This Week in The Journal

Mitochondrial PKA Anchor Protects Neurons from Ischemia

Kyle H. Flippo, Aswini Gnanasekaran, Guy A. Perkins, Ahmad Ajmal, Ronald A. Merrill, et al.

Numerous signaling molecules act by increasing or decreasing cAMP levels, thereby modulating the activity of cAMP-dependent protein kinase A (PKA). But PKA exerts different downstream effects depending on where it is located and what substrates are nearby. These factors are regulated by A-kinase anchoring proteins (AKAPs). AKAP1, for example, tethers PKA to mitochondrial outer membranes, where it phosphorylates dynamin-related protein 1 (Drp1), a major mediator of mitochondrial fission.

Mitochondrial health requires periodic mixing of mitochondrial DNA and other constituents, and this is achieved through mitochondrial fusion. Subsequent fission is required to enable individual mitochondria to move to sites where they are needed. Fission also facilitates degradation of unhealthy mitochondria. Excessive fission is detrimental, however. AKAP1 prevents this by facilitating PKA-dependent phosphorylation of Drp1, preventing it from associating with mitochondrial membranes to initiate fission.

Flippo, Gnanasekaran et al. show that this function of AKAP1 helps protect neurons from ischemic damage. Drp1 phosphorylation was lower and its association with mitochondria was higher in forebrain from AKAP1-deficient than in wild-type mice. Furthermore, mitochondria were smaller—suggesting fission was increased—in AKAP1-deficient hippocampal neurons and astrocytes than in controls. In addition, the function of complex II of the mitochondrial electron transport chain was impaired in AKAP1-deficient neuronal cultures.

Importantly, ischemia produced by occlusion of the middle cerebral artery caused more extensive damage in AKAP1-deficient mice than in controls. Furthermore, glutamate—which contributes to neuronal death after ischemia—led to calcium deregulation more quickly and caused greater production of superoxide in AKAP1-deficient neuronal cultures than in controls. Notably, enhanced glutamate-induced excitotoxicity was rescued in AKAP1-deficient neurons by replacing wild-type Drp1 with a phospho-mimetic form, and it was replicated in wild-type neurons by expressing a non-phosphorylatable form of Drp1.

These results suggest that by enabling PKA-dependent phosphorylation of Drp1, AKAP1 limits mitochondrial fragmentation after ischemia, and thus limits neuronal death. This is consistent with previous, controversial work showing that inhibiting Drp1 reduced ischemic brain injury in rodents. Notably, AKAP1 is degraded under hypoxic conditions, possibly promoting mitochondrial fragmentation. Specific inhibitors of Drp1 might counteract this effect, thus limiting brain damage after stroke.

Acetylcholine, via SK Channels, Regulates CA3–CA1 Coupling

Crescent L. Combe, Carmen C. Canavier, and Sonia Gasparini

(see pages 8110–8127)

Cognitive functions require coordinated activity among neurons in multiple brain areas. This coordination can be seen in oscillations of the local field potential, which are produced by synchronous fluctuation in the activity of numerous neurons. Synchronization across areas is thought to facilitate communication between those areas, and oscillations at different frequencies are thought to underlie different functions. For example, as rats explore an arena, slow-gamma-range (25–50 Hz) oscillations in hippocampal areas CA1 and CA3 are coupled; this coupling has been proposed to promote memory retrieval. In contrast, synchronized oscillations in the fast-gamma range (65–100 Hz) are thought to facilitate communication between CA1 and the entorhinal cortex, possibly to promote memory encoding. Spiking in CA1 becomes phase-locked to fast- instead of slow-gamma oscillations in CA3 when a rat encounters a new object (Zheng et al., 2016 eNeuro 3:1), which might also contribute to encoding.

When CA3 afferents are stimulated at slow-gamma frequencies, most CA1 neurons spike at the same frequency (red in heat map). But when afferents are stimulated at fast-gamma frequency (marked by ellipses), CA1 neurons spike at half that frequency (left). This frequency selectivity is attenuated by cholinergic agonists (right). See Combe et al. for details.

The mechanisms underlying the generation and coupling of gamma oscillations in hippocampus remain unclear. Combe et al. show that the intrinsic properties of CA1 pyramidal cell dendrites have a role. Stimulating CA3 afferents to CA1 at slow-gamma frequencies evoked phase-locked, 1:1 spiking in CA1 pyramidal cells. In contrast, stimulating afferents at fast-gamma frequencies evoked spiking at approximately half the stimulation frequency. CA1 neurons could spike at high frequency when current was injected into the soma, however, suggesting dendritic properties limited the ability of high-frequency afferent input to generate 1:1 spikes.

A computational model suggested that activation of small-conductance calcium-activated potassium (SK) channels was responsible for the frequency selectivity of CA1 spiking. Consistent with this hypothesis, blocking SK channels in hippocampal slices attenuated frequency selectivity in CA1 neurons, so they produced 1:1 spiking over a broader range of afferent-stimulation frequencies. Intriguingly, muscarinic acetylcholine-receptor agonists also reduced frequency selectivity in CA1 neurons, and this effect was occluded by blocking SK channels.

These results suggest that by regulating the activity of dendritic SK channels, acetylcholine—and perhaps other neuromodulators—influences how CA1 neurons respond to inputs from CA3. Notably, hippocampal acetylcholine levels increase in novel environments. This increase might therefore allow hippocampal circuits to switch from memory-retrieval to memory-storage mode. Future work should investigate how such neuromodulation affects CA1 coupling with the entorhinal cortex.