Human TANK-binding kinase 1 is required for early autophagy induction upon herpes simplex virus 1 infection.

To the Editor: Mutations disrupting the Toll-like receptor 3 (TLR3)–dependent–interferon pathway can underlie herpes simplex encephalitis (HSE) of childhood caused by herpes simplex virus 1 (HSV1) infection. These otherwise healthy patients with HSE carry germline mutations in the TLR3-interferon circuit, including TIR domain–containing adapter–inducing IFN-b (TRIF) and TANK-binding kinase 1 (TBK1). Their dermal fibroblasts show impaired interferon production after HSV1 infection and polyinosinic-polycytidylic acid (poly[I:C]) stimulation. A number of these genes (TLR3, TRIF, and TBK1) have also been implicated in the process of autophagy. On the other hand, HSV1 is known to antagonize the antiviral interferon pathway and the autophagy machinery in part through TBK1. Specifically, TBK1 is targeted by the virus-encoded proteins ICP34.5, ICP27, VP24, and UL46, compromising antiviral interferon signaling. In the context of autophagy, TBK1 has been reported to phosphorylate autophagy receptors, such as p62, to promote clearance of intracellular pathogens, including HSV1 in vitro. Here we study the role of autophagy in HSV1 infection using dermal fibroblasts from healthy controls and HSE patients with autosomal dominant TBK1 (p.G159A/WT) and autosomal recessive TRIF (p.R141X/R141X) deficiencies. Despite showing normal autophagy activation after rapamycin and poly(I:C) stimulation, TBK1 fibroblasts showed no induction of autophagy aftermultiple stimuli: cyclic diguanylatemonophosphate (c-di-GMP), HSV1 60mer–double-stranded DNA (dsDNA; 60mer-dsDNA), and HSV1 infection. After rapamycin, LC3B (microtubule-associated protein 1 light chain-3B) punctate signal increased by 3-fold in both control (media, 20.0%; rapamycin, 72.3%) and TRIF (media, 21.8%; rapamycin, 70.0%) fibroblasts and by 6-fold (media, 11.0%; rapamycin, 61.1%) in TBK1 fibroblasts, suggesting that TRIF and TBK1 were not required for rapamycin-induced autophagy (Fig 1, A and B). To assess autophagy induced by means of TLR3, poly(I:C) was used to stimulate fibroblasts, leading to a 12-fold (media, 7.1%; poly[I:C], 86.5%) increase in LC3B puncta in control fibroblasts. TRIF fibroblasts were unable to induce LC3B puncta, implicating TRIF in poly(I:C)-induced autophagy. However, TBK1 fibroblasts showed a moderate 8-fold (media, 4.0%; poly[I:C], 31.3%) induction of autophagy, suggesting its partial role in poly(I:C)-induced autophagy consistent with its partial impairment of poly(I:C)-induced interferon production (Fig 1, A and C). Although the role of the dsRNA TLR3 pathway in regulating autophagy has been documented in other cell lines, its involvement in infection remains elusive. In addition to TLR3interferon signaling, TBK1 is also involved in the HSV1 DNA recognition pathway through stimulator of interferon genes


