The Aging Female Brain Retains More Metabolic Youth Across the Lifespan

Post by: Amanda McFarlan

What's the science?

In humans, the decline in brain metabolism with age is hypothesized to reflect the gradual ending of developmental processes in the brain as it reaches maturation. As such, any factors that influence the developing brain, such as sexual differentiation, are likely to play an important role in the aging process. This week in the PNAS, Goyal and colleagues used a combination of brain imaging techniques and machine learning to investigate the influence of sex on metabolism in the aging human brain.

How did they do it?

The authors included a total of 205 healthy male and female participants between the ages of 20 and 82 years old in the study and identified individuals in two groups: amyloid-negative individuals and asymptomatic amyloid-positive individuals. Amyloid is often found in individuals with mild cognitive impairment and  Alzheimer’s disease, however, these individuals were cognitively normal. All participants underwent a PET scan and structural MRI scan to measure several metabolic markers including total glucose usage, oxygen consumption, cerebral blood flow and aerobic glycolysis (the difference between total glucose and oxygen consumption). These metabolic measurements were normalized across all PET scan sessions for 79 brain regions. The authors used the normalized brain metabolism data from the amyloid-negative individuals to train a machine learning algorithm (random forest regression with bias correction and 10-fold validation) to predict the actual age of the participants, and then tested the ability of the algorithm to accurately predict a participant’s age based on their metabolic profile. The authors then performed three additional analyses: 1) To assess differences in metabolic profiles between males and females, the authors trained their machine learning algorithm on the normalized brain metabolism data from either males or females only, and then used the algorithm to predict the age of members of the opposite sex based on their metabolic profiles. 2) They performed further analyses to determine the impact of each individual metabolic parameter on the observed sex-based differences. 3) Finally, they applied their initial machine learning algorithm from amyloid-negative individuals to the data from amyloid-positive individuals (ages 60-80) to determine the effect of amyloid deposition on metabolism in the aging brain.

What did they find?

The authors found that the machine learning algorithm was able to closely predict an individual’s age based on their metabolic profile. They revealed that when training the machine learning algorithm on data from male participants only (and then used that algorithm to predict the age of females), the predicted metabolic age for females was on average 3.8 years younger compared to males. In support of these findings, the mean metabolic age for males was 2.4 years older compared to females when the machine learning algorithm was trained on data from female participants only (and then used to predict the age of males). Together, these data suggest that the metabolic profile of a female brain is younger compared to that of an age-matched male brain. Further analyses revealed that the sex-based differences in metabolic brain age were more strongly associated with brain glucose use rather than cerebral blood flow or oxygen consumption. Finally, the authors determined that the average metabolic brain age between age-matched amyloid-negative and amyloid-positive participants was not significantly different, suggesting that amyloid deposition does not account for variability in metabolic brain age across individuals.


What's the impact?

The authors provide evidence that the female brain is more youthful compared to the male brain across the lifespan in an in vivo study of brain metabolism. They used machine learning algorithms to show that on average the female brain is a few years younger than the male brain, from a metabolic perspective. These findings provide insight into how sex can impact glucose metabolism and the observed pattern of brain aging in different individuals.


Goyal et al. Persistent metabolic youth in the aging female brain. Proceedings of the National Academy of Sciences of the United States of America (2019). Access the original scientific publication here.

Type I Interferon Protects Neurons from Infectious Prions

Post by Elisa Guma

What's the science?

Prion diseases are progressive neurodegenerative diseases for which no effective treatment is available. They are associated with a buildup of misfolded forms of naturally occurring proteins in the brain, known as prion proteins. Once formed, prion proteins can convert other normal proteins into an abnormal form, causing a chain reaction, leading to accumulation of prions, neuronal death, and progressive cognitive decline. Neuroinflammation is known to be associated with prion diseases, however, the interaction between the immune system and prion accumulation in prion diseases remains unclear. This week in Brain, Ishibashi and colleagues used in vivo and ex vivo prion disease models to understand the protective role of type I interferon (I-IFN), part of the body’s innate immune response, against prion disease.  

How did they do it?

The authors first investigated expression of various inflammatory signaling genes in a prion-infected cell culture (ex vivo model). Next, the authors investigated the potential anti-prion effect of I-IFNs (alpha and beta interferons) in the cell culture model, first by administering the I-IFNs, and then by administering Poly I:C (which activates the innate immune system via I-IFN induction) to see if this could rescue the prion infection. They then investigated the potential protective property of IFN in mice that were prion infected by selectively expressing the IFN-beta gene in the brains of these mice and then measuring the prion proteins expressed in their brains. The authors wanted to confirm that prion suppression was due specifically to IFN signaling, therefore they generated a cell line that did not express IFN receptors and examined prion expression. They also infected normal (wild-type) mice, and mice lacking IFN receptors, and monitored prion protein expression and gliosis in the brain.

The authors also investigated the effects of RO8191, a compound known to bind to the I-IFN receptor, and increase IFN related gene expression and signaling. They first administered RO8191 to cells, and measured prion protein levels. They then tested the efficacy of R08191 treatment in mice, administering treatment from the time of prion infection until death (3x/week). Lastly, the authors tested the blood-brain-barrier (a protective layer between brain tissue and blood vessels connected to the rest of the body) permeability of RO8191 by measuring RO8191 concentration in the brain and spleens of the treated mice.

What did they find?

