Here, we show that OGS of the NAc using ChR2 increased c-Fos expression and glucose metabolism in the region of stimulation. Additionally, we observed increased metabolism in regions inter-connected with the NAc, including the basal ganglia (caudate putamen, globus pallidus, and ventral pallidum) and limbic regions (amygdala, hippocampus) and decreased metabolism in regions of the default mode network or DMN (retrosplenial cortex and anterior cingulate gyrus), and in secondary motor cortex.
Our finding of increased glucose metabolism and c-Fos expression at the site of OGS—the NAc—is consistent with fMRI results reporting an increase in BOLD signal in the area of stimulation (for prior study it was the motor cortex) (Lee, 2012
), as well as immunohistochemistry studies showing increased c-Fos expression at the site of stimulation (Adamantidis et al., 2010
; Lobo et al., 2010
). This indicates that OGS with ChR2 activates neurons in the region where the light is delivered. As predicted, we also observed increased metabolism (activation) in regions that are connected to the NAc; specifically, the hippocampus, caudate/putamen, amygdala, periaqueductal gray, globus pallidus, and ventral pallidum. This is consistent with fMRI studies showing BOLD signal increases in regions neuroanatomically connected to the site of stimulation (Lee et al., 2010
; Lee, 2012
Interestingly, we showed decreased metabolic activity in the retrosplenial cortex (posterior cingulate gyrus), anterior cingulate gyrus, and secondary motor cortex. Both the retrosplenial cortex and the anterior cingulate gyrus are part of the DMN that is deactivated (including decreases in glucose metabolism) when engaging in a task (Raichle and Snyder, 2007
; Pfefferbaum et al., 2011
). The clusters of inhibition that we observed following OGS of the NAc fall within the previously reported DMN of the rat brain (Upadhyay et al., 2011
; Lu et al., 2012
). Moreover, brain imaging studies indicate a correlation between markers of dopamine (DA) neurotransmission and deactivation of the DMN (Tomasi et al., 2009
; Dang et al., 2012
; Sambataro et al., 2013
), which, in conjunction with our findings, suggests that NAc activation facilitates DMN inhibition.
A previous optogenetic fMRI study reported local excitatory responses (positive BOLD signal) following stimulation of the motor cortex with ChR2, flanked by inhibitory responses (negative BOLD signal) in lateral regions; however, no inhibitory responses in non-adjacent regions were reported (Lee et al., 2010
). The NAc is comprised mostly of small spiny GABAergic neurons, which constitute the source of projections out of the NAc, but there are also acetylcholine interneurons that modulate activity of GABAergic neurons. Thus, increased metabolism in NAc is likely to reflect the activation of both sets of neurons, unless OGS also indirectly influences activity of afferent terminals into the region, in which case the increased metabolism could reflect increased glutamate release from cortico-striatal, thalamo-striatal, or amygdo-striatal terminals or DA release from DA terminals (Kegeles et al., 2000
; Fernandez et al., 2006
; Eyjolfsson et al., 2011
; Surmeier and Graybiel, 2012
). Since the GABAergic neurons are the ones that project out of the NAc, the activation in connected regions is likely to reflect indirect circuit-level consequences of NAc activation. Studies that restrict the expression of the ChR2 to the NAc’s two main types of projection neurons, those predominantly expressing DA D1 versus DA D2 receptors, are needed to clarify the circuitry underlying the metabolic changes seen with NAc stimulation. Note that Lee et al. used a vector that was designed to specifically drive the expression of ChR2 in Ca2+
/calmodulin-dependent protein kinase II α (CaMKIIα)-expressing principal cortical neurons, which are excitatory, but not in GABAergic or other inhibitory cell types (Lee et al., 2010
). However, we may have also stimulated cholinergic interneurons in the NAc; for while they constitute less than 1% of NAc neurons (Rymar et al., 2004
), their optogenetic activation and inhibition has been shown to modulate activity in other NAc neurons (Witten et al., 2010
Downstream effects of OGS of the NAc seen in the amygdala and hippocampus are supported by known anatomical and functional connectivity that constitute the limbic system (Parkinson et al., 2000
; Cardinal et al., 2002
; Heidbreder and Groenewegen, 2003
; Morgane et al., 2005
). The activation of the basal ganglia (caudate/putamen, globus pallidus, and ventral pallidum) is also consistent with the basal ganglia modulating or being modulated by activity in the NAc, as well as the hypothesis that the NAc is the interface between limbic and motor systems (Groenewegen and Uylings, 2000
; Heimer, 2003
; Morgane et al., 2005
; Postuma and Dagher, 2006
). The activation of the periaqueductal gray (PAG) could reflect the functional loop between the NAc, PAG, and amygdala that underlies opioid-mediated antinociception (Ma and Han, 1991
; Ma et al., 1992
). Our results are also consistent with findings in the human brain of functional connectivity between the NAc and the amygdala, hippocampus, globus pallidus, caudate/putamen, anterior and posterior cingulate, and precuneus (retrosplenial cortex in the rat) (Cauda et al., 2011
). Clinical studies have shown additional connectivity of the NAc with the orbitofrontal cortex, insula, and midbrain (Di Martino et al., 2008
; Cauda et al., 2011
); however, the lack of an observed effect in these regions in our study may reflect distinct connectivity patterns when the brain is studied during a resting state versus when it is studied during regional activation, as well as species and methodological differences.