To the Editor:
Mutations disrupting the Toll-like receptor 3 (TLR3)-dependent-interferon pathway can underlie herpes simplex encephalitis (HSE) of childhood caused by herpes simplex virus 1 (HSV1) infection. These otherwise healthy patients with HSE carry germline mutations in the TLR3-interferon circuit, including TIR domain-containing adapter-inducing IFN-b (TRIF) and TANK-binding kinase 1 (TBK1). 1,2 Their dermal fibroblasts show impaired interferon production after HSV1 infection and polyinosinic-polycytidylic acid (poly[I:C]) stimulation. A number of these genes (TLR3, TRIF, and TBK1) have also been implicated in the process of autophagy.
On the other hand, HSV1 is known to antagonize the antiviral interferon pathway and the autophagy machinery in part through TBK1. Specifically, TBK1 is targeted by the virus-encoded proteins ICP34.5, ICP27, VP24, and UL46, compromising antiviral interferon signaling. 3,4 In the context of autophagy, TBK1 has been reported to phosphorylate autophagy receptors, such as p62, to promote clearance of intracellular pathogens, including HSV1 in vitro. 5 Here we study the role of autophagy in HSV1 infection using dermal fibroblasts from healthy controls and HSE patients with autosomal dominant TBK1 (p.G159A/WT) and autosomal recessive TRIF (p.R141X/R141X) deficiencies.
Using immunofluorescence imaging, we found that HSV1 infection triggers 2 LC3B phenotypes in control fibroblasts: perinuclear LC3B puncta in infected cells and cytoplasmic LC3B puncta in antigen-negative plaque-neighboring (''antigennegative'') cells (Fig 2, A). Although the former occurs later in infection and is likely the phenomenon termed nuclear envelopederived autophagy because it also stained with LC3A (Fig 2, A), 8 cytoplasmic LC3B formed early in infection (up to 3 hours after infection; Fig 2, B). Strikingly, TBK1 1/2 fibroblasts did not form cytoplasmic LC3B puncta in antigen-negative cells, despite being able to form perinuclear LC3B later in infection (Fig 2, A and B). Furthermore, inhibiting TBK1 in control fibroblasts using BX795 resulted in significant reduction in cytoplasmic LC3B formation (see Fig E1 and this article's Methods section in the Online Repository at www.jacionline.org). Although the lack of early autophagic induction was specific to TBK1 1/2 fibroblasts, TRIF 2/2 fibroblasts only showed delayed induction of autophagy (see Fig E2 in this article's Online Repository at www.jacionline.org), suggesting its partial involvement in HSV1-induced autophagy.
These results show that the 2 types of autophagy differ in localization (cytoplasmic vs perinuclear) and temporal response to HSV1 infection, implying that they have different functions. We decided to focus on the TBK1-dependent early cytoplasmic phenotype as the later perinuclear LC3B, likely nuclear envelopederived autophagy, was induced in all cells and has been reported to be a generalized stress response to viral late protein production. 8 We next sought to understand how the different triggers of autophagy affect HSV1 infection. After pretreatment with poly(I:C), HSV1 replication was significantly reduced in control fibroblasts (nontreated: 5.8 3 10 5 , poly[I:C] treated: 5.5 3 10 4 , P 5 .006), which can be attributed to production of IFN-b (Fig  2, C and D). Consistently, with a low dose of HSV1, no viral plaque was observed in control fibroblasts, which exhibited cytoplasmic puncta in response to the poly(I:C) treatment (Fig 2, C-E and G). Interestingly, poly(I:C)-induced LC3B puncta in TBK1 1/2 fibroblasts was detectable after HSV1 infection. However, this pre-enhanced autophagy and IFN-b production in TBK1 1/2 fibroblasts did not improve cell viability or viral replication in contrast to control fibroblasts (Fig 2, C-G). TRIF 2/2 fibroblasts did not induce autophagy or interferons after poly(I:C) treatment 1 and hence were not protected against HSV1 infection (Fig 2, D). Notably, however, cytoplasmic LC3B puncta was present following HSV1 infection of TRIF 2/2 fibroblasts, confirming that formation of cytoplasmic LC3B puncta is interferon independent (Fig 2, C, D, and G, and see Fig E3, A, and E4).
In conclusion, we show that in addition to its antiviral role in interferon production through TLR3 and STING, 1,2,6 TBK1 induces autophagy following HSV1 infection. We demonstrate that TBK1-induced autophagy occurs early during HSV1 infection in antigen-negative fibroblasts, can be mediated by c-di-GMP or HSV1 dsDNA, and is TLR3 and interferon independent. TBK1 1/2 fibroblasts derived from a patient with HSE harboring a dominant negative mutation had a selective impairment of autophagy induction early in infection represented by the lack of cytoplasmic LC3B puncta formation. We believe that host or virus-induced factors, possibly acting as danger signals, can trigger autophagy in antigen-negative fibroblasts, promoting cell survival without influencing viral replication. This study highlights a possibly cytoprotective role for TBK1 in HSV1induced autophagy, which might serve to control inflammation and has potential implications for patients with HSE.

