Changes in cytokine and chemokine profiles in mouse serum and brain, and in human neural cells, upon tick-borne encephalitis virus infection

Tick-borne encephalitis (TBE) is a severe and potentially lethal neurological infection caused by the TBE virus (TBEV), for which there are presently no specific and effective therapies available. Each year, over 13,000 TBE cases, and numerous TBE-related deaths, are reported in the large forested TBEV-endemic regions of Europe and Asia [1]. TBEV and several other medically important and emerging viral pathogens (e.g., yellow fever virus, West Nile virus, Japanese encephalitis virus, or Zika virus) are members of the genus Flavivirus within the family Flaviviridae, that constitute a group of small enveloped viruses with genomes comprising a single-stranded RNA of positive polarity [2]. TBEV is primarily transmitted to humans by Ixodes sp. ticks and also exhibits alimentary transmission following consumption of unpasteurized milk or milk products from TBEV-infected goats, sheep, and cows. The clinical presentation of TBE in humans ranges from mild flu-like fever to severe encephalitis or encephalomyelitis [1, 3]. The main target organ for TBEV is the host brain, where the virus primarily infects neurons [4]. However, other neural cells, such as astrocytes, are also reportedly susceptible to TBEV infection [5, 6]. Due to their anatomical association with the blood–brain barrier, astrocytes may constitute an important route through which TBEV penetrates the brain [7].

The outcome of TBE is largely determined by the balance between damage caused by viral replication and the immune response [8]. The mammalian immune system employs sophisticated antiviral measures to limit virus replication, but this process can have immunopathological consequences. The protective and pathological roles of the immune reaction must be balanced to support both efficient virus clearing and protection from immune system-mediated damage to brain tissue.

Immune response initiation and the recruitment of immune cells to the brain are primarily mediated by cytokines, chemokines, and interferons. Many studies describe the critical role of type I interferons (IFNs) in the defense against TBEV [9]. TBEV infection reportedly induces type I IFN expression in host cells, and the virus is highly sensitive to host cell pretreatment with IFNs [10‐16]. In virus-infected cells and the surrounding tissues, the antiviral action of IFNs is mediated by the expression of hundreds of IFN-stimulated genes (ISG) [17]. Viperin is an important ISG that strongly inhibits TBEV, targeting the virus at multiple stages of its life cycle [9, 15, 16, 18‐20]. However, little is known about the roles of soluble mediators other than IFNs, particularly other cytokines and chemokines, with regards to protection or immunopathology during TBE.

Our current knowledge about TBEV-associated cytokine and chemokine expression has been primarily obtained through analyses of human samples. Increased levels of various cytokines and chemokines have been detected in serum or cerebrospinal fluid (CSF) samples from TBE patients, and some of these cytokines/chemokines are considered biomarkers that indicate a severe course of infection. Human TBE patients exhibit elevated serum levels of the pro-inflammatory cytokines interleukin (IL)-6, IL-8, IL-12, IL-17A, and IL-17F and of the chemokines CCL3, CXCL2, CXCL10, and CXCL13 and elevated CSF levels of IL-8, IL-10, IL-17F, IL-22, CCL2, CCL3, CCL5/RANTES, CXCL1-2, and CXCL10-13 [21‐27]. Compared to control patients, TBE patients exhibit higher ratios of IL-12:IL-4 and IL-12:IL-10, reflecting the global pro-inflammatory cytokine balance [28]. However, there remain unanswered questions regarding the dynamics of the production of these cytokines/chemokines during the course of TBE, about their antiviral or harmful roles and about the cell types that produce these immunologically important soluble factors within the host body.

In the present study, we aimed to assess the concentrations of different cytokines and chemokines in blood serum and brain tissue samples from mice at different time-points after infection with a lethal dose of TBEV. For these analyses, we employed a multiplex array using a fluorescent bead-based technology. We also used this system to investigate the TBEV-infection-evoked production of multiple cytokines and chemokines by human neural cells, and to compare the cytokine/chemokine communication and induction of antiviral response between naïve and infected neuronal cells and astrocytes.

