Synopsis

Optogenetics has proven to be a highly useful tool to delineate the function of distinct proteins. Here, this approach was combined with fMRI to activate in-vivo selectively two interacting, but supposedly opposing, neuronal circuits of the central lateral amygdala (CEl). A classical fMRI paradigm was chosen to study the influence of the activation of either PKCδ- or somatostatin-expressing neurons on central pain processing, and to identify involved brain networks or areas. PKCδ was found to act preferentially anti-nociceptive by controlling via thalamus higher-order brain regions. Somatostatin on the other hand was shown to interact very closely with brainstem regions, controlling in a “bottom-up”-fashion thalamus, limbic system and cortex.

Introduction

During the last decades, functional magnetic resonance imaging (fMRI) was widely used to shed light on how the brain integrates sensory signals. Whilst we got great insight how the brain deciphers various sensory inputs, we are still far from understanding how it processes more complex, multifaceted signals such as pain.

It is known that not just single brain structures, but instead a vast network of interacting regions are involved in the many aspects of pain: besides sensory, cognitive and motor regions, the limbic system is involved in the development of fear, as aversion is highly protective. But with this quite general idea of pain processing, our knowledge on the level of involved neuronal circuits and their interaction with brain-wide networks thereby shaping the pain response needs further investigation. In this context, we explored central pain processing by combining optogenetics with BOLD fMRI (ofMRI), investigating in detail the distinct role of two neuronal circuits (PKCδ- and somatostatin (SST)-expressing cells) of the central lateral amygdala (CEl), which have shown to be the major player for the output function of the amygdala [1].

Materials and Methods

Two strains of transgenic mice were used:
*PKCδ::cre, expressing cre recombinase in PKCδ-expressing and
*SST::cre, expressing cre recombinase in SST-expressing cells in CEl.

8-12 week old male mice were assigned randomly to control and experimental groups.
The following viral constructs were injected stereotactically at CEl bregma -1.35, lateral 2.75, ventral 4.7:
*PKCδ::GFP - AAV2/5.EF1a.DIO.GFP.WPRE (GFP-PKC; controls)
*PKCδ::ChR2 - AAV2/5.hsyn.hChR2(H134R).eYFP.WPRE (PKC)
*SST::GFP - AAV2/5.EF1a.DIO.GFP.WPRE (GFP-SST; controls)
*SST::ChR2 - AAV2/5.hsyn.hChR2(H134R).eYFP.WPRE (SST).

200-400 μm ferrule-connected optogenetic fiber stubs were implanted with dental cement. After a rest period of 4 weeks (recovery and viral expression), the mice were measured once via fMRI (isoflurane anesthesia; 4.7T Bruker Biospec, matrix 64×64, FOV 15×15mm, voxel-size 0.234x0.234mm, slice thickness 0.5mm, axial, 22 slices) using a GE single-shot EPI sequence (TR=200ms; TEef=25.3ms).

For MR signal detection and good SNR, a 3cm 2x2 phased array head coil with two holes was fitted to allow passage of the implant. The laser fiber for optogenetic stimulation was fixed to the implant afterwards.

During the fMRI session, two different kinds of stimuli were applied alternatingly (interval 100sec):
*noxious heat (50°C; 20sec; dorsal side of right hind paw)
*laser stimulation via implant (473nm; 10mW; 10Hz; 20sec).
Every second heat stimulus was combined with simultaneous laser stimulation, to assess the influence of the optogenetic manipulation on noxious heat processing.

After GLM-analysis (single predictor for each condition), data of the animals were aligned by affine registration. 196 brain structures were identified using a modified Paxinos mouse brain atlas [2]. Influence of the optogenetic stimulation itself and thereof on pain perception were evaluated using functional connectivity (FC)-based network analysis [3].

Results

By comparing the experimental groups PKC and SST with their respective controls, laser stimulation was found to involve different brain regions: activation of PKC showed reduced FC within contralateral sensory cortex, thalamus and extensively within brainstem regions and the cerebellum (Fig.1 left).
Amygdala was found to be influenced mainly ipsilaterally, showing strong negative correlations with sensory cortex and the limbic system (Fig.2 left).

Activation of SST reduced FC within ipsilateral sensory/association cortex and the limbic system (Fig.1 right). Here, amygdala was found to respond also on the contralateral side and showed strong negative correlations with cortical and brainstem regions (Fig.2 right). Interestingly, strong negative correlations were also found between thalamus and brainstem, cortex and brainstem, as well as limbic system and brainstem, indicating that the CEl-SST circuit controls higher-order regions via brainstem.

The influence of laser stimulation on pain processing was evaluated by comparing the heat-only with the heat-laser-condition (Fig.3): For the CEl-PKC circuit, differences in heat processing could be found between thalamus, sensory cortex and hypothalamus, with a slight lateralization to the contralateral side, all reflected by reduced FC compared to the controls. The SST-circuit showed enhanced FC within cortex and cerebellum, and reduced FC within ipsilateral sensory/motor cortex as well as the brainstem.

Conclusion

For pain processing a new and differential modulatory effect of two amygdala circuits could be demonstrated by combining fMRI and optogenetics. PKCδ-expressing neurons were found to regulate pain by controlling thalamus, and thereby higher-order cortical regions. PKCδ is known in the literature to be described as “anti-nociceptive”, which fits nicely a kind of gate-control-mechanism of the thalamus, regulated by the amygdala.

SST-expressing neurons on the other hand controlled strongly the brainstem, and starting therefrom thalamic, limbic and cortical regions. As SST is known to be acting rather pro-nociceptive and fear-promoting, a strong involvement of the brainstem is very likely as it is known to play a major role in emotional processing as well as vegetative and motor responses to distressing stimuli.

Acknowledgements

References

[1] Haubensak, Wulf, et al. "Genetic dissection of an amygdala microcircuit that gates conditioned fear." Nature 468.7321 (2010): 270.

[2]Franklin KBJ, Paxinos G. The Mouse Brain in stereotaxic coordinates. Academic Press, New York, 3 Ed. 2008

[3] Sporns O, Chialvo DR, Kaiser M, Hilgetag CC. Organization, development and function of complex brain networks. Trends in cognitive sciences 2004;8(9):418-425.