Chaudhury vs Tye Papers
Seminar in Biological Psychology
Dr. Rebecca Shansky
The most striking facet of these two pieces of work is that they were both published in Nature in the same year, and both papers featured Karl Deisseroth, a pioneer in the optogenetics field. Furthermore, the papers published arguably contradictory information — although that was accounted for in the discussion portions of the respective letters.
Chaudhury et al took a social defeat approach to the problem. They asked, if we subject mice to a behavioral paradigm where they experience enough stress to remain resilient but have undergone considerable stress, and also optically activate a specific subset of neurons in the ventral tegmental area (VTA) during the same time, what behavior will these mice manifest? Will they remain resilient? Will they exhibit social defeat? And thus, what is the function of this particular subset of neurons? According to previous literature, it had been well established that the dopamine (DA) neurons in the VTA had been shown to play a role in reward behavior, and to be implicated in disorders such as anxiety and depression. Thus, they tackled this issue by utilizing a tyrosine hydroxylase (TH) - cre transgenic mouse, and virally expressing channelrhodopsin (ChR2) or halorhodopsin (NpHR) in TH-positive cells, using a cre-dependent opsin virus (TH is a precursor for catecholamines such as dopamine).
They variably stimulated these dopamine cells in the VTA (tonically and phasically), to observe that phasic — and not tonic — stimulation had a strong effect on a mouse’s motivational and hedonic state after experiencing subthreshold social defeat. Furthermore, they showed that in a mouse that undergoes subthreshold social defeat without optogenetic stimulation, susceptibility can still be induced by phasic stimulation during social interaction or sucrose preference assessments.
After this, they moved away from a transgenic approach to use a pathway-specific approach using a retrograde cre virus. They knew that the nucleus accumbens (NAc) had been implicated in the reward pathway, and so they specifically expressed ChR2 in VTA->NAc direct projections. The issue with this technique (compared to the last one) is the loss of cell-type specificity. The beginning experiments were based on the finding that the specific firing pattern induced in the specific neuronal cell type (VTA DA neurons) induced a susceptibility phenotype, but this was not as directly addressed in this next experiment. Nonetheless, they found that increased phasic firing of VTA->NAc projections accompanied the susceptibility phenotype in both measures (social interaction and sucrose preference), and conversely, inactivation of this projection showed an increase in the resilience phenotype. They followed up with an experiment specifically targeting the VTA->mPFC projections and found that there was no change in the behavioral phenotype in sucrose preference, but there was a change in social interaction when the projection was inhibited. This brings up the point that these two tests may not be the same measure for social defeat. What is the best animal correlate of human depression? Of human anxiety?
Tye et al took a similar approach, also using TH-cre mice to optogenetically stimulate or inhibit DA neurons in the VTA (and furthermore, to specifically target VTA->NAc projections), but they took different behavioral approaches. Instead of using a social defeat model, Tye et al utilized a chronic mild stress model. Both papers addressed that a potential reason for the differential findings may stem from a longstanding idea that chronic mild stress and social defeat recruit the same neural circuits in an opposing manner.
Tye et al found that activation of VTA DA neurons after chronic mild stress induces a resilience phenotype, while inactivation of that ensemble produces the opposing phenotype. Again, this may be due to the behavioral approach.
A good control that they used was the open field test along with optogenetic stimulation, which helped say that the behavioral effects they were seeing were indeed due to motivational changes and not simply to locomotor changes. Another good control was the inactivation of glutamatergic neurons along with dopamine neurons in the NAc, which held a similar purpose — to say that the dopaminergic cells in the NAc were recruited in the manifestation of this phenotype.
A clever experimental approach they took, however, was to pharmacologically inactivate NAc neurons and see that the behavioral effects of VTA activation were nullified. That is to say that there is something about the connectivity of VTA->NAc circuits that may affect behavior. To probe the causality of VTA activation, they concurrently activated VTA DA neurons and recorded from the NAc in vivo during a stressful task (the forced swim test). This was a great way to follow up the claim that the VTA->NAc pathway may indeed be leading to behavioral changes related to motivation.
In the end, they spike sorted the waveforms they had accumulated during the in vivo recordings, and the issue of this sorting technique is that it is usually done manually and is not perfect. Moreover, they found that all the possible responses a cell could have had, they found a cell that did. In other words, a neuron could have responded to VTA activation, could have preceded an increase in escape-related behaviors, or both. And they found cells that exhibited all 3 properties.
The two papers highlight a key talking point in behavioral neuroscience: not all behavioral tests that test for disorders are necessarily the same, or test for the same thing at all. And so interpretations of the data must be taken carefully.
Note: Tye et al used 500nL for all their viral injections. That’s a lot! Did they not worry about diffusion of the virus/aspecific infection to neighboring regions of the brain?