TBEV infection primarily targets neurons, leading to their damage and death [3]. Accumulating evidence suggests that some neuronal damage may also be mediated by the immune response itself [32]. CNS infection leads to T cell recruitment, followed by cytotoxic T cell-induced immunopathogenesis [32]. Chemokines help regulate leukocyte trafficking into the CNS. Leukocytes are not only essential components of cell-mediated immunity and antiviral host defense, but also of immune system-mediated pathology [21]. To date, little is known about the dynamics of cytokine/chemokine production in serum and brain tissue or about the relative contributions of resident CNS cells (e.g., neurons and astrocytes) to inflammation development during TBE. The results of our present study yielded four major findings. First, CXCL10 and G-CSF are the earliest and most robustly upregulated chemokines, exhibiting increased presence in the serum of TBEV-infected mice as early as during the viremic phase. Second, the TBEV-infected brain shows increased levels of numerous chemokines, colony-stimulating factors, and pro-inflammatory cytokines, and these increases correlate with virus growth in the infected brain tissue. Third, TBEV-infected astrocytes, and to a lesser extent neuronal cells, are a potential source of pro-inflammatory cytokines/chemokines in the CNS. Fourth, cytokines/chemokines produced by TBEV-infected astrocytes or neuronal cells reduce virus growth in neuronal cells and increase the survival of infected neuronal cells or astrocytes, respectively.

In the mouse model of TBE, which recapitulates pathogenesis observed in severe/lethal cases of TBE in humans [33], we measured the cytokine/chemokine levels in serum and brain tissue at different stages of infection. In agreement with previous studies [34, 35], peak viremia occurred on day 3 p.i., and the viremic phase was followed by the virus reaching the brain, where the virus titer continuously increased until the end of the experiment. The titers reached in mouse sera are comparable with titers seen in our previous studies [32‐35], and TBEV RNA levels seen in sera of human patients collected in the first phase of the disease [36, 37]. CXCL10 and G-CSF were the earliest and most robustly upregulated chemokines, showing an increased presence in the serum of TBEV-infected mice as early as during the viremic phase. A prior evaluation of chemokine concentrations in the serum of patients with TBE and a control group indicated that CXCL10 may be a good serum biomarker of TBEV-induced CNS inflammation [21]. Our present study demonstrated that increased serum levels of CXCL10 appeared before the virus reached the brain, possibly reflecting viral infection of peripheral tissues and organs during the first phase of disease; however, the exact source of this chemokine remains elusive. On days 7 and 9 p.i., we observed increased concentrations of IL-6 and IL-10. This is in agreement with prior analyses of serum samples from TBE patients, which have revealed increased serum levels of IL-6 [28, 38, 39] and IL-10 at admission, with these increases remaining for 4 weeks after treatment [38, 39]. Overall, comparisons of our recent results in a mouse model with the findings from previous studies of human TBE patients reveal that CXCL10 is the main serum biomarker associated with TBE in both TBEV-infected mice and humans.

Increased CXCL10 levels were detectable in the brain tissue of TBEV-infected mice at the time of viremia and viral neuroinvasion. During these stages of infection, little or no virus could be detected in the brain; thus, the elevated CXCL10 in the brain at this time was likely of peripheral origin. At later time-points, as the neuroinfection progressed, CXCL10 levels in the brain continued to substantially increase. Since serum CXCL10 levels remained relatively stable across these time-points, the increased CXCL10 level in the brain was most likely of neural origin. Increased CXCL10 concentrations have also been reported in the CSF of human TBE patients [21, 22, 27], where CXCL10 plays roles in recruiting CXCR3-expressing T cells into the brain/CSF [27, 40]. The recruitment of antigen-specific CD8⁺ T lymphocytes into the brain parenchyma is a key step in clearing viral infection from the brain [41, 42]; however, these cells can also play pathological roles under some circumstances [32]. Previous studies by our group [34] and others [43‐46] using different viruses and models have demonstrated that excessively high CXCL10 levels in the CNS can be harmful to the host. For example, a study of mice with different susceptibilities to TBEV infection revealed that higher CXCL10 expression in the brain correlated with more severe course of infection and poorer outcome [34]. CXCL10 can cause pathology through excessive recruitment of cytotoxic T cells, and by cytotoxic action on neurons [47].

