When humans have an infection, like a cold or the flu, they are fatigued and achy, have a reduced appetite and elevated body temperature (aka a fever). Remarkably these symptoms of illness are largely shared across the animal kingdom as they represent the body’s natural response to infection that are critical for fighting pathogens and enabling recovery. Although one might assume that these symptoms arise as a side effect of the body’s immune reaction to pathogens, they are in fact orchestrated by the brain. In our new study, published in Nature, we have discovered how neurons in the brain directly read signals from the body’s immune system and how those signals alter neural circuit activity to trigger symptoms of sickness.
Our first goal was to identify the brain areas that may be responsible for generating sickness symptoms, particularly fever. Using mice as our model system, we first induced a mock bacterial infection by injecting the mice with a small amount of bacterial membrane component, called lipopolysaccharides (LPS). The subsequent inflammatory response that follows includes several sickness symptoms such as fever, loss of appetite, increased warmth-seeking behavior, decreased movement, and altered social interactions. We then scanned the brain for increased activity following LPS, indicated by the induced expression of the gene Fos, and saw that a specific area of the hypothalamus called the ventral medial preoptic area (VMPO) was highly activated compared to controls. The VMPO is located near a region of the brain with a ‘leaky’ blood brain barrier, and thus is accessible to circulating immune signals from the body.
To find out whether these neurons were important for generating fever and other sickness symptoms, we used a chemogenetic receptor to artificially activate the LPS-sensitive VMPO (VMPOLPS) neurons and found that there was an increase of 1-1.5 degrees Celsius, which is the typical range for a fever in mice. In addition, the mice no longer had interest in eating food, and they moved to a warmer temperature than usual when placed on a temperature gradient. These results indicated that the VMPOLPS neurons were in fact critical for the generation of several sickness symptoms such as fever, loss of appetite, and warmth seeking.
Interestingly, when we specifically ablated these neurons in absence of LPS injection, body temperature or appetite were not affected, indicating that VMPOLPS neurons are not part of normal homeostasis, or the set-points our bodies use to regulate body temperature and hunger, but instead are likely to alter normal set-points through connections with the circuits that control baseline homeostasic circuits. Indeed, we found that VMPOLPS neurons form direct synaptic connections with areas of the brain known to regulate body temperature and feeding behavior. When we specifically activated those individual connections, we were able to see changes in only body temperature or appetite alone. These experiments indicated that several homeostatic circuits are altered simultaneously, providing a clue as to why multiple symptoms typically arise together when we are sick.
Next, we wondered how VMPOLPS neurons are activated during an infection. Using a combination of single-cell and spatial transcriptomics we were able to genetically identify VMPOLPS neurons and characterize their spatial organization. We found that activated VMPOLPS neurons (i.e expressing Fos) were preferentially located near the blood brain barrier and expressed multiple receptors for receiving immune signals. There has been little evidence so far that immune signals can directly affect neuronal activity but we decided to test that idea. By adding specific immune molecules to slice preparations of the VMPO, such that we can simultaneously record neuronal activity, we found that indeed VMPOLPS were able to sense immune signals and increase their electrical activity. Further study showed us that these signals originated from the epithelial cell types lining the barriers between the blood and the cerebral spinal fluid.
In summary, our experiments uncovered how the brain is able to sense circulating immune signals resulting in changes in neural activity for specific neuronal populations. In addition, we showed that immune-sensitive VMPOLPS neurons represent a hub that coordinates multiple sickness symptoms by altering homeostatic set-points. This work opens new questions as to how other sickness symptoms, such as altered social interactions or sleep patterns, may be generated and whether brain activity is also altered during chronic inflammatory conditions.