A1-reactive astrocytes and IFNAR signaling collectively induce neuronal cell death during infection of IFNAR1−/− mice by severe fever with thrombocytopenia syndrome virus

Severe fever with thrombocytopenia syndrome (SFTS) is caused by the SFTS virus (SFTSV), a bunyavirus that is endemic throughout eastern Asia. SFTSV was first discovered in 2009 in central China¹ and has since infected about 5500 people in China² and approximately 800 each in Japan and South Korea.³ The disease (SFTS) is characterized by fever, thrombocytopenia, and central nervous system (CNS) symptoms including headache, vertigo, coma, and encephalitis^{4, 5} which has impacted as many as 45% of SFTS patients.⁶ The mortality rate caused by SFTSV infection varies by region and by year with reported rates as high as 27%.^{7, 8} There are currently no therapeutics or vaccines available for SFTS treatment and prevention.

While clinical manifestations such as decreased blood flow to the brain resulting from systemic SFTSV infection likely contribute to the disease pathology,^{9, 10} some studies have shown that direct viral infection of the brain may also play an important role in disease pathogenesis. Evidence from a newborn mouse-infection model has shown that SFSTV can infect microglia to cause inflammasome-mediated neuronal cell death,¹¹ which could in part explain how some immunocompromised individuals are more likely to develop severe disease as a result of the infection,^4-6 whereas most patients who succumb to SFTS show symptoms of severe encephalitis before death. Nevertheless, the molecular mechanisms of virus-induced neuronal pathogenesis and the location of virus-induced brain damage have not been fully characterized.

Kim et al. have recently published a paper in Journal of Medical Virology,¹² in which they infected the interferon-alpha receptor knock-out (IFNAR1−/−) adult mice intraperitoneally with SFTSV to deliver the virus systematically and to show that the virus could successfully cross the blood–brain barrier (BBB) to invade the CNS as evidenced by the presence of the viral protein and genomic content being detectable throughout the CNS 4 days after the infection. They showed that SFTSV infected the brainstem and spinal cord of the IFNAR1−/− mice and that A1-reactive astrocytes were activated by virus infection, leading to neuronal cell death. Because the brainstem and spinal cord are regions of the brain that have been associated with respiratory function and motor neurons in IFNAR1−/− mice, SFTSV infection leading to lethal neuroinflammation and neuronal cell death may underlie the cause of disease pathogenesis, pathology, and mortality of some SFTS patients.

Specifically, the authors showed that viral gene expressions were significantly higher in the olfactory bulb, cortex, cerebellum, brainstem, and spinal cord of the IFNAR1−/− mice. They also characterized the immunological effects of SFTSV infection of the brain and showed that IFNAR deficiency also increased neuronal cell death as demonstrated by a significant decrease in viable neurons by Day-4 postinfection. They showed decreased expression levels of most tested pro-inflammatory cytokines, such as IFNγ, interleukin (IL)1α, IL1β, IL6, and tumor necrosis factor-α. Two exceptions were C1q and C3 cytokines, which were higher in the virus-infected IFNAR1−/− mice. Because these two cytokines have been implicated in the activation of the A1-reactive astrocytes, the authors expanded on this finding by investigating astrocyte signaling and found significantly higher levels of activated astrocytes and A1-astrocyte gene expression in the virus-infected IFNAR1−/− mice.

The authors also showed that primary astrocytes of SFTSV-infected IFNAR1−/− mice also induced neuronal cell death through the activation of A1-reactive astrocytes. To do this, they investigated the interactions between infected neurons and microglia or astrocytes by isolating primary neurons, microglia, and astrocytes from embryonic or neonatal wild type (WT) and IFNAR1−/− mouse pups and infecting those cells with SFTSV in culture and incubating the infected neurons with the cell-culture media supernatant collected from infected microglia or astrocytes. Conditioned media collected from infected WT and IFNAR1−/− astrocytes but not those from microglia increased neuronal cell death, which reached statistical significance with infected IFNAR1−/− neurons. Furthermore, conditioned media from infected IFNAR1−/− astrocytes had a greater level of increases in neuronal death than those from WT astrocytes. Culturing uninfected neurons with conditioned media from infected microglia or astrocytes demonstrated that the conditioned media supernatant did not pass the infectious virus onto neurons. However, a consideration that was not being addressed by this cell-culture model was the potential impact of direct cell-cell contact, which could be investigated in future experiments in a coculture model.

It is important to note that while the authors also characterized the rates of morbidity and mortality of several SFTSV genotypes in IFNAR1−/− mice, they did not directly compare those in SFTSV-infected WT mice, and thus, limiting a direct comparison of the potential disease progression due to the virus strains and/or mouse models used. It is also noteworthy that IFNAR1−/− mice have also previously been shown to suffer from enhanced viral breach of the BBB to cause neuronal infection by West Nile virus.¹³ Astrocytes have also been shown to be infected during bunyaviral infection,^14-16 and studies using other neurotropic viruses have found them to be the first sites of productive infection in the brain.^17-21 They can also act as longer-lasting viral reservoirs during chronic viral infection,²¹ making their inflammatory responses to viral infection potentially implicated in neuronal pathogenesis throughout the course of the infection. By contrast, productive bunyaviral infection of microglia has so far only been shown in cell culture^22-24 and in neonatal mice¹¹ or humanized mice¹⁴ despite other studies using bunyaviral infection in adult mice showing microglial activation.^{15, 24} As such, the current study¹² is an important first step in studying SFTSV-induced neuropathogenesis in adult animals.

In summary, the authors of the current study¹² demonstrated that neuronal cell death and immune activation were observed in the brain of adult mice when SFTSV was delivered systemically and could significantly impacted by the loss of IFNAR cellular signaling. Specifically, A1-reactive astrocytes seemed to be more strongly associated with neuronal cell death than microglia. While not directly addressed, it's also possible that A1-reactive astrocytes could be associated with higher levels of SFTSV infection, since the sites of the brain most impacted by IFNAR signaling are in prolonged contact with astrocytes associated with the BBB. This may allow for the high number of susceptible astrocytes concentrated in the BBB to become a site of productive infection upon viral breach of the BBB. Microglia may therefore have a less robust relationship with infection levels and neuronal cell death due to their more diffuse presence throughout the brain. It should also be noted that astrocytes^{25, 26} and microglia^{27, 28} secrete chemokines to recruit peripheral immune cells such as lymphocytes, which have many documented pro- and anti-inflammatory interactions with cells throughout the brain. The contribution of these interactions toward viral control and disease pathogenesis will be another point of study in future investigations of bunyaviral infections.

This adult mouse model described in the current study¹² presents an intriguing alternative to neonatal mouse-infection model to study SFTSV-caused disease pathology and pathogenesis. Because microglia in neonatal mice continue to develop in the first couple weeks of life,^29-31 it is important to develop an alternative animal model, such as the one reported in the current study,¹² to study SFSTV infection in vivo. This will in turn contribute toward future efforts in developing SFSTV therapeutics and vaccines, such as targeting specific cell types in the brain and optimizing their performance in patients, including immunocompromised individuals, who are most prone to develop severe SFTS disease.

The authors declare no conflict of interest.