On day 9 p.i., which was the time of advanced brain infection, infected mouse brains exhibited increased levels of numerous other chemokines, colony-stimulating factors, and pro-inflammatory cytokines. This is in full agreement with the findings of previous studies in mice [34, 35, 48] and of studies analyzing CSF samples from patients with TBE [21, 23, 49, 50]. CSF samples from children or adult TBE patients have exhibited higher levels of CXCL1 [25], IFN-γ, IL-4, IL-6, IL-8 [51], CXCL10, CXCL11, CXCL12, CXCL13 [21], CCL2 [50], CCL3 [49], and other pro-inflammatory soluble factors. Additionally, increased concentrations of CCL5 (RANTES) and IL-5 have been detected in the CSF, but not in the serum, of patients with TBE [23, 26]. Similarly, in the present study, TBEV-infected mice showed increased concentrations of CCL5 (RANTES) and IL-5 on day 9 p.i. only in the brain tissue, with no changes in the serum levels of these molecules throughout the experiment. This is also in agreement with the findings of Zhang et al. [52]. Antagonizing RANTES within the CNS reportedly extended the survival of TBEV-infected mice and reduced the accumulation of infiltrating cells in the brain after TBEV infection [52], indicating that CCL5 (RANTES) causes neuroinflammation and may contribute to brain destruction during TBE [52, 53]. In summary, TBEV infection elicits strong production of various cytokines, chemokines, and colony-stimulating factors in mouse brains. For most of these soluble factors, the results in TBEV-infected mice show a good correlation with the findings in human CSF samples. CXCL10 is one of the earliest and most robustly upregulated chemokines in the TBEV-infected brain.

We next investigated the relative contributions of major resident CNS cells—specifically neuronal cells and astrocytes (HBCA)—as mediators of brain inflammation, by examining the ability of TBEV-infected human neuroblastoma cells (SK-N-SH) to produce key cytokines and chemokines. In response to TBEV infection, SK-N-SH exhibited increased production of IL-8 and CCL5 (RANTES). In contrast, infection of SK-N-SH with WNV reportedly results in a sharp increase of key pro-inflammatory cytokines (e.g., IL-1β, IL-6, IL-8, and TNF-α) at day 2 p.i., which coincides with peak virus replication [54]. It seems likely that neurons may be a main source of cytokines in WNV-associated neuroinflammation [54]. In contrast, TBEV-infected neuronal cells seem to contribute less to the development of brain inflammation in TBE compared to WNV encephalitis.

In our previous study, we found that HBCA were permissive to TBEV infection and that infection with the low-pathogenic TBEV strain Neudoerfl induced markedly increased expression of glial fibrillary acidic protein (GFAP), which is a marker of astrocyte activation [5]. We also observed the upregulated production of matrix metalloproteinase 9, and the upregulated mRNA expression and/or protein production of several key pro-inflammatory cytokines/chemokines, including TNF-α, IFNα, IL-1β, IL-6, IL-8, CXCL10, and CCL4, but not MCP-1 [5]. In our present study, we investigated the production of a broader panel of cytokines/chemokines in HBCA infected with the highly pathogenic TBEV strain Hypr. While HBCA infection with Neudoerfl and primary rat astrocyte infection with the Ljubljana I strain do not affects cell viability [5, 6], infection of HBCA with Hypr led to a slightly advanced CPE and cell death. Our results showed that Hypr infection of HBCA induced increased production of numerous cytokines/chemokines—particularly CCL5 (RANTES), CXCL10, IL-6, and IL-8, which confirmed our previous findings. While TNF-α mRNA expression was upregulated in HBCA infected with the Neudoerfl strain [5], we found that TNF-α remained at undetectable levels in HBCA cultures infected with Hypr strains. Additionally, MCP-1 mRNA expression was not upregulated in Neudoerfl-infected HBCA [5], but MCP-1 protein expression was upregulated in Hypr-infected HBCA.

