Optogenetic activation of inhibitory GABA neurons projecting from the limbic forebrain to the lateral hypothalamus causes this mouse to binge on cheese. The rapid onset and offset of the intense feeding behavior is striking. Credit: Jennings et al. (2013).
The hypothalamus is a collection of discrete nuclei in the vertebrate diencephalon that control a variety of metabolic, neuroendocrine, and circadian functions. Since the 1940s, the ventromedial nucleus (VM) has been known for its important role in satiety — lesions of this nucleus cause rats to become obese, while electrical stimulation of this structure curtails feeding. On the other hand, the lateral hypothalamus (LH) controls hunger. In 1953, Delgado and Anand implanted multilead electrodes into the brains of cats and found that electrical stimulation of the LH caused an increase of food intake to 500-1,000% of control levels.
Sixty years later, Jennings and colleagues (2013) set out to delineate the precise circuitry and neuronal population responsible for these effects using modern optogenetic techniques. They targeted a projection pathway from the bed nucleus of the stria terminalis (BNST), a part of the "extended amygdala" in the limbic forebrain, to the LH. These inhibitory GABAergic neurons synapse onto excitatory glutamatergic neurons in the LH.
This specific cell type was targeted by using a genetically modified mouse line. The mice express a recombination enzyme only in neurons that express the vesicular GABA transporter (vGAT-ires-cre mouse). A viral construct was used to insert a gene that codes for Channelrhodopsin-2, a light-sensitive protein that was fused to yellow fluorescent protein, directly into the BNST via microinjection. The BNST projections are stimulated by exposing them to blue light using specially implanted optical fibers. Since these neurons are GABAeric, they inhibit the postsynaptic LH neurons. This is shown schematically in the figure below.
Fig. 1 (modified from Jennings et al., 2013). VgatBNST→LH circuit activation induces feeding in well-fed mice. (A) Schematic showing VgatBNST→LH circuit targeting.
A different circuit was targeted in control mice, the projection from BNST to the ventral tegmental area (VTA). Activation of the VgatBNST→VTA projection did not induce voracious feeding behavior. But the mice did find it rewarding, which isn't a surprise... the VTA contains the dopaminergic cell bodies of the mesocortical dopamine system.
This is very impressive work in tune with the priorities of the BRAIN Initiative. Unaffiliated expert commenters have noted:
“This is a really important missing piece of the puzzle,” says neuroscientist Seth Blackshaw of Johns Hopkins University in Baltimore. “These are cell types that weren’t even predicted to exist.” A deeper understanding of how the brain orchestrates eating behavior could lead to better treatments for disorders such as anorexia and obesity, he says.
Cynthia Bulik, Distinguished Professor of Eating Disorders at UNC School of Medicine and the Gillings School of Global Public Health, says, “Stuber’s work drills down to the precise biological mechanisms that drive binge eating and will lead us away from stigmatizing explanations that invoke blame and a lack of willpower.”
Finally, we have one minor skeptic:
Previous studies from other groups had shown the opposite of what Stuber's team found: when other researchers activated the LH by exposing it to the neurotransmitter glutamate or by electrically stimulating the neurons, animals would start eating. However, B. Glenn Stanley, a professor at UC Riverside who studies the brain mechanisms of eating behavior, said Stuber's team’s results are not necessarily in conflict with earlier findings. “To see an inconsistency would be an oversimplification,” said Stanley.Stanley noted that Stuber's group focused on a subset of neurons in the LH, those interacting with neurons from the BNST. ...It's possible that some areas of the LH stimulate feeding when they're activated, and others when they are inhibited, Berthoud added. “The lateral hypothalamus is really a big area,” he said, adding that the authors didn't describe precisely where those neurons turned off by the BNST reside.
Two weeks ago, the BRAIN Working Group issued its Interim Report (PDF). High priority areas for 2014 include a focus on cell types and circuit manipulation:
#1. Generate a Census of Cell Types. It is within reach to characterize all cell types in the nervous system, and to develop tools to record, mark, and manipulate these precisely defined neurons in vivo. We envision an integrated, systematic census of neuronal and glial cell types, and new genetic and non-genetic tools to deliver genes, proteins, and chemicals to cells of interest. Priority should be given to methods that can be applied to many animal species and even to humans.