The authors found that prion infection decreased gene expression related to inflammatory signaling, including the I-IFN related gene. Next, they found that treating the prion infected cell line with IFN-beta (and alpha to a lesser extent), or Poly I:C (to stimulate IFN production) significantly reduced the number of prion proteins in the cell line. Introducing the IFN-beta into the brain of prion infected mice was also successful at reducing prion protein expression. The authors also observed that removal of IFN receptor genes significantly increased prion protein levels in cell lines, whereas and reintroduction of the IFN gene to these cells made them less susceptible to prion infection. Similarly, mice whose IFN genes had been knocked out were more susceptible to the prion infection - their lifespan was shortened, they had higher levels of prions in their brain and spleen and higher levels of gliosis (microglia and astrocytes) in their brain.


Next, the authors found that pre-treating prion infected cells and mice with R08191 decreased prion protein levels in the cells and in the brain and spleens of mice by at least 50%. Gliosis was also reduced in many brain regions including the cortex, thalamus and pons. Finally, the authors found that RO8191 had high blood-brain-barrier permeability, suggesting that it may reach and act on the brain, in addition to peripheral tissues.

What's the impact?

This study provides evidence that interferons may play a protective role against prion proteins in both cell lines and mice. Additionally, treatment with the novel small molecule RO8191, known to bind to the I-IFN receptor, was successful at reducing prion protein expression, making it a candidate for treatment. A better understanding of the role of the innate immune system in prion disease may provide ideas for novel therapeutic agents.

Ishibashi et al. Type I interferon protects neurons from prions in in vivo models. Brain (2019). Access the original scientific publication here.

Neuronal Maturation in the Developing Human Brain

Post by Shireen Parimoo

What's the science?

Ribonucleic acid (RNA) sequencing is a technique that is often used to create a genetic profile of cells in the brain. When applied at the level of single cells, RNA sequencing can be used to discern their identity. Studies using RNA sequencing have recently shown that many different types of neurons in the developing human brain arise from radial glia and other progenitor cells (cells that can differentiate into other cells). Although genetic profiles are useful for identifying cell types, other factors like cell physiology and morphology can also provide insight into cell identity. Currently, the physiological properties of various cells in the developing human brain – like how they respond to neurotransmitters – are not known. This week in Neuron, Mayer and colleagues developed a novel technique to combine RNA sequencing with cell imaging measures to identify the genetic and physiological profiles of developing human neocortical cells.

How did they do it?

The authors identified different cell types in tissue samples from the first and second trimester of the developing human neocortex including ventricular radial glia, outer radial glia, intermediate progenitor cells, and newborn neurons. They analyzed an RNA sequencing dataset to identify the gene expression levels of various neurotransmitter receptors and receptor subunits. They used single-molecule fluorescent in-situ hybridization and immunohistochemistry to examine the expression of a serotonergic receptor (HTR2A) and a purinergic receptor (P2RY1) in the radial glia specifically. To investigate the electrophysiological properties of these cells, they applied receptor agonists (molecules that bind to and activate receptors) to tissue slices and measured the change in the resulting electrical currents. Using calcium imaging as a measure of activity, they examined the effect of agonists on single cells from various regions of the neocortical tissue. To map the genetic profiles of neocortical cells with their physiological response profiles, the authors dissociated cells and dosed them with various receptor agonists to record their response, followed by RNA sequencing. They then investigated whether the maturational stage of newborn neurons affects their physiological responses. Finally, they trained two types of machine learning models (a supervised Bayes classifier and an unsupervised classifier) to predict cell identity from their genetic and physiological response profiles.

What did they find?

Cells in the developing brain differentially expressed genes for neurotransmitter receptors and their subunits. For example, the GRIN2A subunit of the NMDA receptor was highly expressed in progenitor cells, whereas the GRIN2B subunit was only expressed in neurons. Similarly, there was a high expression of the HTR2A receptor in radial glial cells. Application of receptor agonists also revealed a distinct response profile for each cell type. Specifically, NMDA receptors induced currents with different properties in neurons and radial glia. As the HTR2A receptor expression was only observed in the radial glial cells, applying agonists only induced currents in the ventricular and outer radial glia. In fact, applying an HTR2A antagonist (inhibitor) disrupted the morphology of the outer radial glia after 72 hours. This means that the various cells in the developing brain not only have distinct genetic and physiological response profiles, but these properties may also affect their morphology.


Neurons had more heterogeneous response profiles than progenitor cells, and this varied by their stage of maturation. The Bayes classifier was able to predict cell identities (in terms of physiological properties)  based on genetic data. However, it identified some cells as immature neurons, even though their physiological responses were more similar to mature neurons. Therefore, different readouts may help to better define neuronal maturation. On the other hand, the unsupervised classifier clustered cell types based on their physiological response profiles. Although genetic and physiological identification generally matched, some genetically identified cell types had several possible physiological responses, which was related to their maturational stage, for instance. This means that the same neuron can respond differentially to neurotransmitter signaling depending on its stage of maturation.

What's the impact?

This is the first study to use multimodal analyses to map the genetic profiles of developing human neocortical cells to their physiological profiles. In particular, the finding of distinct physiological response profiles across cell types is important because it highlights how the function of various cells and neurons in the brain can change as they mature. This has important implications for understanding the functional role of various cells and neurons in the brain.


Mayer et al. Multimodal single-cell analysis reveals physiological maturation in the developing human neocortex. Neuron. (2019). Access the original scientific publication here.