Neuroscience, Physiology and Pharmacology


Neuron-glial interactions and brain energy supply

Professor David Attwell, FRS

Jodrell Professor of Physiology
Tel: +44 (0)20 7679 7342
Email: d.attwell@ucl.ac.uk

Lab Members:

Lorena Arancibia-Carcamo

Narges Bazargani

Elisabeth Engl

Marc Ford

Nicola Hamilton-Whitaker

Matt Hammond-Haley

Renaud Jolivet

Anna Krasnow

Vasiliki Kyrargyri

Christian Madry

Anusha Mishra

Ross Nortley

Fergus O'Farrell

Work in the Attwell lab is funded by a Senior Investigator Award from the Wellcome Trust, and a grant from the Rosetrees Trust.
David Attwell

David Attwell studied physics as an undergraduate in Oxford, and then did a PhD in neuroscience with Julian Jack. After a post-doc in Berkeley with Frank Werblin, he came to UCL.


In my lab we are interested in signalling between neurons and glial cells (oligodendrocytes, astrocytes and microglia), and in how the brain’s energy supply is controlled and determines the computational power of the brain. Recent work on neural-glial signalling has focussed on how activation of glutamate and GABA receptors on oligodendrocytes may be responsible for the mental and physical disability occurring in white matter diseases such as cerebral palsy and spinal cord injury. Our studies of brain energy supply have characterized a new locus (in capillaries) for control of cerebral blood flow, have investigated how the amount of energy the brain receives determines the speed of the neural computations performed, and have studied the consequences of failure of the energy supply in conditions like stroke. We are also interested in how synaptic signalling to neurons is modulated. This work is performed using patch-clamp, calcium imaging and oxygen electrode techniques applied to brain slices and isolated cells, and using mathematical modelling.


Oligodendrocytes myelinate axons, and thus speed action potentials, but in pathological conditions like cerebral palsy, stroke, spinal cord injury and multiple sclerosis oligodendrocytes are killed, leading to mental and physical handicap. As for the neuronal death that occurs in stroke (see below), part of this oligodendrocyte death occurs as a result of glutamate being released by reversal of uptake transporters. We are using patch-clamping (Fig. 1) and immunocytochemistry to study the glutamate-evoked death of oligodendrocytes. Recent findings include the discovery of NMDA receptors on oligodendrocytes (Káradóttir et al., 2005) which play a role in damaging these cells in pathological conditions (Bakiri et al., 2007). Future work will investigate the physiological function of these receptors in controlling the development of oligodendrocytes.

Figure 1
Two oligodendrocytes, after whole-cell patch-clamping with different colour dyes in the pipette Káradóttir et al., 2005

Oligodendrocyte precursor cells

Oligodendrocyte precursor glia transform into myelinating oligodendrocytes during development, but are also present in the adult CNS where they comprise ~5% of the cells and are the main proliferating cell type. Damage to oligodendrocyte precursors, leading to reduced myelination, contributes to mental and physical impairment in periventricular leukomalacia (pre- or perinatal white matter injury leading to cerebral palsy). Adult OPCs may form new myelinating oligodendrocytes in multiple sclerosis and spinal cord injury, and OPC transplants could serve as a basis for therapeutic remyelination. However, the functions of OPCs are poorly understood: they may simply become oligodendrocytes in normal development and constitute a reservoir of cells which replace damaged myelin in the adult CNS, but they might also differentiate into astrocytes and neurons and thus have some stem cell characteristics. Understanding the diversity of NG2-expressing OPCs is crucial for understanding normal brain function, for appreciating the diversity of the brain’s progenitor cell population, and for developing therapeutic strategies to treat demyelinating diseases. We have recently shown that OPCs fall into two classes with different electrophysiological properties, and a different susceptibility to death in pathological conditions (Káradóttir et al., 2008). Strikingly, for glial cells, one of these classes fires action potentials (Fig. 2).

Figure 2
Myelin ensheathing axons entering the cerebellum, with inset of action potentials recorded from an OPC (Káradóttir et al., 2008).

Coupling of neuronal activity to blood flow

When neurons are active they require more energy to pump out ions that enter to produce synaptic and action potentials. Active neurons signal to the vasculature to increase the blood flow. It used to be thought that all of the regulation of flow occurred at the level of precapillary arterioles. We showed, however, that contractile cells called pericytes on capillaries (Fig. 3) also contribute to regulation of blood flow (Peppiatt et al., 2006). Interestingly, contractile signals can propagate from one pericyte to another, possibly spreading back to upstream arterioles.