#4. Develop A Suite of Tools for Circuit Manipulation. By directly activating and inhibiting populations of neurons, neuroscience is progressing from observation to causation, and much more is possible. To enable the immense potential of circuit manipulation, a new generation of tools for optogenetics, pharmacogenetics, and biochemical and electromagnetic modulation should be developed for use in animals and eventually in human patients. Emphasis should be placed on achieving modulation of circuits in patterns that mimic natural activity.
Translation to Treatments for Human Obesity and Binge Eating Disorders
How close are we to applying this knowledge to treat people with severe intractable obesity? Not very. Optogenetics is a very invasive method. That's why development of new technologies is another high priority area of BRAIN (i.e., "advancing innovative neurotechnologies"). Nevertheless, it's a key component that will advance basic knowledge of neurocircuit functioning.
Has the potential for breakthrough treatments been overblown? Or is it all part of generating enthusiasm and public support (and hence $$)? Jennings et al. refrained from mentioning obesity except in the first and last sentences of their paper. In the press, senior author Garret Stuber said, “The study underscores that obesity and other eating disorders have a neurological basis. With further study, we could figure out how to regulate the activity of cells in a specific region of the brain and develop treatments.” [i.e., less restrained]
But what are these treatments?? How are they being developed? In its section on Human Neuroscience and Neurotechnology, the BRAIN Report notes that “The study of human brain function faces major challenges because many experimental approaches applicable to laboratory animals cannot be deployed in humans.” They call for improving the resolution and power of human neuroimaging methods, for instance, and learning more about the cellular mechanisms that generate the hemodynamic signal measured by fMRI.
Below I'll list a few more crucial areas where the animal and human studies can inform each other, using the lateral hypothalamus and obesity as illustrations. Then I'll conclude with speculations on integrating multiple levels (and perspectives).
Deep Brain Stimulation
Currently, the most obvious link between circuit manipulation in animals and humans is the increasing number of indications for deep brain stimulation (DBS). And here, pilot studies of DBS for the treatment of obesity have been ongoing for several years (Whiting et al., 2012). Stimulating electrodes were implanted bilaterally in the LH of three patients with refractory obesity. They were enrolled in a two year FDA-approved study on safety after meeting stringent selection criteria: unsuccessful gastric bypass surgery, weighing over 50% more than ideal body weight (BMI ≥ 40), passing physical and psychological examinations. The results were decidedly mixed. One patient lost no weight, and the other two lost only 12.3% and 16.4%. However, this was a study of safety (not efficacy). No undue adverse events were reported, although non-optimal LH stimulation could produce transient sensations of nausea and feeling too hot, while presumed VM stimulation could produce a transient anxiety or panic response.
Detailed Neuroanatomy of Human Hypothalamus
A new MRI atlas of the human hypothalamus was recently published in NeuroImage (Baroncini et al., 2012). Post-mortem histological sections were processed with Nissl staining of cell bodies and Sudan Black B to identify fiber tracts. These were compared to MRI sections and mapped to onto standard Montreal Neurological Institute (MNI) coordinate space. Developing more detailed anatomical atlases is critical for improving localization accuracy in stereotactic neurosurgery, including implantation of DBS electrodes. Perhaps the Hypothalamic Atlas of the future will include details about gene expression profiles (à la Allen Human Brain Atlas) and cell types within each nucleus.
Animal Models of DBS for Obesity
Studies of DBS have been conducted in rat (Sani et al., 2007) and minipig (Melega et al., 2012) models of obesity. Continuous bilateral stimulation to inhibit the LH caused relative weight loss in rats, while low frequency stimulation of VM was promising in minipigs. Electrode locations and stimulation parameters can be systematically tested in these animal models.
Neurology vs. Psychiatry vs. Clinical Psychology (why can't we all just get along?)
I found it striking that Dr. Stuber said his lab's study reinforces the view that obesity and other eating disorders have a neurological basis. Was his comment meant generically, that they have a basis in the brain? Or did he mean these are neurological disorders like epilepsy, Parkinson's disease, and Alzheimer's disease? These diseases are largely unaffected by your outlook on life, and they're not amenable to psychotherapy. How would a neurological definition of binge eating differ from a psychiatric one? That might depend if you're using DSM-5 or a futuristic biologically-based diagnostic scheme. In the U.S., federally funded research at NIMH is moving in a direction consonant with the circuit view presented here, i.e. the Research Domain Criteria (RDoC) framework.