Other studies have also reported that TBEV infection upregulates CCL5 (RANTES) production at both the mRNA and protein levels in primary astrocytes [52, 53]. Specifically, RANTES production in human primary astrocytes is induced by viral NS5 (particularly the RNA-dependent RNA polymerase domain), but not by other proteins of TBEV. It appears that TBEV NS5 activates the IFN regulatory factor 3 (IRF-3) signaling pathway in a manner dependent on RIG-I/MDA5, which causes the nuclear translocation of IRF-3, and its subsequent binding with the RANTES promoter [53]. As mentioned above, the CCL5 (RANTES)-mediated migration of blood monocytes and T lymphocytes may contribute to brain damage during TBE [53]. HBCA infection with another flavivirus, Zika virus, is associated with only limited activation of immune cytokine/chemokine response, but ZIKV infection induces the highest upregulation of CXCL10, IL-6, 8, 12, and CCL5 (RANTES) [55], which corresponds with the observations in TBEV-infected HBCA. Overall, the available data indicate that astrocytes are an important source of various pro-inflammatory cytokines and chemokines in the TBEV-infected brain. Some of these soluble factors, such as CXCL10 and CCL5 (RANTES), are known to mediate neuronal damage, which can have important pathological consequences during TBE.

We next investigated how the pro-inflammatory cytokines released from TBEV-infected SK-N-SH cells influenced naïve HBCA. Surprisingly, treatment of naïve HBCA with virus-free culture media from TBEV-infected neuroblastoma cells did not increase the production of any of the analyzed soluble factors. A study of WNV demonstrated that neurotoxic mediators released from infected SK-N-SH can activate astrocytes, as measured by a significant increase in GFAP expression, which was also associated with significant upregulation of inflammatory cytokines in naïve astrocytes [54]. On the other hand, we found that treatment of naïve SK-N-SH with virus-free culture media from TBEV-infected astrocytes resulted in increased production of several cytokines and chemokines—particularly the chemokines CXCL10 and IL-8, the growth factors FGF and G-CSF, and a number of pro-inflammatory cytokines. While TBEV infection of SK-N-SH increased the expression of only two soluble factors (IL-8 and CCL5), treatment of the same cells with cytokines/chemokines produced by TBEV-infected HBCA induced much stronger immune activation. Although TBEV-infected HBCA produced some soluble factors with known neurotoxic properties [47, 51, 54], supernatants derived from these cells caused no toxicity towards treated SK-N-SH cells. Notably, this result does not necessarily mean that cytokines/chemokines exert no neurotoxicity during TBE; it is most likely related to the concentration used in our experiments. This issue warrants further experimentation, ideally with primary neurons, which is beyond the aim and purpose of our present study. Interestingly, treatment of SK-N-SH with supernatants from TBEV-infected HBCA (at early time-points), and treatment of HBCA with supernatants from TBEV-infected SK-N-SH led to reduced viral growth and increased survival of the infected cells. Our data do not enable us to pinpoint the specific mediator derived from infected SK-N-SH or HBCA that contributed to reducing viral growth and virus-induced cell mortality; however, CXCL10, RANTES, IL-8, and IL-6 are likely candidates. Our next investigations will focus on the specific roles of these soluble factors in TBEV infection.

Immunopathological reactions in the brain make substantial contributions to the neurological sequelae during flavivirus encephalitis [56]. Considerable research efforts are now focused on targeting pro-inflammatory cytokines and chemokines as a novel and promising therapy for various diseases, including viral neuroinfections [56, 57]. For example, it may be possible to reduce immunopathological reactions in the brain by neutralizing inflammatory cytokines or antagonizing their receptor function [57]. Our present study is important because the results delineate the major inflammatory cytokines and chemokines present in the serum or brain tissue during infection in the TBE mouse model and provide a description of their production kinetics. Moreover, we identified both neuronal cells and astrocytes as one of the potential sources of these soluble factors in the infected brain. Future studies are needed to investigate the suitability of the identified cytokines/chemokines as potential targets of immunomodulatory therapy.