Viral infection and quantification
Human fibroblasts were infected with HSV1-green fluorescent protein (KOS strain with green fluorescent protein-tagged capsid protein VP26) or HSV1 (strain 17AR1) at various MOIs and time points for immunoblot and immunofluorescence experiments. After 1 hour of infection in DMEM supplemented with 2% FBS, the virus was removed, and fresh media added with 1% human serum. Viral titers were determined by infecting a confluent monolayer of Vero cells in a 12-or 96-well plate and performing a plaque assay or calculating the 50% end point (TCID50/mL). E3

Cell viability assay
Cells were plated in a flat-bottom 96-well plate in triplicate at a density of 0.18 3 10 6 cells/mL in 10% FBS-supplemented DMEM. Fibroblasts were pretreated for 16 hours before infection with HSV1 (MOI 1) for 24 hours. Fibroblast viability was measured using the CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (MTS) Kit (Promega, Madison, Wis) and performed according to the manufacturer's instructions. Cell viability was determined by normalizing to noninfected cells of each cell line.

Statistical analysis and software
Immunofluorescence images were analyzed with Icy software. Densitometric analyses of immunoblots were carried out with ImageJ software (National Institutes of Health, Bethesda, Md). Statistical significance was assessed by using 2-way ANOVA or the Student t test with Prism 7 software (GraphPad Software, La Jolla, Calif). Numbers of fluorescent cells or LC3B puncta-positive cells were quantified with ImageJ software. All experiments were performed at least 3 times.
FIG E1. TBK1 inhibition reduced cytoplasmic LC3B puncta formation but did not affect perinuclear LC3B formation after HSV1 infection. A, Control and TBK1 1/2 fibroblasts were pretreated with 1 mmol/L BX795 for 16 hours before infection with HSV1 (MOI 10) for the indicated length of time. Cells were fixed and stained for LC3B (green), ICP4 (red), or both. 49-6-Diamidino-2-phenylindole dihydrochloride (DAPI; blue) was used for nuclear staining. The scale bar of each representative image is 20 mm. The inset represents the magnified view of the indicated area and has a scale bar of 10 mm. White arrows indicate cytoplasmic LC3B, whereas yellow arrows indicate perinuclear LC3B. B, The percentage of cells positive for cytoplasmic LC3B puncta in Fig E1, A, was counted on at least 100 cells. Images are representative of 3 independent experiments (n 5 3). Data are represented as means 6 SEMs and analyzed by using 2-way ANOVA. **P < .01.

FIG E2
. TRIF 2/2 fibroblasts showed delayed cytoplasmic LC3B puncta formation. A, Fibroblasts grown on coverslips were infected with HSV1 (MOI 10) for indicated lengths of time before being fixed and stained for LC3B (green), HSV1 ICP4 (red), or both. 49-6-Diamidino-2-phenylindole dihydrochloride (DAPI; blue) was used as the nuclear stain. The scale bar of each representative image is 20 mm. The inset represents the magnified view of the indicated area and has a scale bar of 10 mm. White arrows indicate cytoplasmic LC3B, whereas yellow arrows indicate perinuclear LC3B. B, The percentage of cells positive for cytoplasmic LC3B puncta was counted on at least 100 cells. Images are representative of 3 independent experiments (n 5 3). Data are represented as means 6 SEMs and were analyzed by using 2-way ANOVA. ***P < .001 and ****P < .0001.
FIG E5. Interferon-induced autophagy in fibroblasts. A, Control, TRIF 2/2 , and TBK1 1/2 fibroblasts were stimulated with 1 3 10 5 IU/mL IFN-a-2A for 24 hours before being fixed and stained for LC3B (green). 49-6-Diamidino-2-phenylindole dihydrochloride (DAPI; blue) was used as the nuclear stain. The scale bar of each representative image is 20 mm. The inset represents the magnified view of the indicated area and has a scale bar of 10 mm. White arrows indicate LC3B puncta. Images are representative of 3 independent experiments (n 5 3).