Moving in the opposite direction, the British Psychological Society's Division of Clinical Psychology recently issued a Position Statement on the Classification of Behaviour and Experience in Relation to Functional Psychiatric Diagnoses, which advocated “a paradigm shift in how we understand mental distress towards one which is no longer based on diagnosis and a ‘disease’ model.” [a rather non-biological view]
Behaviors Don't Exist in a Subcortical Vacuum
Why is it so hard to land somewhere in the middle? What's completely missing from BRAIN is the recognition that social, emotional, and environmental factors can influence the onset of binge eating and other psychiatric disorders in humans. This complicates the utopia of circuit manipulation to cure eating disorders. What are the external factors (e.g., social pressure, stressors, negative evaluation) and internal states (e.g., emotional, cognitive, endocrine) that precipitate the onset (or end) of a binge? Certainly nothing as dramatic as optogenetic activation or inactivation.
The Report pays lip service to the prefrontal cortex and top-down control of behavior (p. 23), but modeling such influences is beyond the scope of the Initiative. But hey! you might say. The plan is ambitious and grandiose enough as it is... you can't even begin to scratch the surface with $100 million in 2014.
But why not push the speculation a bit further? If you're going to anticipate dreaming up entirely new technologies, why not shoot for the stars and identify top-down inputs to BNST neurons?
The Rise of the Circuit-Based Human
Vaughan Bell recently wrote an insightful article about Changing brains: why neuroscience is ending the Prozac era, in which he said:
As the Prozac nation fades, the empire of the circuit-based human will rise, probably to the point where dinner party chatter will include the misplaced jargon of systems neuroscience.
Ironically, Stuber the optogenetics researcher speculated that drug development might be the best and most practical way to go in the future:
Stuber said his group is going to focus on characterizing the cells in the circuit to see “what makes them special.” One appealing direction is to develop pharmacotherapies to tweak the activity of these cells and potentially change feeding behaviors long term.
The BRAIN Interim Report presents many reasons for optimism, but it is in part a political document, produced under the auspices of the Obama White House. It proposes a visionary collaborative world of neuroscience research where competition for scarce resources is unknown. It's an idealized blueprint where all the start-up funds aren't spent on establishing an infrastructure. All of these wonderful ideas are to be implemented in fiscal year 2014, which starts tomorrow... if the federal government doesn't shut down, that is.
Even in my Neurocomplimentary guise, I suppose my depressive realism still gets the best of me.
Baroncini M, Jissendi P, Balland E, Besson P, Pruvo JP, Francke JP, Dewailly D, Blond S, Prevot V. (2012). MRI atlas of the human hypothalamus. Neuroimage 59(1):168-80.
DELGADO JM & ANAND BK (1953). Increase of food intake induced by electrical stimulation of the lateral hypothalamus. The American journal of physiology, 172 (1), 162-8. PMID: 13030733
Jennings JH, Rizzi G, Stamatakis AM, Ung RL, & Stuber GD (2013). The inhibitory circuit architecture of the lateral hypothalamus orchestrates feeding. Science, 341 (6153), 1517-21. PMID: 24072922
Melega WP, Lacan G, Gorgulho AA, Behnke EJ, De Salles AA. (2012). Hypothalamic deep brain stimulation reduces weight gain in an obesity-animal model. PLoS One 7(1):e30672.
Sani S, Jobe K, Smith A, Kordower JH, Bakay RA.(2007). Deep brain stimulation for treatment of obesity in rats. J Neurosurg. 107(4):809-13.
Whiting DM, Tomycz ND, Bailes J, de Jonge L, Lecoultre V, Wilent B, Alcindor D, Prostko ER, Cheng BC, Angle C, Cantella D, Whiting BB, Mizes JS, Finnis KW, Ravussin E, & Oh MY (2013). Lateral hypothalamic area deep brain stimulation for refractory obesity: a pilot study with preliminary data on safety, body weight, and energy metabolism. Journal of neurosurgery, 119 (1), 56-63. PMID: 23560573
A well fed mouse from the experiment consuming bacon and donuts, despite already having its energy requirements met. Credit: Josh Jennings.