erev, -95.3 ± 2.2 mV

K+ equilibrium potential (-95.8 mV)

                                                                                         Inward current

The chord conductance was                                                                                                             The chord
measured at two different                                                                                                             conductance of the
potentials that were equidistant                                                                                                      DHPG-suppressed
from Erev (Fig. 2b).                                                                                                                  currents was higher
                                                                                                                                      for hyperpolarized
                                                                                                                                      potentials (Fig. 2h)
In all cells tested, the conductance
was significantly higher for the
most negative potential (Fig. 2d).

                                                                                                                   the suppressive effect of DHPG was
The effect of baclofen was abolished in a voltage-independent manner in the presence of Ba2+ (Fig. 2c and d) and   blocked by the superfusion of Ba2+ at low
in a voltage-dependent manner in the presence of Cs+ (Fig. 2d), which is characteristic of Kir currents.           concentrations (Fig. 2g–h).
• Together, these results indicate that
  activation of GABAB and mGlu receptors
  exert antagonistic regulation—
  enhancement and inhibition,
  respectively—on a Kir current in deep
   The morphological evidence showing that a single DHN
    received both glutamatergic and GABAergic synaptic
            contacts and expressed Kir channel

Triple immunostaining showed a dendrite of a plateau-
generating DHN injected with biocytin (blue) receives
both VgluT-a marker of GABAergic fibers - (green,
open arrowhead) and GAD65-a marker of
glutamatergic fiber - (red, filled arrowhead) containing

Kir3.1 immunoreactivity reveals that Kir3 channels
(green, arrows) are expressed in the soma of a biocytin-
injected DHN (blue)
• These observations suggest that the Kir
  current that is described in
  electrophysiological experiments and
  modulated by both mGlu and GABAB
  receptors is of the Kir3 family.
• Finally, the author addressed the question of whether the
  three firing modes of deep DHNs correspond to different
  properties of sensory information transfer and thus may
  define different functional states. The author measured
  the input–output relationships between primary afferent
  spiking and the output firing of a single DHN with tonic,
  plateau or oscillatory firing patterns. To precisely control
  the afferent spiking and avoid complex polysynaptic
  effects associated with dorsal root stimulation, the author
  designed a simple canonical circuit using the hybrid
  network method. That is, a computer model of primary
  nociceptor discharge was connected to an intracellularly
  recorded DHN through an artificial excitatory synapse.
                              State-dependent capabilities of
                                    information transfer
The modeled nociceptive fiber discharge produced
slowly adapting responses to depolarizing stimuli
(frequency range, 4–32 Hz).
                                                         The corresponding cross-correlogram (Fig. 4d)
Every spike triggered an artificial synaptic current     yielded a broad peak of small amplitude,
injected through the recording pipette (Fig. 4a),        indicating that the recorded DHN was not
mimicking realistic AMPA excitatory postsynaptic         responding in a one-to-one manner to every
potentials (EPSPs) in the DHN (Fig. 4b).                 presynaptic spike.
                                                         Conversely, a plateau-generating neuron
The spike transfer was quantified using cross-
                                                         reacted to the same input with a higher
correlation analysis between the afferent                frequency, accelerating spike train (Fig. 4e),
discharge and the DHN response.                          and had a more precise correlation as shown
                                                         by the higher and narrower peak in the cross
                                                         correlogram (Fig. 4f).
                                                         The averaged CC (n = 6) was significantly
             a mean input firing                         higher (0.26) for plateau-generating neurons
                                                         (Fig. 4i) compared to tonic neurons (0.08). The
             frequency of 24 Hz                          lower CI for plateau neurons (0.1 versus 0.23 in
                                                         tonic mode)

                                                         These results suggest that expression of active
                                                         plateau properties in dorsal horn relay neurons
                                                         significantly increased the transmission of
                                                         afferent single spikes.

                                                         The responses obtained during spontaneous
                                                         oscillations and rhythmic bursting of DHNs (Fig.
                                                         4g) were characterized by both a low CC (Fig.
                                                         4h and i, 0.11) and a low CI (Fig. 4j, 0.07),
                                                         indicating that the oscillatory firing mode was
                                                         filtering out most of the afferent activity.

                                                         Analysis of the delay between the occurrence of
                                                         an input spike and the generation of a
                                                         correlated output spike (Fig. 4k) revealed a
                                                         shorter latency for the plateau firing mode
                                                         compared to the tonic mode. The mean delay in
                                                         the oscillatory mode was not significantly
                                                         different from that in the tonic mode, but did
                                                         show a higher variability (larger s.e.m.).
              Chord conductance?

• For type of channel at a single time, we can define the
  conductance through the channel as the inverse of the
  resistance - i.e. g=I/V. However, at a microscopic level
  channel behave in a stochastic and binary way, and so
  we cannot predict the behaviour of individual channels in
  this way. The chord conductance is a measure of the
  permeability of all of one type of channel for a
  particular cell, allowing us to predict the actual
  amount of current that will flow across the whole
• Chord conductance lets us describe the macroscopic
  behaviour of the channels, giving an indication of the
  permeability of a class of channels for a given voltage.
• Potassium is outward, sodium is inward and Cl- is inward.
• Then given that :
•     gK+ = 0.78 ìsiemens; gNa+ = 0.06 ìsiemens; gCl- = 0.25
• use the chord conductance equation, but need to determine
  equilibrium potentials for each ion using the Nernst Equation
• Ena = ~ +82.5 mV; Ek = ~ -70mV; Ecl = ~-48 mV
• Calculate the membrane potential, and Em = ~ -61mV
• B. Determine the necessary change in gNa+ if the recorded
  membrane potential is +34 mV. gNa = ~ 2.0 usiemens
•          (Assuming all other quantities remain constant.)
• C. What is the physiological importance for a membrane that can
  shift it conductance between the states given in A and B?
     cross-correlation analysis
• Two indexes were computed:
 (i) the correlation coefficient (CC), which indicates the ratio of input
   spikes that are transmitted as output spikes in the DHN neuron and
   thus characterizes the global efficacy of input–output spike transfer .
 (ii) the contribution index (CI), which quantifies the percentage of
   output spikes that were precisely correlated with afferent input
   spikes. It estimates the probability that a DHN spike was triggered
   by an input spike rather than generated spontaneously
                                        Input spike
                                                                  CC : 3/10 = 0.3

                                        Output spikes             CI : 3/5 = 0.6

           Output spikes were precisely correlated with afferent input spikes.

To top