In our study, NAc stimulation increased metabolism in the contralateral amygdala and hippocampus, which is consistent with a lateralization of NAc connections with these two other regions (Cauda et al. 2011
); however, we did not see ipsilateral activation of these limbic regions. It is possible that this contralateral activation of downstream regions represents a compensatory mechanism for maintaining homeostasis. A similar effect has been demonstrated, such that unilateral stimulation of the subthalamic nucleus results in contralateral activation of downstream basal ganglia circuitry (Parent and Hazrati, 1995
; Liu et al., 2002
; Arai et al., 2008
). We further explored c-Fos activation in some of the contralaterally activated regions, as well as control regions with ipsilateral and bilateral activation, following OGS of the NAc to determine whether these could be artifacts. Our findings from c-Fos immunolabeling corroborate our μPET findings, showing increased expression in the contralateral amygdala and globus pallidus, ipsilateral secondary somatosensory cortex, and bilateral periaqueductal gray.
In our study, we also showed decreased metabolic activity in the motor cortex following stimulation of the NAc that might reflect the fact that during OGS the rats were placed in a small cage, which was used to prevent stimulation-induced locomotor hyperactivity. Thus, the restricted containment may have required that the animals inhibit motoric brain regions that otherwise would have been activated had they had the space to move. However, a previous study of unilateral OGS of the NAc also failed to increase basal locomotor activity even when mice were in a larger space (Lobo et al., 2010
), which indicates that unilateral OGS of the NAc may be insufficient to elicit significant increases in locomotor activation. Previous studies have shown that locomotion-stimulating drugs are less effective when microinjected unilaterally into the NAc, compared to bilaterally (Jackson et al., 1975
; Essman et al., 1993
; Schildein et al., 1998
). Additionally, the observed contralateral activation of downstream regions, previously proposed to be a compensatory mechanism, may be the neurobiological substrate by which the brain is homeostatically regulating behavioral output in response to unilateral stimulation of the NAc.
A limitation of our study is that since the vector we used equally infects all neurons, we cannot distinguish which neuronal cell type(s) drive the increases in metabolism in the NAc and its effects on downstream brain regions. As well, rats were food-deprived overnight (12h), accordingly to standard protocols, prior to FDG PET scans to attain consistency in blood glucose levels as abnormal blood glucose levels interfere with FDG uptake (Wong et al., 2011
; Fueger et al., 2006
). While longer food deprivation (24-48h) is known to be a stressor and to affect behavioral sensitivity to both natural and food rewards (Levine et al., 1995
; Shalev et al., 2003b
; Shalev et al., 2003a
), it does not affect c-Fos immunoreactivity in the NAc or other regions of the reward pathway (Shalev et al., 2003a
). This may be less of a concern in the present study as food deprivation was brief (12h). Lastly, although a potential confound could have been motor activation induced by OGS, this is unlikely to be the case since there were no differences in locomotor activation between the GFP and ChR2 rats during the stimulation session.
This study shows that ChR2-mediated OGS of the NAc results in activation of the NAc (as measured by both glucose metabolism and c-Fos expression), in addition to activation and inhibition of downstream projection regions. These results also provide evidence of the feasibility of using μPET with FDG in conjunction with OGS to map connectivity in the awake, behaving rat brain.