Figure 3
(a) Schematic of blood flow control by arterioles and by pericytes. (b) Pericytes (red) on cerebellar blood vessels (green). From Peppiatt et al. (2006)

Constraints on information processing imposed by brain energetics

The brain is 2% of the body's mass but uses 20% of its energy. Consequently it is likely that energy use has constrained the brain's evolution, helping to determine the wiring pattern, signal coding and synapse properties of the neurons. Understanding the brain's energy use is important both for understanding how the brain has evolved and for understanding functional imaging signals, which detect the mismatch between oxygen supply and oxygen use in small volumes of brain tissue. Theoretical calculations suggest that much of the brain's energy is employed in reversing the ion movements producing synaptic currents and action potentials (Attwell & Laughlin, 2001). Consequently energy constraints are predicted to lead to minimization of axon length, signalling which shows adaptation (so that not so many action potentials need to be transmitted), minimization of postsynaptic currents, and distributed coding. Current work is investigating how the properties of synapses are optimised to process information in the face of the high energy demand of neural circuits (Attwell and Gibb, 2005), and analysing the allocation of brain energy use to conscious and unconscious information processing (Schölvinck et al., 2008).

Neurotransmitter transporters and stroke

We have used patch-clamp techniques to study transporters for the main excitatory neurotransmitter glutamate (Brew & Attwell, 1987), investigating in particular the ion movements which drive glutamate transport (Barbour, Brew & Attwell, 1988, Bouvier et al., 1992; Levy et al., 1998). We have demonstrated (Fig. 4) that glutamate transporters can run backwards when ion gradients run down in conditions like stroke (Szatkowski, Barbour & Attwell, 1990; Attwell et al., 1993), releasing enough glutamate to activate receptors in nearby neurons (Billups & Attwell, 1996). Indeed, in the first few minutes of a stroke, reversed operation of glutamate transporters is the main mechanism by which the extracellular glutamate concentration of glutamate is raised to levels which trigger the death of neurons, leading to mental and physical handicap (Rossi, Oshima & Attwell, 2000; Szatkowski & Attwell, 1994). A similar reversal of GABA transporters leads to GABA release early in stroke (Allen, Rossi & Attwell, 2004). Current work is focussed on the factors controlling glutamate dynamics early in a stroke, and the downstream consequences of glutamate and GABA release.

Figure 4
Using a detector neuron to sense glutamate release from a glial cell. Billups & Attwell (1996)

Neuromodulation: excitatory neurotransmission

Modulation of synaptic transmission plays a key role in adapting the behaviour of the nervous system to prevailing conditions. We are interested in how conventional neurotransmitters, unusual intercellular messengers like arachidonic acid, endocannabinoids and intracellular interacting proteins, can alter the behaviour of ion channels and transporters (Barbour, Szatkowski, Ingledew & Attwell, 1989; Miller, Sarantis, Traynelis & Attwell, 1992; Marie et al., 2002; Marcaggi, P. & Attwell, D. 2005. A major focus of our work is on how glutamate uptake controls synaptic transmission (Sarantis et al., 1993; Takahashi, Sarantis & Attwell, 1996; Auger & Attwell, 2000), and on the input-output relationship of excitatory synapses (Attwell et al., 1987; Sarantis, Everett & Attwell, 1988).

Neuromodulation: inhibitory neurotransmission

One focus of our work on inhibitory synaptic transmission is to understand how spillover of neurotransmitter GABA at the Golgi cell to granule cell synapse in the cerebellum affects the information processing carried out by the cerebellum (Fig. 5: Rossi & Hamann, 1998; Hamann, Rossi & Attwell, 2002). Surprisingly, at this synapse most of the synaptic current is produced by GABA acting on cells which are not anatomically postsynaptic to the releasing cell. This depends on the granule cells expressing very high affinity GABAa receptors (containing alpha6 subunits) in non-synaptic locations, and the amount of GABA reaching the receptors is controlled by GABA transporters. This extrasynaptic inhibition mediates more than 97% of the inhibition received by granule cells, and plays an important role in determining how the cerebellum stores motor commands. Other work on inhibitory transmission is aimed at understanding how inhibition of retinal bipolar cells is modulated (Billups et al., 2000).

Figure 5
A cerebellar Purkinje cell filled with dye, during studies of cerebellar information processing. Hamann, Rossi & Attwell (2002)


We currently have active collaborations with the labs of:

Selected Publications